Hacker News new | past | comments | ask | show | jobs | submit login
Patio11's Law: The software economy is bigger than you think (secondbreakfast.co)
1035 points by secondbreakfast on May 13, 2020 | hide | past | favorite | 422 comments



I remember reading this article about SAP a few months ago, and it just blew my mind: https://retool.com/blog/erp-for-engineers/

Here was an entire market segment I'd never heard of, and just the market leader is worth $163B. And on top of that, the technology is so totally alien to anything I'd ever heard of before. It's like an alternate-history kind of thing. Like if we discovered a continent somewhere in the Pacific that had never been in contact with our civilization and had steam-powered airships or some crap.

And I'm sure this is just one of many such industries, even just within the umbrella of software.


That is fascinating to me, because virtually all my friends have heard of SAP outside of IT - as users or recipients. My engineer friend uses it for supply chain management, a beer store hq clerk helped implement it, construction accountant has used it, etc.

At the same time, I guess you have to be at a company of a certain size for it to be relevant to your life... But that perhaps proves the point - HN is very starup focused, and occasionally in a bit of a bubble (which is fine - I come here to get away from my daily mainstream IT grind and see excited people talk about cool things:), yet a hundred reasonably successful startups won't employ or touch near as many people as a single "quietly" successful traditional behemoth.


And if it helps understand the ecosystem and its seeking oddness, this is the crucial part of article:

"Implementing ERP isn’t just a purchasing decision: it’s committing to redoing how you handle operations. Installing the software is the easy part; adjusting the whole company’s workflow is the most of the work."

Don't think of erp and implementing it as a software development challenge. Sure, Developers can fscrew up*, an erp implementation but they cannot make it succeed - it is fundamentally a business transformation endeavour and must be undertaken as such to have a chance.


A good way for your organization to completely botch an ERP implementation is to fight the system every step of the way because your business is 'special'. Many people end up abusing customizations to the point of trying to emulate their old system in the new one, and nobody wins.


You just brought back some painful memories. At an old company of mine, I helped migrate a legacy system to SAP (of all things!), and they fought it every step of the way.

Well, management did. Every other phrase they uttered was "The old system did it this way" or "I used to just press F2". The pushback to learning something new was unreal.

Nevermind the fact that multi-step processes were reduced to a simple button click. Or that inventory and customer management were orders of magnitude better. The old system!! The old system did it another way!


I feel like you could replace "ERP implementation" with "business transformation project involving vendor software" and have a new Law.

The same with the parent commenter's thoughts, "Sure, developers can screw up [business transformation projects], but they cannot make them succeed."


> A good way for your organization to completely botch an ERP implementation is to fight the system every step of the way because your business is 'special'.

More generally, a good way to botch a business transformation project is to select technical solutions without a keen grasp of how they fit your business need, and to bullheadedly stick with those solutions as friction is revealed with the business needs.

Both distorting the business around vendor software and ending up continuously fighting with the vendor software throughout th effort (both of which are commonly cited “how to fail” routes) are clear (though superficially opposite) symptoms of that same failure occurring upstream in the process.


Written like a successful salesperson for competing ERP solution X, well done!


This is why it's important to tackle this from both ends:

- From the ERP end: you want it to be as flexible as possible so that it can reflect how the business actually operates and should operate in the future

- From the organization's end: you want to identify not only how the business currently operates, but how you would want it to operate in an ideal world


This reminds me of something I read while helping a business major friend with a case study a while back. One of the readings she had as background mentioned that companies like SAP that sell ERP are almost selling the process/workflow more than the actual software.

Part of me wonders if these things are more of a "no one ever got fired for buying IBM" thing than anything else. Are people buying the workflow because it's what they really need, or are they buying it because everyone else uses it?


1. You're absolutely right - you're buying software, a set of best practices, workflows, and a proven way to implement and use them.

2. HR, Finance, all the backoffice stuff - that's not your competitive advantage. It's not your core business. You just want to do it as efficiently and quietly as possible.

3. Therefore, in ideal case, a company realizes it is not a special snowflake, buys an ERP, implements it as vanilla as possible, and quietly goes on focusing on what IS their competitive advantage / core business... rather than spending years, millions of dollars, and tremendous risk, by trying to customize ERP to some obscure different process they just "must" have :-/


that happens a lot with enterprise software, too much customization and you end up painting yourself in a corner or run into roadblock after roadblock of platform gotchas. I do SF implementations these days and have seen it go great and have seen it turn into a total nightmare.

however, not once have I ever heard of an SAP implementation going well. It's a massive undertaking and takes total commitment from so many different people and roles in an enterprise. You need genius PMs and BAs along with a genius client team before you even talk about the actual technical people.


>>not once have I ever heard of an SAP implementation going well.

Sometimes it depends who you ask; I found that hands-on people on the ground and SMEs tend to be very negative about it, for personally valid reasons - it's a lot of work, it doesn't go smoothly, and it disrupts and changes what and how they do it.

However, there are quite a few ERP implementations where the top-level stakeholders feel the results are worthwhile the disruption and pain, and from that perspective it "went well" - the business objectives were overall met, whether it be capabilities or speed or reporting or integration/centralization or headcount or complexity reduction, all of it ultimately boiling to money at some level...


Now you're describing an implementation that is disruptive to the majority of employees "on the ground" and only worthwhile to the top-level stakeholders.

Only the top-level stakeholders felt the results were worthwhile the disruption and pain -- that they did not suffer.

But have you ever heard of one going well ?

As in, not just worthwhile to a tiny minority of the people on the top?

Yeah I bet it makes stuff more efficient, but clearly there's a disparity between the people suffering "disruptions and pain" and the ones who ultimately feel it has been worthwhile.


Another way to look at it is, the top-level stakeholders are responsible for the business and not just themselves or the lives of their employees. An effective leader is thinking about what is best for the business. Not just themselves or any one part of the business.

And maybe, just maybe, what is right for the business is bad for the majority of the staff. At least in the short term. I know that is a bitter pill to swallow, but it is reality.


> And maybe, just maybe, what is right for the business is bad for the majority of the staff. At least in the short term. I know that is a bitter pill to swallow, but it is reality.

Perhaps an example more palatable on HN would be Uber replacing all its drivers with software. Clearly better for the business.


The process often derails at exactly the “implement it as vanilla as possible” stage. Usually because someone doesn’t want to change what they do and someone else doesn’t care enough to make them change.


Another view of this - if the true purchase is of a group of workflows, are they purchasing for the things they know they need or the assurance that behemoth IBM/SAP/whatever has a process for whatever they didn't think about? Purchasing to minimise some future operational risk.


This is probably an understated benefit a lot of us in tech don't think about.

If you're selling widgets, and selling 100 units a month, there are growing pains on the path to 100 units a day and 100 units a minute that you can't anticipate, and that billion-dollar retailers have already figured out. So I think to the extent that physical B2C sales is pretty constant across industries, you know that if you're big enough to justify the investment in SAP or a similar ERP platform, there exists a process that you'll need 0.5, 1, or 5 years from now that they have and you can't really articulate today.


Sounds like Agile to me.


Some industries need your system to be certified for a software system to consume their data/services. You don't have much choice.


A lot of HN topics are about tooling for startups. Those tools are sometimes startups themselves, which can can make HN feel like a bit of a bubble I think.

Software for software developers can be a huge focus here, but it's just scratching the surface of software for the rest of the world. I was startled when I applied for a job at a company in my city, and they told me their product was entirely API connectors for the local banks. In fact all of the companies I've worked for here have been making software for local industries! From solar installers, energy efficient lighting manufacturers, local commerce and POS systems, mechanics software, creative market places, local lead generation.

I think if you want to start a business from scratch and you're not in SV, it's worth targeting those problems instead of problems that SV/startups have. You might not get sexy VC capital, but you might be able to jumpstart with some clients up front.


You won't need sexy VC capital if you can build a business that runs a good profit. Honestly I'm getting more and more cynical of the SF bubble that doesn't seem to be about making great software, but about getting a lot of investor money and going public - prioritizing short-term profits at the expense of the consumer and employees.


SAP was particularly successful because not only did they make money, but so did the Systems Integration partners.

At Accenture, for example, there were projects where 30-40 full time consultants were flown in from all over the country to do big SAP implementations.

...so you could see how Systems Integration partners heavily sold SAP products.


But my Blockchain Deep Learning Dog Walking app is going to change the world!! I went to Stanford so I must be right! /s


Dog walking is not a viable matchmaking platform business model because successful matches cut out the matchmaker, you didn't get the memo?


Humor fella.. he is kidding!


Have you seen your local software? It is rarely about making great software.. good enough.

Open source tries to make great software. Does the best job.


When good enough software solves the right business problem, it frees up resources to solve the next most important business problem. Good enough software runs many great businesses.

The one unicorn I worked for had so much cash to spend, there were at least 3 teams of 8-10 engineers on the same floor trying to solve similar business and infrastructure problems with different combinations of open source software.


Agreed. Running a company with a sustainable business model and healthy profits, not being propped up by VC money would probably do the local economy of one such a city perhaps more good.


"My engineer friend uses it for supply chain management, a beer store hq clerk helped implement it, construction accountant has used it, etc."

Is SAP something that can be somehow scaled down? if this is used for a beer store, it has to be really large store, if that's what is written in the cited article is true:

"A basic installation of SAP has 20,000 database tables, 3,000 of which are configuration tables. In those tables, there are ~8,000 configuration decisions you need before even getting started. And that’s why SAP Configuration Specialist is an actual job title!"


Ah,I guess context would help - "The Beer Store" is Ontario's [Corrected From "Canada"] monopoly province-wide beer retail chain.

You are absolutely right, a single retail store would not use such an ERP :)


Er, an Ontario-wide beer retail chain, owned by the province-wide beer wholesale monopoly, in turn owned by a cartel of foreign breweries.

Let me guess, you live in Toronto?


> an Ontario-wide beer retail chain, owned by the province-wide beer wholesale monopoly, in turn owned by a cartel of foreign breweries.

good thing we have SAP to facilitate the development of large scale business, such as monopolies and cartels.


I don't think the existence of SAP is particularly relevant to a monopoly that exists because of statute. Although I don't live in such a state, I know there are places in the US where all the liquor stores are government run.


Corrected! Thx :)



To add some context (for your curiosity), I'm not in The Valley and I'd never worked at a true "startup" until very recently, but I have only ever worked at small (<100 people) companies. Except, that is, for the summer I interned at Hewlett-Packard. Not sure how it never came up there; maybe they just have their own equivalent system.

I had actually heard the name SAP come up in passing at my last company, who had lots of ginormous clients. But all I'd really gathered before reading the article was "it's some kind of database, or something".


> Except, that is, for the summer I interned at Hewlett-Packard. Not sure how it never came up there; maybe they just have their own equivalent system.

When was this? Leo Apotheker became CEO of HP in 2010, after 22 years at SAP (two as CEO). He then presided over a sequence of colossal fuck-ups, lost 40% market cap, and was fired less than a year later.

HP definitely knows about SAP, but I wouldn't be surprised if they're not big fans.

Correction: 22 years at SAP, not as CEO.


He ran sales arms of SAP for 20 years, not SAP itself. After he became CEO of SAP itself he was fired after 2 years, then he joined HP amd he was fired from that after one year. Pretty rough end to his career there..


... or maybe it's the perfect ending?

I'm not sure about his SAP earnings, but he sold HP an 11 month CEO for upwards of 27.5M[0]. That sounds like the perfect way to end your sales career ;-)

[0]https://www.nytimes.com/2011/10/01/business/lets-stop-reward...


Promoted to the level of his incompetence perhaps.


Both hp and SAP are absolutely gargantuan companies (try guessing how many employees each had by the end of his tenure- unless you actually know I'll bet you'll underestimate it). They are also old (for computer related companies at least) and have lots of products for lots of clients.

Not to say he wasn't bad at it (I honestly don't know) but it's a hard job to get right.


SAP goes deep. It goes as far as to the factory floor when the technician is measuring if the piece is within tolerances and from Mitutoyo caliper sends the data into some kind of database, or something ;)And there it is, on a screen from 1994, SAP proudly on the screen. It's everywhere.


I worked for EDS from late 90s to early 00s and SAP was _everywhere_. Projects were scoped and framed as "how can we fit this into SAP?" - and not "how can SAP help this project?"

SAP is/was virulent in nature. If you used it for one thing it damn near forced you to use it for everything. A significant chunk of the work EDS were undertaking (as IT outsource) was migrating to SAP.

Our timesheets, project plans, data, hell even a huge chunk of our app development was done in SAP. We were paying through the nose for SAP contractors too.


That... explains a bit of the herding cats commercial.


> At the same time, I guess you have to be at a company of a certain size for it to be relevant to your life... But that perhaps proves the point - HN is very starup focused

That’s pretty dismissive. I work at a company with ~80B in revenue, I never interacted or heard of SAP in work context. In fact it is only incidentally that I’ve heard of them previously. There are hundred of thousands of engineers working in companies with 10-11 digit capitalizations that will never interact or hear about SAP.

Edit: to clarify because I am getting a bunch of downvotes. My point is not about SAP’s penetration, it is that lots of engineers (even in companies that use SAP) might not encounter SAP products. The perspective “of you haven’t worked with SAP you are in a small org” is wrong.


-“ The perspective “of you haven’t worked with SAP you are in a small org” is wrong.”

I would fully agree with that, my generalization was kind of the opposite - if you work at a small company you are unlikely to use or work with SAP.

My other generalization was that HN was focused on startups, their technical culture, the cool and interesting, and almost by the definition not on boring bread and butter aspects. Does not mean that people/users themselves do not work there - I have spent twenty years on PeopleSoft/Oracle myself, but that is not why i come here and yes, the overall zeitgeist on HN toward e.g. Oracle is pretty consistent :->

FWIW, comment was not meant to be dismissive. I was genuinely fascinated, late at the night, at the notion that IT industry professionals might not be exposed to a behemoth such as SAP or even ERPs at large, and was pondering as to how that might come to be; there's absolutely personal bias & hubris in it, which is why I tried to understand where it came from and what's the diff between background of my friends and colleagues, vs perhaps some other IT professionals :)


> The perspective “of you haven’t worked with SAP you are in a small org” is wrong.

The fact that you are arguing against this might be why you are getting downvoted. The original comment was saying “if you are in a small org you probably haven’t worked with SAP”, which is not the same thing.

Edit for further clarification: I can see how it might look like they were arguing what you stated, but it was more of a Bayesian inference. Many HN readers are at a small company, therefore a plausible reason that an arbitrary HN reader might not have heard of SAP would be that they are in the large pool of HN readers that are {in small companies and, therefore, do not work with SAP}


You know, I actually like your Bayesian interpretation. My main dislike was the dismissive tone in OPs phrasing, but I think his proposition might be OK.


There is nothing remotely dismissive in the (100% accurate) sentiment that small organizations have no need or desire to use large Enterprise-among-Enterprises applications such as SAP.


Your firm (KP) does use SAP's BusinessObjects business intelligence software along with SAS analytics tools and to "support its data analysis activities against it's Oracle EHR system".

Since you are in the machine learning silo, your perspective may be a little skewed.


To be fair BusinessObjects is not SAP. It's just one of its acquisitions.


Well, if you look at it that way then Powerpoint is just one of microsoft’s acquisitions. After a certain period an acquisition becomes fully native to that business. In my view BusinessObjects passed that point with SAP a while ago.


Sorry, you are right, I should have been more specific. What I meant was BusinessObjects is only the reporting component they acquired to complement their product suite. It's not "the product" SAP sells (i.e. ERP).


"SAP" isn't a product either, it's a company name. When certain people talk about SAP they actually mean R3 or ECC. R3 is dead and the sun is setting on ECC which is being replaced by S4/HANA (which some people will still call SAP). SAP has a lot of products and it's main products have a lot of modules so two people using "SAP" may not even recognise each others versions.

It would not surprise me if KP were using ECC for their finances or some other area the parent comment is not familiar with.


Sure, but lots of people don’t interact with it and would not know about it. The fact a company uses it doesn’t mean most engineers would know about it.


Companies that make things often have components managed by the ERP system. This flows down to the engineers. A software shop never deals with such things.


There are also alternate vendors of ERP. Microsoft Dynamics NAV for example. At my current shop I manage INFOR VISUAL ERP.


...Microsoft Dynamics GP, Microsoft Dynamics AX, all entirely different products for different sectors. Then there is Netsuite, Sage (lots of different products), SAP Business One (small SAP), Oracle ERP, JD Edwards...plus 1000 niche providers who are often migrating to NAV rather than their own product. It may be a shock to a FAANG, but the ethos is very much, 'move slow and don't break anything'. 10 year old versions still supported, that sort of thing. Another reason Windows is popular in enterprises is that ERP clients written in VB6 (or whatever) still run on Windows 10!


Windows is popular in enterprises for like, a million reasons. As a Windows Systems Administrator, The entire Windows ecosystem is built around business and business processes and management.


And was that a FAAMNG? Presumably the larger, "new-wave" tech companies roll their own solutions of some kind


While Microsoft does sell their own SAP competitor they actually use SAP for their main accounting.

https://www.erpsoftwareblog.com/2012/11/why-does-microsoft-h...


I'm pretty sure I've seen job adverts from Google for people with experience of Oracle Hyperion.

What would be the point of Google writing their own financial consolidation software?


You'd think so... I thought so once... but again, perhaps not necessarily - ERP is NOT their core competency.

I.e. HR & Payroll legislative rules and best practices, business & organizational process transformation etc, are not what Google Engineers and IT necessarily have experience in, and if they allow hubris to override that fact, it might not go great.

Internets indicate that Google uses Oracle, Apple uses SAP.

This info may be outdated - for example, Facebook claims they built their own ERP just recently; I do wonder what that "backend database" part is though, or how much HR/Pay/Finance, as opposed to inventory/datacentre management, is custom: https://techcrunch.com/2019/08/15/how-facebook-does-it/


I work for a large healthcare company. My point is not whether we use it (as another poster pointed out we apparently use a product that was acquired by SAP), it’s that lots of people might not know all of the business software that is used for various reasons.


How many more letters are we going to add to this acronym?


Do you fill in a timesheet using a computer ? If so, there will be SAP or another ERP system behind it.


SO what ERP system are you using? What do you sell? How do you manufacture? How do you handle inventory and finances?

You're either using SAP and you don't know it or you're using Oracle.

EDIT: just looked at your linkedin profile. I can see that your company literally has SAP consultants working for them full time. I don't know which SAP products you are using but you definitely are an SAP customer.


> You're either using SAP and you don't know it or you're using Oracle.

Or you're using NetSuite (which nowadays is Oracle, granted, but a rather different product from what most people think of when they talk about Oracle or SAP in an ERP context), or you're using Microsoft's ERP (Dynamics something or other? Never used it), or you're using something developed in-house (very likely running on an AS/400). Or hell, you might be running multiple ERPs in a weird limbo state mid-transition (e.g. Thermo Fisher, which was using AS/400 and SAP (or maybe it was Manhattan? Or both? My memory's fuzzy...) - the latter of which might very well have been partially implemented on AS/400 in order to integrate with the existing ERP - as of a couple years ago when I was working for them).

It's worth noting, too, that there are plenty of large and arguably-dysfunctional companies not using an ERP at all, or (like with my current employer) it's relegated to very specific uses because other functionality got offloaded to other systems (for example, it's relatively common to run a WMS separate from the ERP and integrate between the two, as it is with an eCommerce system; in my employer's case, all three are separate).


I have heard of sap in tons of contexts in undergrad business classes, news articles, and things like that. Though I have never come across an implementation in the wild. Though my experience in large enterprise is limited.


I'm in Germany and of course everyone here has at least heard of SAP and knows it's huge, but I guess not many people know just how huge. SAP consultancy is a big market here with lots and lots of job openings. I knew a few people who did that, it's usually a very well paid job, but it's well paid because otherwise no one in his right mind would do it. From what I heard, it's incredibly boring while at the same time incredibly stressful when you're on-site. Also, I've never met anyone who actually liked SAP.


I've heard very similar things about being a SAP consultant here in Europe - basically, it sounds like a job for someone who:

- understands technology, but is not a big fan of it

- can communicate with clients

- can handle stress well

- is into salary maximization


Doesn't mind living out of a suitcase.


A friend was telling me how they make over $200 an hour. For a job that doesn't require a technical chops of a fullstack dev, that sounds crazy to me but doable.


SAP consulting is like some sort of mythical club,where only the chosen ones can join. SAP isn't something one can learn from his bedroom and then go around advising multinationals.You have to somehow get trained on a job where it is used. So the vast majority of SAP consultants pretty much owe to the fact that they lucked into it.It is however, pretty lucrative though.


I liked SAP, especially compared to what I have used since. We have never met, so your point is still true! It might be some time before Germany lets us English types in again!


Want your mind to be blown even more? I worked for a multinational company, I would not be surpised that this company is among the top 10 SAP customers worldwide. Well, everybody hated hated hated SAP, it was a huge hindrance instead of any help.Any use beyond of a "dumb" database was a huge PITA, and I am not fully convinced that all in all it helped more than damaged the business. It was clunky, unintutive, slow, ugly, archaic and a long etc. There were these guys, that were trained as "experts" but they frankly behave more like witch doctors. SAP is coasting just by its name and legacy but from my POV is ripe for the taking. It is not easy though, ERP systems are hard, and changing them ever harder, but in my fantasy world I would like to develop some program and would be very happy to skim 0.00001% of their income.


SAP's ugly green, old style-UI screens are wonderfully optimised for data-entry. You can be ultra-fast. It's like VIM. It's like driving at 300 mph.

Its all the new fangled dumbed-down, web-ui's that is sheer crap and leaves your head scratching.

Well..the old UI's were coded by old-school, C programmers who started their career in the 70s and 80s. They even wrote a multi-process Erlang style kernel to back the application programming. Performance came naturally from their gut.

Nowadays, we get Electron and congratulate ourselves.


Saw a big write-up somewhere about this. It's easy to learn the muscle memories for various patterns on a text only UI, and you can become extremely fast.

But almost no one is fast with a mouse based GUI


Excel.

Indeed.


I did once get an explanation of how SAP screens work when a former employer was considering buying SAP ERP - it reminded me at the time of how IBM mainframe terminals work.


I worked at an $8bn multinational and we ran the whole operation through SAP from quote to cash... and we really liked it. It was consistent, and the 'ugly' green-screen style data entry is wonderfully efficient. I hear this all the time though, about all ERPs. It is all about how you use them.


Same experience here I work as a SAP Developer at a €14bn multinational.

I had to work in the store for a month and got good feed back from the workers there.

We try the smooth the user experience with web apps for tasks that are done often so it's not to confusing.

But the impotent data entry task like order entry and so on are done the old interface because it's more efficient.


Same with accounts generation.

Almost certainly most operations accounts and tax filings could be fully auto generated from their bookkeeping - if the bookkeeping was good enough.

You could do that by constraining the bookkeeping and adding some intelligence to the reporting.

So why do bookkeeping firms like Xero stop at basic bookkeeping? Because they sell via accountants - who make a lot of money piping data from one computer system into another to change the format a bit.

Enabling certain people to maintain their mystique is lower hanging fruit than pushing for full automation.

Plus who wants to program accounts systems...


'Almost certainly most operations accounts and tax filings could be fully auto generated from their bookkeeping - if the bookkeeping was good enough.'

I firmly disagree. I am supposed to be reviewing accounts right now....accruals, prepayments, deferred income, the difference between capital and expense, capital allowances, R&D allowances, dividends vs salaries, consolidation of groups. The average bookkeeper is not expected to work at this level. That all requires judgement and experience and expertise. That is why Xero stop at the easy bit


That sounds all pretty trivial.


I wonder if you have worked as an accountant? Or at least took some intro courses in the area? Or at a bare minimum took a look at common corporate financial statements?

This is such a typical dismissive reaction by a person who has ZERO insights into the domain.

If it is so trivial, why don't you do it and disrupt the market? Some very trivial millions are waiting for you to just come and get them, what's holding you back?


I don't do it because I care more about my time these days than about making millions :-) And why would I want to do something that is so obviously boring?


Trivial stuff takes what, an hour, a week, a month at maximum to do? So you are trading the opportunity to be financially independent for the rest of your life for what, your free time on a span of one month? Yeah, that does sound reasonable.


Good, then we agree :-)


Ostensibly you’d rather spend your time on HN than making millions.


Definitely! There is no doubt about that :-)


Trivial in the sense of integer multiplication being easy; non-trivial in that it's full of weird & wonderful business rules, many of which are barely written, and strange interactions between different markets and contracts etc.

Half of it is just the business & people skills to extract the necessary information. Harder than you'd think, and that's just asking people for stuff!


For basic bookkeeping, yes, you can generate all the needed reports and tax filings from the books. I'm in Europe and I've had a couple of companies where I had no accountant - I just entered all the transactions in the software and it generated all the necessary filings for me. Payroll is also auto-generated, taxes handled etc.

For complex accounting and analysis you still need accountants and business analysts. It's a different type of job in that case though: instead of handling the common cases, you handle the edge cases that the software is unable to deal with.


Harder than it sounds. A complete drop-in replacement requires an almost unbelievable amount of internationalisation and research. And even if you can get all of that lined up, B2B sales and marketing - and inertia - are a huge challenge.

It's not that it couldn't be done, but it's very much a moon shot and it's not going to happen without someone charismatic and experienced driving it and selling it.

Specific niche feature SAP-a-likes are more plausible on a smaller scale, but sales and marketing will always be the real challenge.


Good luck! You're time to market and the sunk cost fallacy are overwhelmingly against you.


There are smaller ERPs too, but I think often the real alternative is building something in-house tailored to your specific business.


That's basically what Salesforce did, but they ended up skimming a lot higher % than that


I've never seen an et cetera qualified :D


> but in my fantasy world I would like to develop some program and would be very happy to skim 0.00001% of their income.

Which is exactly what SFDC, NetSuite, etc have done. There's definitely room here.


Workday is in the process of doing this. I wish them all the best, but I can't help thinking that we're just replacing the old goat with a new goat.


I did that and was successful to some degree, but it's not easy. I would've done something else if I knew that back then.


You skimmed a small part of SAP’s market? Like, a specific niche that you built a better tool for?

I can only imagine that if you create any tool worth competing with SAP, your customers also have to be huge companies.


Yes, we competed with SAP in a specific industry and honestly our tools were better and more user friendly, and no, they were not all huge companies, which is one of the reasons we were able to eat SAP's cake because we were much cheaper. I believe we got the job 90% of the time when we were shortlisted with SAP.

We lost some customers to SAP later on because they become bigger and "consultants" told them that they should switch to SAP because, you know, all big companies use SAP. I haven't heard of any success stories from those customers yet (this is not entirely SAP's fault though).


Which SAP product did you compete with?


Coming from the opposite direction (mgmt consulting for big corps -> startup), I was very familiar with the ERP concept, but am still super confused by the ERP market. Every so often I look for an ERP for our company, but am always disappointed by the options. It seems there's an uncanny valley between Quickbooks and SAP, but maybe the issue is just one of discoverability.

We've slowly stitched together an assortment of SaaS tools, custom built modules, and Excel files in the exact way I swore we would never do. If anyone's been through this hell before, I would appreciate some help.


I have been doing ERP consulting for about a decade now and I tend to disagree. I think we have a lot ph ERP options now in the paid space apart from just Quickbooks and SAP. You need to pick and choose based upon what is the most complex operation for a company- be it multi currency financials, complex supply chain or some other parameter. Then you can pick from lowest complexity to highest in the following manner

* Quick books

* Sage

* Dynamics / NetSuite

* JD Edwards / Infor / Manhattan

* Workday

* Oracle EBS / SAP

Each of those on the same line are similar in features and complexity.

This is obviously not a complete list - just the ones I have dealt with.


Quickbooks will do you fine up to $100M in revenue or 8-10+ years in operation. When you cross either of those bridges, you should think about implementing a "real" ERP.

But this doesn't happen until your team (accounting/finance/CEO/CFO) start running into roadblocks in either piecing together financial statements, or the month close process starts taking weeks of manual labor, or there's a HUGE accounting issue and everyone discovers the forecast is way off and the company has been burning cash faster than expected.

At this point you have the organizational momentum to do the crazy amount of work required in getting an ERP up and running, which requires getting most of your business flowing into it so that it'll even be remotely useful (otherwise everyone goes back to Excel).

What needs to flow into it? Everything you're doing by hand to update your three sheets. You need to get revenue/income in there (sales orders, purchase orders etc.), product costs if you have them, payroll, bills (internet, office supplies). Hopefully you use something like Bill.com, and this becomes a (slightly) simpler process.

Maybe you have vendors in Europe where you pay in Euro, so you now also have to figure out how you're going to convert currencies. Maybe you have warranty on your products, so you have to balance accrual accounts. This list goes on, and on, and on.

Every company is different, and you might not even recognize the overlap between two ERP launches besides the fact that they're using SAP.


>Quickbooks will do you fine up to $100M in revenue or 8-10+ years in operation. When you cross either of those bridges, you should think about implementing a "real" ERP.

I think that very much depends on what market you are in. Sure if you are a Saas, but not if you are a European distributor with stock in two countries with different currencies. Stock adds massive complexity. You might be reaching for a lower end ERP at $5m of turnover.


Stock in two countries is nothing that can't be managed by hand without an ERP. I know, because I've done it before.

You usually reach in inflection point where your manual process sans-ERP is costing the company real money, and the ability to make quick(er) decisions is hampered by your current process.


It also depends to an extent on your legal accounting / reporting requirements.

Friends who had businesses in both Poland and Singapore tell me that the difference was like night and day.


Wait until you start accounting for stock compensation.


Yeah I thought the same, I've worked with, against and for SAP since I turned 21. My whole career has spent working on ERP on various platforms.


SAP is a German company so here everybody knows about it. There is even a startup strategy called "unbundling SAP" where you pick one feature of SAP, do it 10x better and target a smaller but still huge market. Examples for such companies are: Personio (500M valuation), Cellonis (2.5B valuation)


Sounds similar to the "unbundling Zapier" article that made its rounds here on HN a couple days ago. Definitely seems viable.


mind sharing a link of this article, can't seem to find it



A friend of mine worked for Sun in the late 90's and mentioned SAP to me.

He told me companies would buy top-of-the-line (overpriced) hardware and pay a premium to get SAP up and running, because they would plug their company into it and recoup their costs almost overnight. It would just do say currency conversions or other super-boring non-core-business things better and save a pile of cash.

You can't save yourself rich, but sometimes the difference between a profitable quarter and a horrendous loss might be your accounting software or your tax software...


Oh yes, there are tons of giants creating software for health applications you never heard of, defense purposes you cannot hear about, logistic you would never consider (managing competitions, port services, private jet traffic, truck fleets...) and so on.

And those don't use $LATEST_TECH, $FANCY_FRAMEWORK or $EXOTIC_LANGUAGE.

They use $BORING_TECH, $OLD_LANGUAGE, $DEPRECATED_FRAMEWORK.

They are not in the cloud, don't update often, have to deal with legacy stacks, stacked on a stack of legacy stacks, and administrative burden from left to right.

Those are the ones that pay for an Oracle licence, and sync it to dozens of excel sheets from an FTP and a lone Access DB, where all auth is manage by a Tomcat service actually proxying to Active Directory and their custom CRM.

And they certainly don't have a team of 10x programmers. In fact, most are windows dev that never used the terminals and don't have admin rights on their machine.

I should know, I make half of my income going from company to company to train their team.

They are everywhere.

They make billions.


SAP is amongst the least fun software category which is ERP systems.

Highly customised, to the point that upgrades are impossible.

Extremely complex systems. So complex that vendor lock in is assured, and the lock in means big prices.

And yet a good business because ERP systems are at the core of any medium to big business.

It's grindy, boring, challenging without being fun work, often using non standard technologies that have been developed by the ERP systems vendor, which help ensure lock in.

ERP companies like SAP and Oracle are the absolute best at making money on services - they know how to charge huge huge amounts and get away with it. There's TONS of money there.

I wouldn't want to work in that area though.


My sister-in-law graduated with a degree in Industrial Engineering, and got exposed-to/proficient-in SAP at one of her first jobs. Now, she does SAP consulting work, much of it from home (even before COVID-19), at a significant pay bump and highly flexible hours.


The crazy thing is for most of the world people use Windows and SAP and firewalled intranets and no external email. Apple laptops and free access to the internet and SV culture is a tiny minority of what the world's computing looks like.


I’m nominally a DevOps engineer at a really big Germany-based auto parts (among other things) manufacturer, specializing in Azure and Kubernetes.

My real job many days is helping devs make peace with our various web proxies - even the special, relatively open, non-SSL-intercepting “dev proxy” requires CNTLM, which just celebrated its eighth anniversary on its current build. And of course, my little darlings take dependencies on it for things for which the masses absolutely not have access to it.


as a designer working in enterprise side application development - so much this. So many design portfolios are geared towards consumer side apps but the world is powered by the most dry boring stuff.

[I finally got around to making an account after lurking for so long. Your post was the catalyst ]


I'd imagine the smart startup engineer folks will never know o this market because they're either working at some place like Google or doing their own thing, and one would need to have worked at a less sophisticated blue chip company to be exposed to these flawed systems.


Some people also worked in finance. That can pay well, but has plenty of SAP or similar crap to deal with.


When I was starting school, my dad had informed me that SAP was HUGE, and consultants with SAP skills were pulling in 200k, back when even google employees were still waiting to IPO.


I've got a - maybe silly - question: Do you think that software engineers read about the business of technology?

I mean, some of these behemoths (or used-to-be behemoths) of technology, like Oracle and Intel and Microsoft (Ok - everyone knows them) and Dell and Apple (mostly the 2nd one) and places like SAP (which is arguably the biggest database and logistics and supply chain (sans the EDI middle-ware vendors on the planet) and Synnix (which is arguably the largest ODM supplier of white label servers to the largest data centers in the world), and, well, Wistron? BBG? (Submarine - https://www.visualcapitalist.com/submarine-cables/) Satellites? Aerospace? What it takes to wire our world together?

This is not a criticism - it is simply curiousity - as a tech person - you don't read up on (or do you think people do or do not read up on?) the way the world works technologically?

My son graduated with an engineering degree recently and his friends always want me to take them to lunch (free food!), but also so I can tell them how the world is connected (power, telecom, SDN, Satellites (launch - cradle-to-grave), and I'm always a little amazed because I grew up libraries of books and they've got google, but want to be lectured?

I guess I don't understand. If you have a moment, could you edumacate me?


>I grew up libraries of books and they've got google, but want to be lectured?

Your son and his friends know you. Your experience is lived, and real, as real as you sitting next to them. Reading books is abstract, the language can be obscure, archaic, verbose, flowery, dense, etc. Your son and friends know how to read your emotions and communicate effectively with you. An hour of your stories could likely replace dozens to hundreds of hours of reading because you can tailor your stories on fly the in a way no author ever can.


Thank you for that. (I'm not being glib.)

EDIT: I think your learning style and mine are much different. I remember when I had to do a biological proposal to DARPA (I guess I didn't have to), I went to the University of Texas bookstore and got every book they had on molecular biology - and read them over two days at a Starbucks on the deck (beautiful day, but hot), and then wrote the proposal and had the interview. (I'm an EE,CS and Physics guy who knows jack-shit about molecular biology). I never figured I ever had time to talk to anybody as long as I had 30 vertical centimeters of books on the floor that had to be processed by dawn. Maybe it's a generational thing. Maybe I'm just broken.


Whatever you are, you're not broken if you can actually process 30cm of books in a few hours.


I do read about these things, but I can also spend my time with lots of other stimulating/important stuff. Priorities, maybe?


Fair enough - I was just curious if the average modern software developer read books by Andy Grove ("Paranoid"), or Crossing the Chasm, or the biography of Steve Jobs, or kept track of what's up with the top 250 global tech companies on planet Earth every week for 20 years.

Now, that's a little unfair because I read really fast.

But - I guess I was just curious about how a tech person couldn't know about SAP.

NOT as a criticism, btw -- just I don't know how this could be? They are not a secret company. I just (apparently) don't know how young software engineers are wired? (Yeah yeah - Ok boomer, I get it)

But, because people actually do matter to me - I would like to understand better how they are wired...

EDIT: I am also stimulated by my healthcare projects and my energy projects and my stupid stupid project to build a carbon-nanotube elevator to the sky (which may have more funding than practical deliverables!), but I am completely stimulated. I assumed you were too, so it wasn't intended to be an either/or.

EDIT2: If you don't know the answer to the question I was asking, don't worry about it. Nobody else appears to know either. Just because I have a question doesn't mean that there is an answer. Thank you for engaging and have a good night.


I think it’s a bit of a personality type thing. I’m 30, know a fair bit about SAP and ERP systems in general despite never having really used any (small electronics company I work for uses a really crap MRP from a tiny company in New Zealand, which I suspect was probably selected because the operations manager was mates with somebody who implemented it it and/or because it was cheap and our accountant is a penny-pincher).

But for me, I just find my self researching all sorts of things that aren’t even related to my job. From aviation to renewable energy to database systems to carrier network hardware to finance and economics to chemical process engineering, etc. I just find myself almost infinitely curious, and meet others like that occasionally but it doesn’t seem to be how most people are wired. I don’t think it needs to be the norm to be for most jobs, but I find it’s a huge advantage for the kind of higher level product architect/systems engineering roles I’m doing more of.

In some ways it’s detrimental - I find it really difficult to keep motivated to do the deeper-level nitty-gritty engineering tasks because I’m always more interested in everything going on at a higher level in our products, so when involved in hiring I don’t try and select for people like me. I think tech companies need some but also people who are happy to focus on deeper-level work.


Thank you for that.


I would guess - at least for a lot of developers in my cohort - we came into tech at a time when every few years there was some big improvement to software productivity. Staying current with language features, new languages, and new paradigms could be the difference between getting that next job or not. Especially since people younger in their career are often expected (yes, even in tech) to be current with all the latest and greatest. A new graduate with only C experience is going to have a much much smaller pool of potential jobs than one with React and NodeJS. We've gotten trained (and paid) to assume the old isn't effective, and the new must be better. Obviously this isn't always the case, but I think it leads to a culture of obsessing about the new hotness and ignoring everything else.


That makes perfect sense. They are looking for jobs. (Which is not bad!)

"We" were going to build a new world. So I guess we were some kind of idiots. After I sold my first "game" on an Apple 2 in 1984 (they paid me a lot), I went to work a semiconductor manufacture, and AT&T and Apple and General Magic and Motorola were going to re-invent the world. Good engineering work - bad vision. Didn't work.

Meanwhile in Japan, where I worked, I translated Nintendo Games (In addition to the adobe and lucas arts stuff, which was more for Sega) to the 6502 (hey, I'm old), and made money trying to make the meet the new.

I think it was easier back then. I think there was less intergenerational hostility.

Why do we have inter-generational hostility? I know how to bring the old arcades alive! :-) (Actually I know how to activate satellites - if we want low-cost back-haul - packets here - packets there - There are 0.1% utilized GPRS networks out there.

AND Now we are getting into topics that you don't want to hear about.

I do really appreciate your willingness to share tonight. You broke open a couple of pre-conceptions. And I guess that's the reason you talk to people.

Thank you.


This is something a lot of us lack. Do you know of any resources which can help in learning about the business side of the technology?


Start here: https://www.andykessler.com/andy_kessler/excerpts/How_We_Got...

Message me after. I am at wmarchibald@gmail.com

EDIT: Oh! Also "Eat People" by same author - Kessler : https://www.amazon.com/dp/B0049U4KC6/

Semi-edited: Everybody else will say Zero to whatever, and the Hard things about Hard Things, and Jimmie's book Good to Great and Crossing the Chasm, and you might call me mad (and I may be!), I am the only person who will recommend a book called "Eat People". (It is a good book, though)

EDIT2: U - akjha - Please let me know what you thought. You have my email address.


Please let me know how those two resources treated you and we can go from there. (Part of it will require that I know where you want you career to go)


Hey! I'm David, the editor of the post you linked to. If there are any other topics you're interested in learning about, please let me know! We wrote another one on Salesforce (https://retool.com/blog/salesforce-for-engineers/) that was well received on HN too (https://news.ycombinator.com/item?id=20277115). We were thinking about writing about Marketo / marketing automation solutions next. (I'm david AT retool.com)

(As an aside — when we started Retool, we ourselves didn't expect the "custom internal tools" to be such a big market. As it turns out, we now have customers that spend $400M+ / year on building custom, homegrown, internal tools! While we [or any single vendor] certainly can't capture all of it, our goal is to capture some percent of it.)


I worked at NetSuite for a little over 4.5 years (2 years before the Oracle acquisition), and it also blew my mind how much companies were spending on NetSuite customizations. It's a huge world out there for sure.


Great article David! Thanks for taking the time to include some screenshots and videos. This is the kind of company-funded-blog-content that we all love to see.


Where is most of your adoption coming from? Which verticals? Which horizontal groups?


Did you ever look at pega, nintex and Appian?


You really have to be living in an alternate universe if you haven't heard about SAP. To me it feels like every other job involves some work with it.


SAP was just nominated (together with Deutsche Telekom) the contractor for the official German Vivid tracking app (Covid-warn-App), just saying


I really don't get why they did not award an agency with a proven track record of delivering apps. If you choose one of the big ones, you still have high rates, but at least there is a chance of actually getting it done.

SAP is known for incredibly long ongoing projects and Deutsche Telekom has a track record of destroying acquired online businesses. I would bet 10:1, that this project will fail.


Probably because SAP is a huge political force in Germany more than anything.


I work in B2B software and almost all of our clients use SAP. It's absolutely massive.


My company is using SAP so it is very familiar to me. But before I joined my company, I have never heard of SAP. There are some good industry process knowledge that gives them an advantage and the software is good once you get used to it. But it can definitely be 10X better (there is a big opportunity to make it easier to customize or integrate with other software) . However, once you become their customer, the switching cost is very high even though there is better software in the market. That makes it difficult for startup to break into this market.


most people would be similarly surprised that $1T-market-cap microsoft is a business & enterprise software company, rather than a consumer one. it's ~80-90% of their $130B in yearly revenue (hard to get more precise because of the way they report their numbers).


That's not surprising to me personally, but I can see how it might be to the average person


yes, but a decent chunk of it competes directly against SAP.


> ..78% of the world’s food all flow through SAP?

Having worked in food industry as IT about 20y ago, we used to have fragmented system portfolio, and were struggling to pull data from one system to another. The around 2002 management decided to switch to SAP. I remember working on the storage/capacity/CPU aspect of it (and mid-migration I left the company and completely missed on the SAP train in my career). That food company was providing milk, yogurt, cheese, juice, ice-cream to about 25mil people (customer base). That's SAP for you right there. So that 78% doesn't surprise me at all. I believe the remaining 23% is too small to invest that amount of $$$ nad just hasn't done it yet.

When you get the ability to "track every cucumber" from pick to sell, cut down on fraud, errors, be able to calculate every bit of profit, loss, why would you go anywhere else? I understand the initial cost may be hefty, but once you set it up (good and well), you can stop focusing on that, and see how to grow the business parts of your business, and not the admin.


See this sounds good, and is why many execs are sold on it.

But because it is so massive, it's impossible to find talent that can actually implement it properly. That's IF your company doesn't require a ridiculous amount of customization.

Projects always run over-budget and over-time. Vendor lock-in becomes a huge problem. These companies solely exist to keep customers using their software and paying their service fees. They then provide just enough marketing buzzwords to justify it to execs who justify it to other execs.

If SAP is done right, it is very powerful, but it is seldom done right.


> Oracle and SAP are so dominant that even Microsoft uses SAP instead of their own ERP offering, Microsoft Dynamics.

This is insane. If it's hard for a company like Microsoft to move away from SAP and shift to their own solution (hence saving face in addition to the $$$), I can't imagine anyone switches software in the ERP world


Not so mad, MS Dynamics GP and Nav don't play in the same league as SAP/Oracle. AX is closer, but not there.


I work for SAP and it's surprising to me that not even other engineers in the Bay Area have heard of the company despite its campus in Palo Alto and their sponsorship of the "SAP Center" in San Jose. Do you file expenses? You probably use SAP Concur. Do you order office materials or equipment? You may use SAP Ariba. Have you taken part in interviewing a candidate coworker, or have you done yearly career goal planning? You may have used SuccessFactors for that. And this doesn't even count SAP proper's offerings (HANA/S4, etc).


Concur, Ariba and SuccessFactors are acquisitions, and only SuccessFactors is fully integrated under the SAP Umbrella. Here in Prague Concur and Ariba have their standalone subsidiary companies, with Separate HQs in different parts of town.


In my opinion the civilization in question does not have steam-powered airships unfortunately, but it does have some really overpriced crap.


Believe it or not, there was business software in the 90s. They were called *RP. ERP = Enterprise Resource Planning, MRP = Material Resource Planning, CRM = Customer Relationship Management. Home-grown code hacks to run a business cannot do what these companies can do, that’s why they’re expensive.


It seems that the market is big enough to allow unicorns making "plugins" like https://en.m.wikipedia.org/wiki/Robotic_process_automation to emerge.


Thanks for sharing, very interesting to read.

> basic installation of SAP has 20,000 database tables, 3,000 of which are configuration tables. In those tables, there are ~8,000 configuration decisions you need before even getting started

Wow, this is nuts. 20,000 database tables?


They've been adding tables since 1980...


I did an internship at the SAP HQ in Germany more than a decade back.

The work was exceedingly boring. But they had excellent lunch. Better than most Google offices.


You work in IT and you've never heard of SAP?

That's pretty wild, like not knowing who Microsoft or IBM are.


I haven’t read of such things in my country but am hopeful that this will happen soon here in India.


TfL, BP, Shell etc run on SAP. Had good fun messing around with SAP over there


ExxonMobil runs on it as well.


I clicked on the link not having heard of ERP or SAP and expecting to have my mind blown, too.

But, isn't that software almost the reason why computers and databases and programming happened? To help businesses with there boring business stuff, like inventory and orders and accounting?


This is kind of how I felt when I started to learn Haskell


It's like Salesforce, that crap is ubiquitous in large companies. Have you ever used it? It's a nightmarishly clunky piece of trash and yet Salesforce is a 150B behemoth.


Why would you use ERP instead of a regular SQL database?

edit Got a lot of vote downs. This was a genuine question from someone completely ignorant on what ERP is. Not trying to say that SQL is better, but I'm asking this because I'm ignorant of what ERP is and the description makes it appear as if it's a database.

What's the difference? would be better wording I guess.


Try getting your average office worker to deal with an SQL database directly. No chance unless it only has a single table. A database without an interface tailored for a specific purpose is useless for office staff. I know this from experience having once tried to get office staff using pgAdmin to update the shiny new database I had designed for them.


What's funny is that this is what SQL was originally designed for. That's why it reads so much like English. Of course, CLIs were too.


Because it's your business logic. And your GUI builder. And your tax system. And your legal report generator. Etc.


Yes I've heard that often businesses mould themselves to the software, rather than the other way around. The argument is that decades of exposure means that, if not entirely optimal, it at least represents competent processes for running a very large and complicated business. Rolling your own multinational tax compliance in SQL would be madness.


Why would you use an ecommerce platform instead of a regular SQL database?


When you're doing eCommerce.


My point exactly.


I don't know about SAP in particular, but ERP solutions frequently use SQL on the backend. I work with Microsoft Dynamics and it's just a SQL database with a special front-end for constructing workflows. I suspect Oracle ERP would be similar and have Oracle on the backend as well.


Because the ERP solution comes with a gaggle of sales people, sales eng, and bizdev folks to convince your co's decision makers that it is best thing since butter.

SQL doesn't.


"SQL doesn't"... you are correct, SQL can't figure out how much tax you need to pay across 9 tax jurisdictions. SQL can't tell me the best way for my warehouse picking robots to move across the floor based on the layout of the warehouse and what others products are being picked right this second plus over the next few minutes. The list is really endless of what SAP will do and SQL won't. It would take decades to replicate what SAP does.


I mean I guess you can throw all that functionality in a bunch of triggers but that would slow down your database to a crawl every time one record updated.


Have the triggers publish to service broker!


[flagged]


[flagged]


I used to implement ERP systems. I don't anymore.

SAP released new IFRS tax compliance features last year to match the latest government regulations; they then gave this is over 4000 customers for free. It cost SAP around 1.2 million EURO to build, test and roll out. Are you saying that those 4000 customers (companies) should have just built their own tax compliance finance systems? That would have cost 4000 * 1.2 million which would be a total economic spend of four billion eight hundred million... how about you put your ego on the shelf and listen to other people?

You're advocating building massive software applications over using what's already available and you're advocating that all companies do this... you completely ignored my argument and you have no idea what you're talking about. Just because you write crappy webapps in JS that doesn't mean you understand the software market at all.


Sounds like SAP does custom App building. Do they use a custom underlying database technology as well?

Would it be equivalent to contracting some software system to a a company that would build the system using SQL?


The metal economy is huge. Almost every product requires metal. But metallurgists don't take home a big fraction of it.

What's special about the software economy is that people who write the software take home a large fraction of it. Maybe more than half, depending how you count. Which is unprecedented.

It's not just because software can be zero marginal cost. Musicians don't take home half of music revenue. In most creative industries, the creatives take home less than 1/10th.

Enjoy it while it lasts.


Good software is a force multiplier than enables everything else to be more effective.

A traditional mindset applied to software results in the businesses we often see people complaining about. "We make the product but Sales gets all the money!" and that's very similar to your metallurgists. Code for hire as a product of labour. It takes an army of people to sell, to advertise, to make me want what it is you have (just like with music).

But a business who looks at software as complementary will see the value. A dollar spent brings many dollars back, just indirectly. It enhances the machinery that's already in place, and allows it to do entirely new things.


" people who write the software take home a large fraction of it. Maybe more than half, "

Where do you get that figure? The vast majority of people writing software are not paid enough salary + benefits to come anywhere close to 50% of the revenue their work is responsible for earning.


It's a good point. It varies widely depending on the industry, but even in the software industry a good rule of thumb is that IT costs about 10% of revenue (some places are as high as 30%, some as low as 2%). But that includes all the management (not executive management) and support staff. The spend for sales and marketing is much higher than development (though sometimes most of that goes out of the company, paying for sales channels, etc).


> IT costs about 10% of revenue (some places are as high as 30%, some as low as 2%).

That’s the total expense, right? Not just employee remuneration.


Yes. However, it may or may not include equity depending on how things are set up.


Do you mean employee equity? Isn't that usually considered part of total remuneration?


It’s not a cost as a function of revenue the same way cash, perks, and other benefit are.

It’s also not clear that typical grants are enough to close the gap to “more than half” as the op remarked.


In companies that actually manage to get to revenue, may be. But if you take together all the software companies, including many that have not yet and may never get to revenue at all, I think it would be much closer.


How so? The vast majority of software workers aren’t at VC or self-funded startups, nor do they work for small boutiques developing web/mobile apps or managing eCommerce deployments for local mom-n-pops. They’re employed at larger companies and defense contractors where the overwhelming majority of revenue goes to pay all workers (of which they’re a small fraction), expenses such as leases, or profit for executives.


Musicians write music. Developers write and maintain software.

Imagine there's no Version, it's easy if you try...

Really, no concept of version: you buy a piece of software as it is, take it or leave it! Either it works enough for you or it doesn't. No bugfix in the next release because there is NO next release. The musician, ehm... developer, has switched to the new shiny thing. You modified yoour environment slightly and you need a tiny change on the software to adapt? Too bad it does not exist!

Really, the point I'm trying to make is that comparing software developers to musicians doesn't hold a lot of water. Both professions include creativity, but a song shipped is as immutable as a piece of hardware. Adaptabillity is what makes software as worth as it is!


One could argue the comparison still holds for developers of video games - as long as the game's code is not touched upon release. Used to be the norm, nowadays admittedly seems to be disappearing.


Also the amount of skilled professionals needed to write a song and to write a game is slightly put - different.

A song surely requires a skilled professional. But a game required dozens of those, with years of experience and lots fo time.

Creative industries get way lower pay because the skill level needed is low compared to many others, like law, medicine etc.

If you are talented you can paint great pictures or be a great guitar player in 1-2 years.

To be a good doctor you need what? like 10? 20 years?


As both a professional programmer and a serious musician I have to _strongly_ disagree with this statement and implore you to consider your biases. I can only assume that you're making this assertion because you either a) don't play an instrument very well, or b) don't play an instrument at all.

First of all, even though "all in one" songwriter/producers exist today, you don't see them in the spotlight until they've already worked on their craft for many, many years. But in most cases, the commercial songs we listen to involve AT MINIMUM 1+ songwriters, 1+ producers, vocalist(s), professional instrumentalists (if it's not a band and/or a song where everything is programmed), an audio engineer, and probably a mastering engineer. And that's on top of the scaffolding of a bunch of other peoples' work (beatmakers, sample packs, recording studios/labels, etc etc). Each one of these people may not be formally educated, but you can bet they've spent years with trial and error honing their craft.

> If you are talented you can paint great pictures or be a great guitar player in 1-2 years.

What makes you think that? And if we're going to have such a low bar for "great", can't we say the same thing with programming? I mean, I know people who go to boot camp for 1 year, work for another, and then call themselves intermediate programmers and get paid $100k+. Same goes for law. I'm sure some real smart and performative people could become great lawyers within a year. The only reason that would never happen is because you literally have to go to law school for X years as mandated by law.

Don't mean to have such an aggressive reaction here, but I am really tired of the crafts being devalued so offhandedly.


Like games and software, the staff needed to make a song varies a lot.

If you want your [game, software, song, ...] to reach a certain level of polish and fame, you need a similar number of people working on it. Mixing, mastering, recording, composing. You need someone who's very good with sheet music and knowledgeable about all parts of an orchestra if you have any orchestral parts you want played on real instruments.

And so on.

Can I make a song in a week that'll impress a few tens of thousands people? Sure. People write in to me all the time saying "this is just what I needed!" They don't notice I banged it out in Ableton in an hour because it's what they needed. It's what I needed too, so it was easy to make it good enough. Just like you can make a game or a piece of software in a week that'll scratch a lot of people's itches even if it's not perfect.

Recently, I got back into photography, and it's the same. I can take a great picture on my own and impress a lot of people, but most famous photographers have a staff that rivals the biggest non-unicorn SV startups. Editors, lighters, second (or third) shooters, general A/V crew, marketing, office staff, etc.


Sorry, you unknown unknowns are leaking.


> Developers write and maintain software.

Which explains why they are so in love with evergreen software that constantly changes things largely for the sake of changing them I suppose.


Yes. The commodification of professional work is worth pointing out. I'm surprised by the percentage of people in this thread who think the SW industry is exceptional. It tells me a lot about the fictions people want to tell themselves in an industry that is still relatively novel.

By contrast I'm a HW builder (mechanical engineer / industrial designer / design manager). I make good comp but maybe 70% of what my similarly leveled colleagues in SW do. The vast majority of the arguments in here still apply to my work, with a little less focus on maintenance. And yet my job role's comp is in a slow decline. Why? Because it's fully commoditized.

Anecdote: I used to work for a HW focused VC leading their internal incubator/consultancy's physical design team. One of our ML focused investments got acquired for a ton of money and every other startup in the building wanted to know how well compensated the team was going to be. It's odd to see some people at the water cooler suddenly be worth millions when your startup is still grinding, it makes for stark contrast and a very real "what if" to the onlookers. One of the Sr MEs at a struggling startup was clearly jealous and lamented to me, "man, screw machine learning, how come nobody is paying me a million bucks for my work?" To which I said "let me ask you something, when you were studying mech-e in school, did you ever use steam tables?" yeah. "Did you learn about the Carnot cycle?" yeah. "Have you considered the fact that our industry is 200 years old? It stands to reason that our jobs are not hot shit anymore. Be glad you like your work and are still paid well; ignore the new kids if you can."


Metal economy is huge, but there are inputs. Software engineers have no inputs into our projects besides our work and some hardware and electricity. A developer is 99% of the cost of a piece of software.


In other words, your value-added as a software engineer is significantly higher, as a fraction of the money in a transaction, than the metalworker's.

This is because of leverage. Implement software that speeds up a process by 20% once, and every time that process is used from then on the speedup adds value.

We're lucky to be working in a field that enables us to add a lot of value, for relatively little capital costs, and pocket a good share of the difference.


They listed music as a counter-example to this. No inputs = zero marginal cost.


You've missed coffee!


What would cause it to not last outside of precedent? In all those cases, it sounds like middlemen exist that corner the end-market that you gotta go through to make big money. How would that even work with software in the age of the internet, open source, and commodity cloud infra?


I don't know. Maybe this time is different.

I have this book from 1880 about steam locomotives, where the preface talks about how all the most ambitious people are going into steam engineering. Which they were, for a while.

The answer might be, until something better comes along.


AI-driven software development down the road could theoretically commoditize developer skills.

I sometimes wonder how many decisions I make per day that an AI could replicate in some repeatable way. Who knows, but nothing lasts forever.


We should technically be the last to be automated. We’ll destroy other labor systematically over decades, and then finally ourselves.

We seem to be subconsciously precocious about our fate too, as we deprecate things at a faster and faster rate.


I think you might be surprised, it’s possible some types of programming could be automated away. Honestly I would consider writing programs more straightforward to automate than a lot of systems that need to interact with the physics map world, or with people.


But programming is essentially about interacting with people more so than interacting with computers, even now. It's less so that the job will go away than that its nature will change, such that the proportion of time interacting with people goes up, and the time currently spent "coding" will instead be spent interacting with AI which writes the actual code. But this AI won't be smart enough for a while to talk to a normal human who doesn't know how to talk to it and translate requirements into the right form. Hence the need for the programmer. Programmers will make use of the automation to just become much more productive.


> We should technically be the last to be automated.

I disagree, developer work will be automated as soon as it is technologically feasible, especially given the high cost of software development labor. I foresee AI software development platforms starting to take over new application development at some point, building applications that humans have minimal access to the source of. Existing apps by humans will likely need human maintenance until AI is able to reason about larger legacy applications.

We’re already starting to see no-code apps taking off, and AI will drive the proliferation of that technology as it becomes more sophisticated.


I have doubts about AI driven end results, but I could totally see AI created domain specific languages


Maybe that analogy would be accurate if you could make new and pleasing music out of hundreds of small samples of other music, and also 90% of the musicians were used to either getting their samples via torrent or picking one album and playing it in an orchestra of 5,000 for the next 40 years


That's literally how most of music production works. You torrent sample libraries, get used to a couple of genres that you do well, and then use the same tools all over again for every client.


> What's special about the software economy is that people who write the software take home a large fraction of it

I don't think this is true. Take all the non-tech companies that rely on software and hire tons of engineers, such as insurance and banking.


My hypothesis for why tech has high salaries is stock compensation and fast growth. For example, Google stock is nearly 2x what it was 3-4 years ago. If you have a four year grant and had about half your comp in stock, then 2xing the stock means you're making 50% more now, assuming you didn't get any promotions or anything. If you're trying to hire folks away from these situations, your pre-stock-growth starting offer has to compete with their post-stock-growth comp, so starting offers have pressure to grow over time.


Curious what fields return what fractions of benefits / value / profits to what participants, and how you'd measure and classify those.

There's SIC / NAICS / ISIC codes, or tier classifications, ranging from three (Qesnay) to five (Clark (1940), Hatt and Foote (1953), and Bell (1973)):

[A]n economy's major sectors, as delineated by Clark (1940), Hatt and Foote (1953), and Bell (1973), correspond to major stages in the essential life process. The primary sector -- agriculture, fishing, lumber, mining, oil and gas -- represents the extraction of matter from the environment to produce energy, including the calories to sustain individual organisms. The secondary sector -- processing primary goods, as in construction and manufacturing -- represents the synthesis of matter and energy into more organized forms (negentropy). The tertiary sector, including transportation and utilities, represents the infrastructure for distributing matter and energy about the system, while the quaternary sector -- trade, finance, insurance, and real estate -- constitutes a parallel infrastructure for the collection, processing, and distribution of information that is necessary in all living system for the control of material flows. Finally, the "highest" of all sectors in its remove from the physical environment is the quinary sector, including government, law, and education, representing the societal programming -- socialization, education, law making -- and collective or representative decision making to effect control.

-- James Beniger, *The Control Revolution,8 p. 179.

For participants, perhaps: unskilled labour, skilled labour, professionals, managers, finance, consumers?


And just because in other industries people are used to taking smaller portions of revenue they generated, doesn't mean that we should do same.

Also unlike in some creative industries, software engineers tend to have higher education and are better able to understand their economic contributions.


First of all, software is not a creative industry. Creatives always took home a small fraction for many obvious reasons.

Software solves hard problems it is constantly evolving and updating so you need lots of personnel to maintain it, with a lot of knowledge so it will never be cheap, and great engineers to create it in the first place.

Take away a bunch of artists nothing will change. Take away a bunch of critical software - the world will stop.

So no, it will last and it will last long, as the world becomes more and more technological.

It is unprecedented because software and computer world in general is a huge breakthrough and is only 30 years old. Music on the other hand is thousands of years.


> Take away a bunch of artists nothing will change. . Take away a bunch of critical software - the world will stop.

Is that right?

Is there any evidence that a society can exist without art?

There’s an overwhelming body of evidence that suggests it can do so without software.


Not agreeing with GP, but I think your comparison here is a bit flawed.

Art is a broad concept - it doesn't suggest any specific form or delivery. Software is fairly concrete - it has to be run on a computer.

A more appropriate comparison would be art versus scientific research, or art versus engineering. I think that's a more interesting question, to which I would say: perhaps we could survive without either, but I can't imagine we'd thrive without both.


> Musicians don't take home half of music revenue. In most creative industries, the creatives take home less than 1/10th.

Perhaps the difference is that musicians don't change a completed work. There is no maintenance cost. A song doesn't become useless and obsolete if you don't pay the musician to continually improve it.

For static software, I'd expect it to look more similar. How much of Tetris went to the people who wrote the software?


Most of the people writing software for the companies listed in this article are freelance workers in lower wage companies. They make close to, and sometimes less than minimum wage for high wage nations.

Startups pay well because speed is so important and faang pays well because their margins are enormous and the risk of those shrinking is worth paying top dollar for the absolute best talent in the world.


Having read patio11/kalzumeus a lot (excelent blog posts and talks!), I'd say if he has a law named after him, it should be "charge more" or some fancier title like "you are not charging enough".


Lol, I agree that should be the real patio11 law


Or maybe a rule or such for general peeps that doesn't necessarily involve some special coffee-drinking in trendy Oregon? And what the heck is a "bed-and-breakfast type coffee tray"?

Or, Patio's Law, rephrsased: "Look at the very significant things you encounter with fresh eyes. You already know how to get to the crux of issues. When you get there, FRESH EYES!".


> And what the heck is a "bed-and-breakfast type coffee tray"?

A tray loaded with tea and coffee making equipment and supplies, like you get in hotel rooms. Not a tray for carrying things around.


A friend has described these to me as "dark matter companies". They are everywhere, and they permeate industries, but you never see them. They work in small, unmarked offices outside of the standard tech hubs. Some of these companies make billions, but you've never heard of them, and they'd like to keep it that way.

In other cases, they're small companies that own a market that's too small for a large competitor to go after. $10m/yr will not make VCs happy. But it will certainly pay nicely to a team of 12.


It reminds me of Hidden Champions.

>Hidden champions are relatively small but highly successful companies that are concealed behind a curtain of inconspicuousness, invisibility, and sometimes secrecy.

https://en.m.wikipedia.org/wiki/Hidden_champions


I’ve had four jobs in my life and have always been stunned by what is still done by hand. Every company had a task which could be heavily automated but was instead consuming 4-30 people.

So if you think the software market is large, consider all the stuff software never reached.


We have an "internal systems" team at my current company to solve this problem. The product manager for the team knows everyone and keeps tabs on what's being done that could be automated. It's not used to replace people (at least it hasn't been yet) rather it's used to allow the team to stay the same size and scale up their operations to meeting the growing business demands.

I think many companies could benefit from a team like this.


One of my four jobs was on such a team. Unfortunately, they canned it (after I left) as they couldn't figure out what it did.

If my team of 8 devs had just one dedicated to improving infrastructure and processes, we could probably increase our overall output by 30% in a year.


Ha, I'm happy this exists. I created this department at the company I founded at and directed it for some time. We call it the internal tools department. It still exists today and is run by a former employee of the department. It's very valuable for us.


This is great. How do they deal with the on-going maintenance of the tools they build? Do they move to the main product team or does the internal systems team continue to maintain them?


I wrote a piece of code last year in probably no more than a few hours.Showed it to a few guys in my team.Instead of sharing my excitement,they were like "if you run this,what are we going to do then?". She was right,if I'd do it, the size of the team would be reduced by half.But then,what I'd get? Do I get a bonus the size of those employee's salary for 6 month? No.I'd get a grand or two and instead and a pat on a shoulder,while those guys won't find a job for momths.So what's the point?


Within the company, there is often no point. You want to be the guy selling it outside the company. You can easily claim tens of thousands for the same work then.


Management has to be ok with reducing headcount. You would think that all managers would be incentivized to made their groups more efficient but there is a high correlation between headcount and political power within an organization and people don't like to give that up for free.


You're in a delicate situation, since you'd be to blame for your co-workers potentially loosing their jobs. Nevertheless, doing nothing is also risky, since someone else from another department could show up and automate away your jobs leaving you in a hard place, and you won't have any control over that.

If you propose the automation yourself, and plan it's adoption within your team, you may have a chance of taking advantage of it, and even try to bring in more responsibilities within the company to your team to secure your teammates jobs.


The point is getting that thousand bucks before someone else does. Or to make the company efficient enough to survive in a market where the other company's already efficient.


helping your company save costs (at scale) is almost always stupid compared to selling that knowledge to other companies.


Isn't really anything your team could do? Usually there are lot of fronts people could work on and bring more revenue to the company. Try to find those to redirect your team and implement your solution.


I've seen this even at technology companies. It's like our devs hours are valuable so we can't invest them in something everyone is comfortable doing with a spreadsheet already.


Internal access to developers is absurdly structured a lot of the time.

One company a friend worked for billed internal developers at twice what IBM would charge for projects.

At other companies, dev time is allocated without any kind of cost analysis whatsoever. It reminded our Romanian business analyst of the Soviet Union.


This happens to me a lot, but it's basically a kind of technical debt. Most of the time the upfront time investment to automate something properly versus doing it manually is hard to justify. Especially if it's a task that only I understand and now I need to document the problem and the automation instead of just the problem.

Business logic doesn't make for elegant APIs that I can hide behind a script and throw on a server, even if it's repetitive or tedious.


There's also the risk of allocating a team of folks with spare cycles, having the get half way through the work, then getting pulled into billable work and leaving it unfinsihed.


We are always looking for ideas!

Can you give a few examples, please?


Sure

A tool where you are given a picture/scan of a table and it spits out a table that can be pasted into Excel. This was at a bank where numerous people spent half their days manually typing in scanned documents into the same table in Excel. You would double a $70,000 a year financial analyst's productivity with this.

A portal to manage employee and employer documents. I.e where both you and your employer can easily access your T$ (Canadian tax form), offer letter, certifications, police checks, etc.

Software to manage which options a client ordered and what their selection was of that option as well as autogenerate the appropriate SQL queries from a client config template. The current process consisted of a mega word document with people fiddling around manually with it.

Happy to provide more.


That first one can be done with teseract ocr pretty cheaply. I think a lot of people just don't know how much software can do. You just need to standardize the form and then there's a human data validation step in there.


Several years ago I came across a pile of interesting-to-me US government data that consisted of scanned PDFs of paper forms (some filled out by hand, some with a typewriter, more recently with a computer). The paper forms change in layout and level of detail a few times over several decades, there are hundreds of sub-schema variations within each time period (yes, that means that a particular variant might have a single record from 1983, a dozen scattered over time, or thousands across several decades), some variations are due to a simple category misspelling, others are more substantial.

Some of the scanned forms have marginal notes (probably okay to ignore these), but I've noticed some cases where the form had prefilled data crossed out with handwritten data replacing it.

Please let me know if a system exists that can deal with this idiosyncratic mess and coaxed to produce structured data of some sort without basically using MTurk behind the scenes. I would be interested in re-releasing the data into the public domain.


Can you not use software for the 80% that are amenable to it and humans for the other 20%?


The challenge we faced (for the two days we tried to hack together a solution) wasn’t so much the OCR but the fact that we couldn’t standardize the form as we were onboarding people with the documents from other banks.

Our requirement was that the the software would need to read the number of separating lines in the table and output accordingly.

Maybe it can do this, idk. I was an intern at the time.

We were using tesseract for our attempt too.


I worked somewhere that did something similar. Their product would scan and extract data from documents. Even working with the banks and companies who produced the documents, there were very stringent requirements (like identifying marks on pages to tell what document type we are looking at, pages being fed in the correct order, etc). Plus the ability to handle degraded, damaged, or bad copies. It's apparently one of those things that seems easy at a glance but the edge cases end up employing a whole department (they had people to review the scans and press OK or correct the data). It was a legacy system and I think a newer rewrite could get further but I'm wary about saying it could be fully automated.


90% automated is sometimes 90% as good as fully automated. It's not always the case, there are cases where hitting that 100% enables you to do things that were just not possible at all before, but sometimes it is. Recognizing those cases where a 90% solution is a 90% improvement at a 10% cost is a valuable skill.


Yes it's also part of the reason I prefer coding for myself. I can ignore a lot of "what ifs" and edge cases because I'm always there to pick up the mess if something happens. I get to focus on the stuff that makes an impact and forget the idiot-proofing :)

Or said another way, it can be fun and also effective to "Do Things That Don't Scale" [http://paulgraham.com/ds.html].


There are commercial OCR software packages that can do it. FineReader is one example (and it's been around for ages).


FineReader seems like it would have done the job.


We used tesseract ocr, plus other image processing libraries to improve ocr's chances, and although we were close to usable results there were still many challenges. One major problem was that different data sources had completely different page layouts, and often times the original documents were out of reach and the client only had available the scans of fax transmissions!

The bottom line was that we were able to achieve very good results consistently, but the margin of error was not acceptable for our specific use case.


Not OP but I had a friend who has interned at a few record labels. He's not really a programmer, but knows enough about programming to know what kind of task should be automated. He described a portion of his job to me that he and a bunch of other entry-level workers spent a few hours a day doing (basically scraping and storing data off the internet to be sent to business people for analysis).

I worked with him to develop a Python script do basically automate half his job in a few hours that could have saved the company tens of thousands of dollars. The company didn't want to dedicate/hire an actual programmer to maintain it, so they kept paying employees to manually scrape the web.

The crazy part is the next summer he interned at a different company and literally the exact same thing happened. He re-made the original script, saved tons of work, and the new company still didn't end up using it after he left.


Are these companies irrational? Perhaps, but there might also be an important point.

A manual process is robust, fault tolerant, and easy to adjust to new circumstances or changing business requirements.

A Python script is fixed, and they day it breaks down due to some slight change in scraped data, you have to pull in a developer to fix it. For a company without dedicated staff to support such automation, it is a high risk.

Ad-hoc automation can turn workflows into Rube Goldberg machines. A Python script here, an Excel macro there - soon nobody knows whats going on and nobody dares change anything.

Developers often think only one step ahead: It was manual, now it is automated with a script. Win! A business have to consider that larger picture - future support, maintainability, risk etc.


Exactly - intern labor is cheap, flexible, and there's a "wax-on wax-off" mentality where they may learn by proxy but are also in a safe environment where they won't hurt themselves or the company. A script and the associated automation could become the expectation or could lead to a key-man-dependency-risk.

It may be lucrative for the friend to take the script and turn it into a product. but then they'll need to market, sell, document, optimize, and do a million other tasks.

There has always been a dead zone between "automated script that makes me work a bit more efficiently" and "product that can make everyone with this job work a bit more efficiently". Fortunately, new technology is narrowing this gap.


Oh, many companies have only themselves to blame for being blown away by technology.


Does he get paid at the internships?


Yep, but not as well as tech internships. Still, the company was throwing money down the drain because they didn't know that even a semi-competent programmer could automate those jobs.


Bulk downloading and resizing photos in weird spreadsheets.

Managing attributes and taxonomies for ecommerce.

Capitalizable hour analysis for software projects.

That's just stuff from the last few weeks. None of these are hard to automate but the workflow integration is the tricky bit.


Do you know why they don't just use an existing API that does image resizing automatically? Presume it's the integration challenge, but I'm curious if you know /why/.

Also, what does "Capitalize hour analysis" mean?


I doubt they know such an API even exists (or what an API is!). Most companies aren't run by hackers, assuming there's an API that does the thing they want to do isn't the natural response.


This does not just apply to software..., tons of small and medium sized businesses and opportunities out there. For engineers you will most likely never notice them if you're caught up in the Silicon Valley / VC thing.

Of course it requires getting out of your comfort zone... a lot of dev's just want to exchange labor for cash and grind out code because it is comfortable and cozy (right now).

There are tons of small towns and cities with their own economies. And a lot of these places it means a lot if the people purchasing your product or service can get on a phone and talk with someone 20 or 30 minutes away max.


It wasn’t until I got a job in consulting and started traveling to what are often dismissed as the “fly-over states” that I got to see billion-dollar B2B businesses that were basically unheard of outside of their market. The B2B e-commerce market alone is over a trillion dollars.

There’s a lot of opportunities for ERP developers and other tech roles that don’t exist in SV.


I spoke to a guy recently who works for a company doing $300 million revenue/year as a wholesale distributor of internet networking equipment (think finger optic cables and equipment for housing them underground). Their customers are the ISPs.

The company is privately owned and not venture backed so you'll never hear of them unless you are in that business.


> The company is privately owned and not venture backed so you'll never hear of them unless you are in that business.

Is the implication that venture backing needs growth so they get heavily involved in the media circus? I suppose that makes sense. I've always liked these "shadow industries", I've run into a few personally looking for things like special cut metal or wholesale pie crust or curtain rails.


Venture capitalists are often looking for billion dollar companies. If you found a company and grow it organically to where it's worth $25 million, then sure, you are no Bezos, but you are fairly successful (financially) by any standard. But venture capital doesn't care about $25 million companies, because their business model involves investing in so many companies that fail that a $25m success doesn't really move the needle for them. They want to see B's not M's in their companies' valuations.

Meanwhile, the media doesn't have much to work with when reporting on privately held, non-venture backed companies. There's no official valuation unless the company sells (sure, once you get to be the size of Cargill or Bloomberg, reporters will try to calculate some sort of valuation, but very few companies fall in this category), and the owners typically have no good reason to want publicity from the media, especially if they are selling B2B.


the types of opportunities that lead to SaaS businesses, are extroadinarily hard to find. I've gone to dozens of dozens of networking events and talked to dozens of people from a variety of industries. I've done informational interviews, etc. It's very difficult to find problems that don't have good enough solutions. Most of the time people are pretty happy to continue dealign with the problem the way they have been.


But that's why when you _do_ find a solution, it's worth money.

But i agree that it's fundamentally a people business - you're trying to convince a business owner to change their way of working for the better. You will need to present evidence that it is _indeed_ better - that's even harder than it sounds. And in most cases, SaaS providers are very biased on the value provided vs charged. I tend to think business owners as conservative, and therefore, the new SaaS product will have to provide extraordinary value to be considered.


That might be just the entry point for me, since I'm having a hard time at landing my 1st job as junior. I feel like I'll have a good time and lots of fun anyways. Any advice on how to approach clients? (I can do (full-stack) web development and solve problems/ optimize solutions with algorithms and meta-heuristics.)


If you can't land a job as a junior programmer, you don't want to go around to businesses and end up responsible for promises you can't keep.


What a lot of these businesses need is very simple stuff. Plenty of small businesses would have use for what would consider “toy” programs.

A friend of mine built a text message to email system for this rural delivery guy for the purpose of letting his office assistant manage invoices, but also allowing him to fulfill orders on the go without talking on the phone as his customers can just text him. Charges him $500 a month for that.

That kind of thing is trivial to do, but can still have a lot of value. Remember, the small businesses don’t automate anything. Excel is frequently beyond them.


> a text message to email system ... letting his office assistant manage invoices ... fulfill orders on the go without talking on the phone as his customers can just text him

Oh man... that's what I'm after, problems like that. Would be so neat even if it's small to see it get used(users) and the analytics/small income.


I don't think that is often trivial and it wouldn't surprise me to see junior programmers struggle with things like that, but maybe the average junior programmer is better than I expect.


V1 skimped on many of the things we might consider requirements such as preventing SQL injection. For a long time the username and password for this service consisted of hard coded if statements for each client. Plenty of plaintext passwords. Default admin logins. Even seen a frontend password checker with jQuery.

So it isn't software we would consider production acceptable. The code quality is atrocious. None of these people I know who do this would ever get hired as a software engineer by any company that does a coding interview.

But you pair their can-do attitude with a willingness to sell and tell them to use Django so they can get away with not knowing a lot about key software concepts, and you can make a living off selling niche atrocious software.

Plus, 90% of what most small businesses need could be replicated in Wordpress with plugins.Plenty of money there too.


> in Wordpress with plugins

My first dev job we just crapped out WordPress sites, I mean WP was just a "headless CMS" pretty much, since it has auth/users/routing/editor out of the box. So we just built pretty pages/uis. It was interesting, in that level, the owners of the company were making bank. There was something I made that cost $30K and I was like wow... I was not paid anywhere near that. This other thing a single person developed was over 6 figures... I think that's an example of selling value over quantified time(within reason).

Oh and with WooCommerce you again had this "headless" shopping system and you could throw your own styles on it.


Oh yes.

My aunt paid $8000 for nothing more than a Wordpress gallery for her business and paid $100 for every change she wanted to make.

People who don't know tech will pay a lot for it.


Man whoever sold that sounds like they made bank. I'm not sure what upkeep you have to deal with regarding Wordpress security but yeah. Sounds like it's low-risk enough generally eg. just show pictures.


You periodically need to update the PHP version to keep up. I have a dormant personal website and maybe 1-2 years I have to go update the PHP when I get an email about it.


That sounds nice/convenient. I only recently started adding "down detectors" to some of my sites regarding the automated notification of any upkeep, cool stuff.


but how do you find these small businesses that need these types of things?


Lots of cases really are obvious.

One of my friends lives in a rural area where most businesses never got beyond the Church bulletin board and the phone book.

Suddenly they became cottage country and none of the local businesses were getting any business as they wanted to see websites or at least written down phone numbers and they don't use phonebook.

Take a walk through the small business section of your town and see who has a website. If they have a website, try to get them on Shopify. Lots of desperate closed businesses who will take a gamble on that right now.

The reason tech has not arrived is that there is no person standing in front of them making it extremely basic for them.


> If they have a website, try to get them on Shopify

So ... they then pay you for helping them get on Shopify? I can see how one might be able to convince a business owner that they could benefit from using Shopify, but is there room in there for a third party to sell configuration services?

Please forgive my being a noob with regard to Shopify. Their "getting started" page makes it seem like it's something the business owner just does on their own, so how does one come across as a _useful_ helper for that process?


> Their "getting started" page makes it seem like it's something the business owner just does on their own

You are vastly overestimating their computer savvy in many cases. Lots of small businesses barely have email. So even with the self onboarding option, many businesses need help.

But what I was thinking was more in line with this:

https://developers.shopify.com/custom-development

I am admittedly not a Shopify developer. I just have friends who make decent coin doing it while not being very good coders in general.


It is due to the fact that I have no, at least professional, experience. Well, at least that's what recruiters have been saying to me in the end of the interviewing process, after I got their take-home projects done. Working for small business seems the way to make money, while getting the chance to leave some paper trail (if this, hopefully, counts as professional experience).


I'm only a couple of years down the road from where you are, currently looking for my third dev job, having started with a 5 month stint in testing, now doing full-stack web stuff.

If you're having trouble getting your foot in the door, like I did, I'd recommend looking at software developer in test jobs. Depending on the company, you'll get a chance to touch a few different codebases (in my case, I got to learn Swift and Objective-C, and sharpen my Java), and you'll get to work on a team, using common tools you might not have been exposed to yet (e.g. Jira). It's a great, relatively low-stakes way to get into the industry.


It was hard for me to start too(no degree). I freelanced first online, it was not easy/I did not make enough to live on(had to wash plates), but that was "proof" of paid work at the time then I got into an agency doing front end work. Now I'm a contractor SWE at a "local" corp, got picked up by a recruiter through Linked In.


This thread seems to suggest that there are heaps of 20$ bills lying all over the world and accessible to any software engineer.

That they are not being picked up suggests that there's some other constraint. Either extracting this value is not really worth one's time, or the bottleneck is not the ability to automate things but some other rare skill.


There is a huge amount of these $20 bills lying around in fact. The thing is they're not that easy to pick up. These ERP systems are often fairly bespoke, designed for the workflow and needs of specific industries, or even specific niches in industries.

I work for one of these bespoke ERP companies, our niche is CNC machine shops (huge industy you don't hear much about). Our software was grown organically out of tools we made to manage our own CNC shop, spun out into its own company. That deep domain knowledge is what allows us to make a competitive product.

If this kind of stuff interests you we're hiring: https://www.proshoperp.com/company/careers/


> These ERP systems are often fairly bespoke

That is the paradox with ERP: they are sold as off-the-shelf solutions but customization costs twice the licenses and always ends up producing a bespoke monster.


I'd liken them to LEGOs. SAP sells modules (bricks) and integration (services to make the bricks do what you need).

They also sell some easy systems, like a LEGO car, but usually it's better to strip them down and reuse the parts of what you need is a house.


But the end result always requires you to cut/reshape the darn Lego bricks themselves, because the modules never quite fit the use case perfectly, so consultants are brought in to hack it all into some sort of workable shape.


> This thread seems to suggest that there are heaps of 20$ bills lying all over the world and accessible to any software engineer.

I argue that they are. It is just that the dollar bills are stuck in the mud or hidden under a leaf or just down paths you rarely take.

The bottleneck is knowing about them. You need to be the kind of person who thinks everything is done poorly and then thinks of ways to do it better.

The other bottleneck is that extracting them frequently requires annoying human interaction, but if you are willing to do it endlessly, there is plenty to do. I'm not and I lose out on a lot of projects as I don't want to do that.

But I know some absolutely terrible software developers (i.e. expect to see your password in the source code kind of developer) who outearn most devs I know simply by selling vast amounts of solutions to the thousands of people they meet every year.

Spend the day talking to random shop owners. Spend the night hacking together a fix for them. Collect an ongoing fee.


This. I think one of the places where young entrepreneurs get stuck is in wanting to build something that gets noticed as much (or more) than building something that makes them financially successful. There are thousands of problems that need to be solved and can be solved by software. And almost all of them are super boring and niche specific.


Engineers are unaware of the needs of these businesses, and often, the businesses are unaware that it's possible to solve some of their problems with software. That's why successful engineer-founders often have prior industry experience, or a co-founder who is an expert in the given industry.

I remember working on a construction survey crew in high school, where all survey measurements were recorded by hand in notebooks, and then manually transferred to Excel to perform some trivial operations. The office had hundreds of notebooks lying around. The leadership had no idea it was possible to have an application built to solve something like this. Perhaps they're using something now — I sure hope so.


Maybe it's also because some software sounds dull and old fashioned so people don't talk much about it.

I am into ERP, WMS, TMS and BI software. You don't hear much about this on HN but worldwide it's huge because it is what logistics run on.

Most of the time it's also old software and I believe there is a lot of money to be made. For example machine learning can be used to estimate goods that have to be purchased to keep optimal stock.


For software people, the missing link is knowledge of a domain. Not domain expertise, the rare part is merely knowing that domains even exist to be searched for $20 bills.

The article mentions actuarial software for funeral homes. In retrospect, of course that's a thing, but asked to name low-hanging domains for lifestyle software businesses, few of us would ever arrive at that answer, let alone in the first 500 guesses. And is it sexy enough that we'd stake our livelihood on it?

I'm reminded of this XKCD comic: https://xkcd.com/1425/

People in these small niches can't easily tell the difference between things that are too hard to be worth the money vs things that are solved by a for loop in an afternoon, and they don't talk about their work in a way that makes it obvious that there's gold in the hills.

This is one reason that user research is an entire field that gets paid almost like engineering: framing the problem can be just as hard as the problem itself.


It's also funny how nowadays such bird detection would possibly be just another few hours.


I really like that XKCD strip. A problem that was once hard now can be solved by a teenager with Tensorflow and a bunch of pictures.


Software engineering skills alone are definitely not enough.

There are a ton of these heaps of bills, sure, but they aren't lying around on the floor so much as on hilltops and mountains, to abuse an analogy.

It may help to think of these opportunities as a form of technology arbitrage. In order to take advantage, one or more people have to a) understand the technology, b) understand the problem domain c) understand the workflows and d) identify and communicate effectively to the people who can actually make the decisions.

How many people do you know who are both competent and interested in all of these aspects? It's not that any of these skills are particularly rare, but they are rare in the same person and there aren't natural communities, for the most part, of people who can do all aspects.

Another issue is sometimes (often, really) the technology aspect is the smallest and least interesting part of the problem.


To me it seems that large parts of 'average' office jobs are things that a software developer would automate so they took a tenth of the time. Ctrl-C/Ctrl-V and undo are advanced computing to many people.

But of course, the people doing those jobs aren't going to go out of their way to make 90% of their job redundant.


A lot of it is developers not understanding, being aware of (or simply being uninterested in) the problems other businesses face.

And businesses not knowing what _can_ be done with software to commission developers to solve it for them (or if they do, the lack of empathy from the developer ends up with them building a half-assed solution).


The ones I've seen are a right-place-right-time kind of things. One company I saw scanned and collected auto parts instructions to be searchable via a web UI. I'm not going to come up with that out of the blue. These people are close to a problem and figured out an automation step that adds a lot of value.


Personally the biggest difficulty I have had is getting ops people to change their habits. Software is only useful if people use it and retrofitting software onto existing processes is infinitely more challenging than building a software driven company from the ground up.


>> "heaps of 20$ bills lying all over the place" This isn't the case from my experience. I've networked with and interviewed dozens of people. It's incredibly difficult to find worthwhile problems that can be automated and aren't already automated. And, even here in the bay area, people tend to react pretty negatively the moment you even mention "repetitive processes". No one likes automation: they associate it with getting fired. So, if you ask people about this, they tend to clam up and stop talking pretty quickly.

And when you get right down to it, and start having the type of deep dive talks need to have with potential customers, you'll quickly discover there's already solutions to almost every conceivable problem out there, even in the some of the most esoteric fields you can imagine.


The Bay Area would have far fewer easy problems as every 2nd person is a developer. It is fished out.

> people tend to react pretty negatively the moment you even mention "repetitive processes". No one likes automation: they associate it with getting fired.

Those aren't the people you should be talking to.

> you'll quickly discover there's already solutions to almost every conceivable problem out there, even in the some of the most esoteric fields you can imagine.

That is irrelevant for most as you sell them the implementation of that solution then. It is a $20 bill, not a Y-Combinator worthy idea.


Did you know there's software to tell you how to unspool fiberoptic cable when laying underwater lines?

It's one massive long integral math equation.

It rents (yes, rents) out for $20K a week.

There's also coffee logistics software. There's software for private club management(tennis, polo, etc), moving and storage, even software just for spa's.

There is so much opportunity out there, the trick is finding it, and finding a working sales method.


It's sales and marketing.


There seems to be more people doing sales/marketing than software engineers (though I don't have any numbers).

Why aren't they picking up the bills?


Because once they get involved they're often not worth picking up.

I think there's a lot of stuff out there that would provide value to both parties if they could simply connect. Something like Gumtree or Craigslist is an attempt to capitalise on this.

e.g. I have a broken machine. It'll be useful to someone for spare parts, to fix and resell (or for fun), as a placeholder (since it sort of works) until they get something better. I'd probably give it away. But dealing with all the time wasters on such platforms makes me want to just toss it.

It's similar for some ideas where there is definitely profit to be had, but simply connecting with the customer, convincing them that it will help them, convincing them you're actually honest, etc etc, all costs more than it's worth.


There's few, if any, people who can sell these things to a new audience, and they're all picking up their own trillion dollar bills.


Two things:

1) Being a non-technical founder is expensive

2) Being savvy at online marketing means you probably were able to figure out that the cost of customer acquisition was higher than the price people would be willing to pay for the product

The reason businesses do inefficient things is because humans are stupid habit-driven animals. To break them out of their habits, you have to assault them with expensive messaging and literally shake them until they realize you have built a better way.


These bills dont exist, they are an oasis. I dont believe in the idea that easy millions is just laying around


There are millions laying around.

But it takes a long long time to pick them up.

"overnight successes" can take years.


There are tons of things that are worth the time to automate but aren't because they aren't worth the additional cost to sell the automation.

Just because something can be done automatically doesn't mean companies can magically discover it and pay for it. You still have to sell it to them.


I work for a Fortune 150 manufacturing company. When I started working here there was no IT group at all (it was all outsourced). The CEO and President said many times that we were not a software/technology company and that would never be a core focus. Nearly 2 decades later, technology is now listed as number 2 of our top 5 core focus. We are working on converting all contractors into full time employees. The IT group is no longer outsourced and takes up 20-30 buildings, out of 80 buildings on campus. We are still a manufacturing company, but software (websites) has become a huge part of the company with thousands of employees working on it.


I have learned to avoid companies that don´t understand that they are a software/technology company. You can't scale up if don't automate. IMO, It is just an excuse to do mediocre technical work.


What do you manufacture?


What kind of manufacturing applications does IT manage?


Mostly sales websites and design applications, not manufacturing applications.


Not just software, but everything. Look around you. Everything around you was made somewhere. The example I like to use is shower curtain rings. Those plastic or metal things that hold your shower curtain onto the rod. Somewhere, there's a factory that makes those. The person who owns that factory is probably rich.

My first job out of college was working for a company that makes tilt-tray sorter systems. These are giant machines that go in warehouses and sort packages into, for example, different bays to go on to trucks to different places. It's like a little roller coaster for boxes. It was something that I had never even considered the existence of before I worked there. But someone has to make it.


When I have kids, I'm gonna play a game with them where we look at random stuff and try to make up how it got there, and then compare that to how it really got there. I bet if you did this consistently with kids during their formative years, it would be a great way to teach them about the world and the economy. My parents just had a huge book room with encyclopedias that I was expected to pick up and read.



This is a political tract. The only actual information that it conveys could be obtained more easily by googling "How are pencils made?"


...

Something can just be interesting to read. Especially for kids.


How would figure out how it really got there though?

That you can, for the most part, peel back layers and trace steps back to an origin is one of my favorite things about code.

But for a heterogeneous system of opaque supply chains in the physical world?

For any given random stuff, I don’t know that I could make reliable progress in that endeavor.


Sending someone at <wherever> an email saying "I'm working on a school project. Can you tell me about X, please?" probably has a higher success-rate than you'd think.


> For any given random stuff, I don’t know that I could make reliable progress in that endeavor.

I bet you'd get better at it after doing it for a while, and that in itself is a valuable skill.


I’m expecting our first in a few months, so I’ll pass this great idea along to my husband, a mechanical engineer.


All the best of luck with the birth, from all at HN. Remember to create them an account :-)


That actually sounds like a good idea. Favoriting this for later.


You know the ends of shoeleaces, those little plastic bits? Roughly a 800,000,000/year industry

Cardboard box industry in USA is bigger than the NFL, NHL, NBA, and the other major sports leagues combined


If you had kids of the right age, into Phineas and Ferb, you'd know those little plastic bits are called aglets ;_)


i’m really sorry to do this to you, but also, if you’re a recent grad of the right age ;)


> Roughly a 800,000,000/year industry

there's 7 billion people on this planet, and most of them wear at least 1 pair of shoes (so 4 end bits), which means there's about 4x7,000,000,000 pieces of these things.

I'm surprised it's only a 800,000,000/year industry! That means they are able to produce it so cheap (i.e., less than 2c per bit, which is an over-estimate)!


Yeah, now multiply 28 billion by... three cents.

Also, just because an industry brings in lots of revenue doesn’t make it a good industry. I’m sure there’s trillions of dollars exchanging hands in the plastic bag market, but I’m also sure that the margins are getting squeezed so tight that even a PhD in OR couldn’t find a percentage point to optimize. The margin per bag is essentially zero.


> The margin per bag is essentially zero.

that's a good outcome imho (and it's not zero, it's just very low). But this means that the value of the bag and the cost of production is very, very close, meaning it's available to as many people as possible!


You assume that on average, each person buys 1 pair of laced shoes per year. How realistic is this? Maybe affluent consumers in developed countries can afford 1 pair of shoe per year, but I doubt it's affordable for consumers in developing countries (where most of the world's population are). Even in developed countries, how many of those shoes have laces? Many women's shoes don't have any laces, for instance.


You know, it's called estimation. Maybe come up with better number?


Those are called aglets, and when I just now googled for “aglet manufacturers”, I got way more hits than I expected.


I like the TV Series “Rotten” on netflix... it basically shows every food, fruit, meat, vegetable, you name it, is actually heavily controlled by cartels and mafias.

I had no idea there were mafia cartels that had wars about avocados.


Huh, hadn't considered this. But it absolutely makes sense. United Fruit Co. used to be larger than Google in today-dollars.

"The Fish That Ate the Whale" by: Rich Cohen is a phenomenal read.


Oh yeah, for sure. It's all just land, which is controlled by mafias or corporate-type entities. United Fruit being the major wakeup call for everyone who reads about that situation.


I assume you mean that town in Mexico? They had a quite a few articles in the US press about them.


Avocados were a big driver of crime in New Zealand pre-COVID. So high-value that orchards were being raided and stripped overnight. I can fully imagine violence around this.


yes, and the global supply chain is a weird place. Because I also know that these industries seem to spring up overnight and disappear just as quick.

From an american point of view it doesn't make sense for the amount of capital investment to happen so quickly, but that does seem to be true. Entire products with complex supply chains.

Entire industries seem to be born and then completely die much much quicker than I thought possible.

I would love to understand how all that works more deeply. Seems interesting.


This is somewhat true. However, unlike a regional coffee/fruit cutting business software still has fewer geographic barriers. This means winners will take (relatively) more of the market compared to more traditional businesses (and thus small software business is somewhat less stable).


I keep running into organizations or industries that don't spend money on tech. They don't put it in the budget, or they choose to continue to use volunteers to accomplish tasks that really should be done with software.

I am in situation after situation where I want to create software to make someone's life easier, but I know that in doing so, I won't make any money and it will make my life harder.

I wish there was a way to make money creating software that could really help organizations that won't pay for the tech they need.


the fact that organizations don't want software "they need" shows that those industry's companies are not competing against each other. If they were and the software really did make a big difference in their bottom line, then the ones that used the software would be stealing market share from those that don't use it. unfortunately, with a burgeoning govt and ever more companies that have monopolies, there's getting to be less and less competition and efficiency in the world


A lot of it really comes down to explaining that labor cost is equally a cash cost compared to spending on software.


For some bigger companies it's not, software being capital expenses and labor being operational expenses. They're taxed and budgeted separately.


That's because this stuff isn't valuable because it's software, it's valuable because of "business value" and expensive because of "customer support".

Seriously. I can list tons of things that are reasonably profitable, but I, personally, don't want to do the hard work of customer support so I won't even start.


> Seriously. I can list tons of things that are reasonably profitable, but I, personally, don't want to do the hard work of customer support so I won't even start.

This in itself is a business idea. We need an Uber for Sales Support Staff that entrepreneurs can just rent by the hour.


I'd never use that because support is mostly dealing with company specific problems.


That will work at some point in the future, but it won't work right now, because B2B sales is still all about human relationships.


"Thousands of people are making millions in software and they do not work for any of the tech companies you have heard about in the news" may be a better title.


I’ve been saying this for years. There are THOUSANDS of small businesses out there who are mostly being ignored.

Many, many, many companies still see shared excel spreadsheets on a windows shared drive as a technical innovation.


> Many, many, many companies still see shared excel spreadsheets on a windows shared drive as a technical innovation.

Not just small ones...


There are hidden "unicorns" everywhere. Sometime ago I heard Germany had a lot of "small" companies that are leaders in their market niche but you don't have it in the radar.

For instance: Hohner (the Harmonica manufacturer), or Flexi (dog leashes)

Every country has some of this hidden unicorns; but it is easier to spot them from the distance.


For what it's worth, those German Mittelstand companies are the exact opposite of what's usually referred to as a unicorn. They are small to medium-size, multi-generational family-run and family-owned businesses that focus on a very narrow set of products, such as just making gears for automative applications, or ball bearings, or blades. The people working there often work there their whole life, with the understanding that the company doesn't take any gambles and will do its utmost to not lay them off even in recessions such as right now.


Doesn't unicorn just mean $1B+ valuation?


And that hasn't IPO'd yet.


I think it has to be a startup, i.e. recently founded, hyper growth, usually VC-funded.

A company worth a lot of money that has existed for 80 years isn't a unicorn, it's just a company worth $1B+.


When is a startup not a startup anymore. Are Facebook and Google still startups?


I like Steve Blank's definition of a start-up: "a temporary organization used to search for a repeatable and scalable business model"

Once you're executing on a scalable business model, you're a regular business.


In my opinion they are not. They were and became unicorns, but they're not any longer, they are regular Fortune 500 corporations.


Heh, in Finnish the general term "fleksi" means an adjustable dog leash. That's how ubiquitous the Flexi is in that category. Like "styroksi" for EPS (from Styrox) or "kyprokki" for drywall (from Gyproc).


This is because of a few factors, mostly availability bias. We find it easy to think of high-profile examples of tech companies, so we assume they are the industry.

Previous comment with fancier wording: https://news.ycombinator.com/item?id=17994976


> “Wait until you find out about the $100 million company that chops all the pre-cut fruit you buy in supermarkets,” he said.

Imagine you harvest 1 fruit, chop it and package it, and then duplicate the result millions of times. Then each time you sell your chopped fruit, it's somehow teletransported directly to the customer.

That's software.


This is why things like "most used languages" coming out of places like StackOverflow are laughable to think they represent the entire software industry.


I have to think there are lots and lots of industries ripe for disruption or even for their first taste of software, and it's just that they're outside the radar of the vast majority of engineers. Industries you never ever hear about unless you're an insider. Maybe we need some mechanism for cross-pollinating people from "deep industry" with entrepreneurs. Many of the "visible" ones have already been saturated to death with startups, and it's just because those are what software people happen to be aware of.


maybe they don't want software people f#cking up their lives?

I know it's hard for the avg hackernews to understand but software isn't always a blessing but many times a curse bestowed upon workers by someone in the mngmt chain who has been sold on a solution.


I was going to go home to the US and visit one or two companies per day that I felt were underserved. This was supposed to be early March. Didn’t go as planned.


Yes I agree with this, but it's not that surprising. I never thought that Σ(companies mentioned in TechCrunch articles) = whole software economy, or even 1% of it.


This is similar to Gell-Mann amnesia in that I had relatively little to do with writing this law ;) (For what it's worth, I think it's likely descriptively accurate, but not sure it's an observation worth the namespace.)


It's a shame it wasn't called "McKenzie's Law". Has a much more serious ring to it. ;)


Motivation to maybe get off my arse and actually build something useful.


Jesus Christ, it's almost like the US economy is mostly ran by small businesses, what the fuck.


Actually the trend is in the opposite direction, towards more enterprise size businesses.


Both are probably true - it's probably true that most of the US economy is SMEs, and that over time enterprises become a larger factor of the economy


I googled where IHOP gets their coffee since it's so pleasant. Royal Cup!


Wow! Very cool. They are a fantastic company.


Admin software for schools - ran into an ex teacher making an absolute killing selling some admin software to schools. And he’s one guy selling to schools in one state!


More patio11 laws: https://threadreaderapp.com/thread/936615043126370306.html

A variant of the current law also makes an appearance:

"The amount of money flowing through capitalism would astound you. The number and variety of firms participating in the economy would astound you."

My picks are:

"Startups are (by necessity) filled with generalists; big companies are filled with specialists. People underestimate how effective a generalist can be at things which are done by specialists. People underestimate how deep specialties can run. These are simultaneously true."

"You radically underestimate both a) how much you know that other people do not and b) the instrumental benefits to you of publishing it."


> a) how much you know that other people do not

this is a particularly confounding problem, and i haven't figured out a way to combat it.

it's so bad, that i'll often learn something new when scrolling twitter or clicking through wikipedia, and then immediately assume that 1) i'm the last person to the party, everybody else already knows, and 2) forget what it was like not-knowing this fact mere seconds before.


Basecamp is a famous example of this.

It's a pretty popular project management software, 56 employees at this point if I remember correctly, probably millions in revenue. It's a tiny, bootstrapped company, no investors breathing down their necks, no IPOs, nothing. And they're one of the few companies who actually have principles, are committed to privacy etc.


Because it’s almost impossible to switch out of Basecamp and use a competitor.

The vast majority of companies like this have perfected the lock-in strategy to ensure that their customers can never leave.


> Because it’s almost impossible to switch out of Basecamp and use a competitor.

Is that unique to Basecamp or does it apply to all similar software? I haven't Basecamp all that much. What exactly do they do to ensure this?


I do recall Basecamp from day one allowed you to export your data which not many other people did. What you do with that and how you get it into something else is another story but also not really their problem.


I remember this too, but the second point is the crux of the usability and lock in. It kind of relies on another product having an importer to ingest it to another product/service. Or you have to pay a consultant to do this for you (write a converter and run it to do the import). Either way it does introduce some friction to leaving and using another service.

There's also the idea that if you provide an easy way to exit the service then people are much more likely to buy the service. Since they know they can easily get out of it if they don't like it.


I don't think this is correct. They have stated in several of their books that they expect their customers to outgrow BaseCamp.


So what you're saying is it's not because they have a good product?


>> software companies he’s never heard of in Oklahoma pocketing $10m/year in profit

Every industry seems to be full of these. The last place I worked is using timeclock software that doesn't look like it's been updated since the Windows 3.1 days. They probably paid someone $50k at the time to develop it, and the minimum installation of it costs $25k.


When talking about the size of an economy, it's more useful to give actual statistical calculations indicating hypothesised/alluded size than to offer anecdotes. Has anyone actually studied this and mathematised their thinking? One can't/shouldn't make business decisions based on anecdotes or bromides.


I was surprised to find out one of my coworkers had been maintaining the point of sale (POS) machines for a local chain of comic book shops, running hypercard (yes, that hypercard) since the 1990s. The POS machines would sync their databases via write ahead logs over dialup


Nice to see HN considered "mainstream tech coverage"...


And yet I have no ideas. Maybe I should talk to my parents, they might have seen something in their old jobs that could have used automation.


Corollary: the most popular hot-in-SV language has nothing on C++, Java, or even C#. They're used widely inside the dark matter.


I call these companies RHBs: rich, huge and boring.


The statement reminds me of GEB:

"Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law."

https://en.wikipedia.org/wiki/Hofstadter%27s_law


Therefore it follows: everything is software and we are all living in a simulation.


Thanks for mentioning ConvertKit! We're happily building our niche product for creators and growing quickly.

Just a couple corrections, at the end of 2019 we were just under $20M ARR (Not $24M). Today we are at $22M ARR and 30,000 customers.

You can see all our numbers (updated in realtime) here on Baremetrics: http://convertkit.baremetrics.com

I'm happy to answer any questions.


Very surprised he hadn't heard of ConvertKit. It's usually in the running whenever I google for email marketing/automation software. Price seems reasonable too (but not for us, we're more newsletter-based rather than conversion-based so it doesn't really work out).


Can I throw my hat into the ring? I run an email marketing platform (https://emailoctopus.com) that’s affordable and aimed at entrepreneurs/makers with newsletters. Shoot me an email if you’d like a discount.


Sure no problem, I have heard of it before, there are just a lot of options out there and it takes time to dig into each one when I only help out part time. Having a "start for free" option is definite a good idea and I probably will check it out soon.


i have now! my business partner introduced me to it, and we are happy convertkit customers. fantastic product.

the crazy thing to me is none of my friends have heard of them.


hey nathan - big fan of what y'all are doing. also love boise (i lived in idaho last year). updated those numbers and added attribution to baremetric


why did you make it public?


This is hilariously self indulgent to the point of being cringeworthy.

Along with the comments about basic products being worth billions to sell.

Industries are deeper and wider than we assume, supposedly this is revelation for some?

Is there a name for laws that gratuitously reference themselves?


I dunno man, sounds like bullshit to me. Some big brained engineers spent decades programming fizz buzz in brain fuck. This is temple OS for the We Work Generation.


What gets you infinite upvotes on Reddit doesn't necessarily work here.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: