>The most interesting reason why now is a good time to start a service relative to other times is automation enabled by AI
A good time to start a business is when you know a lot of people with some pain who could become your customers. Obviously, you should know how to solve their pain better (i.e. making a better design) and more effective than others.
There are lots of existing businesses that are not stupid. Of course a lot will try to integrate AI into their workflows. They have clients, you don't.
> How do you find that first customer? Talk to the people you would have recruited as your first users if you were starting a product company.
How to find customers? Go find them.
> Instead of convincing them to use your product to solve their problem, just solve their problem yourself. Charge more than you would’ve if you’d sold them a product!
They have some pain, and you are trying to convince them to pay more because you are spending your time to solve their problem yourself, but you will use AI.
There are existing businesses that already have pipelines and experience solving their problem, with or without AI.
Not really, my point is very different! I love people starting and doing things, making products even for a small audience.
Starting a business because you used an AI is pointless. There are way more things you have to know and also don't forget that other people likely know something about AI and/or consulted to incorporate it to their existing business. It just can't be a single/biggest reason to start a business.
Really confused to see the conclusion of "It's a good time to start a service business because of AI". AI can't perform, or even semi-perform, a single one of the example services the author used except maybe "design agency".
But otherwise, it does seem that starting a service business is sort of the ultimate "do things that don't scale" strategy.
A service business is about selling employee time. Anything you can use to speed up the non-core tasks goes straight to the bottom line percentage. After that anything you can use to have management "run" more people also goes straight to the bottom line, as a multiplier.
So there is tons of potential in AI-based tools. Is it there yet? I doubt it. But perhaps yes. Anyone here getting traction in AI for business routine efficiency? Back office? Accounting? Time-keeping? Task dispatching? Email requests dispatching? Audit?
But also possibly, having a sales team running the simple cases (mostly) straight through an AI "junior engineer" before even talking to the more expensive guy.
Even if AI doesn't help with any aspect of the service itself there are a ton of efficiency gains to be had from "the rest" of the work, which will make service margins less ugly. Every business still neeeds some form of sales and/or marketing, support, and back office administration.
And for a lot of services there'll be creative ways to squeeze efficiencies from tech/AI. For example roofing businesses used to send a guy to your house to give you an evaluation + quote; increasingly, they now send drones (https://www.roofer.com/).
The claim that there's a lot of efficiency to gain from _tech_ isn't interesting. I'm guessing when the author talks about AI, they're not suggesting to build a fleet of drones.
AI increases productivity. If you are selling your time that means higher margins. It doesn't have to do everything. If it increases your productivity by 10% that is all profit.
Caveat: I'm approaching this from a 'indie hacker' mindset. The article mentions $10 million funding rounds, so they're probably approaching this from a different mindset.
The article distinguishes between a product business and a service business. In the former, the founders find a problem, build a solution, and then sell that solution to people who have the problem. In the latter, the founders find people who have a problem, and then build a solution for that problem.
In a product business, the risk is that you solve the wrong problem. In a service business, the risk is that you never find a problem to solve.
I used to be a schoolteacher, in schools where teachers tended to be on the verge of retirement. Most of my colleagues used computers because one day their typewriter had been taken away and a computer put in its place. Their technology use tended to be quite basic: emails and basic word processing. Most workflows were unchanged since the typewriter and mimeograph days. Almost all planning was still done on paper, on especially-printed planning books. (We were individually asked whether we wanted week-to-a-page or day-to-a-page when the stationery order was being prepared for the next school year.)
(Lesson planning is not a problem to be solved; I'm just using it as an example of the kind of mindset I was working alongside.)
This is a problem-rich environment filled with people who don't know that many of their problems can be solved by computers. You can't ask these people what problems they want solved, because they don't know that their problems can be solved. You have to have something to show them and point to and talk about before they can grasp the idea that it's even possible to automate some of their work.
However, I speak from a position of privilege. I spent years working in that environment, and I walked away with a list of problems - a list of potential products. If you've only worked in software, you won't know any good problems. You'll know some bad problems, and they'll be bad because your potential customers - other software engineers - are just as capable of building their own solutions as you. In that case, a service business is probably no more risky than a product business.
If you start a product business, your moat is the insider knowledge you have of how to solve the problem you solve. If you start a service business, your moat is your reputation for solving people's problems. If you have no insider knowledge, you start your business with no moat at all, and have to dig it while building your castle, whereas the founder with inside knowledge just has to build their castle within their pre-dug moat.
Is the kernel of this advice to start with a known problem which is already solved with a paid service, and then just sell that service more efficiently? That's how I'm interpreting the "skip the R&D phase" idea. In that case I think the question becomes "why is this a better strategy for a startup", since this framing implies going after well-established markets with (at first) an undifferentiated product. Both 'productize' and 'automate' sound like things that incumbents would be able to do more effectively than startups, and likely has already happened in most cases silently in the background.
> "Is the kernel of this advice to start with a known problem which is already solved with a paid service, and then just sell that service more efficiently?"
Yep, this captures what I'm saying.
> "why is this a better strategy for a startup"
If you can build a differentiated product it's great to do that, but it's a lot more difficult/expensive than it used to be.
Incumbent product companies will pursue full automation, but probably won't compete directly with service businesses because it's not their wheelhouse. E.g. GitHub built Copilot and maybe they'll try to ship a "fully automated" AI software engineer. But until their AI software engineer is actually fully automated (which I would guess is 10 years away at least), they won't be competing with software development agencies.
We started selling services in earnest alongside our SaaS product this year, and it’s working out well. We didn’t think of doing this during ZIRP; it was relatively easier to invest in more rapid product development.
Palantir is a classic example of what is a hybrid services => product business. It was a _very_ long road for them to get to a point where their product could actually scale without massive implementation effort, but I think they’re finally getting there (they offer a cheap self service option now AFAIK)
One of the interesting bits of information that came out of Microsoft's acquisition of Github was that at the time, the majority of Github's revenue was enterprise deals. Enterprise is a code word that product businesses frequently use for services - yeah sure you have a product, but here are these guys who want to use it to the tune of $10M/year but you will need to do custom development just for them, so what is that? That looks a lot like a conventional service business running inside the same company as your product business.
Microsoft actually seems to specialize in that sort of thing (selling a software product, and then offering some quite expensive, bespoke enterprise consulting services for it). But they are not the only ones, looks around the industry and you'll see it everywhere.
The only real drawback to this model when you're small is that it dilutes your focus and multiplies complexity and cost because you have to manage multiple very different lines of business simultaneously. But if you can overcome those challenges it can make a lot of sense, be an expert at a particular problem, sell a product for the average Joes who have that problem, do bespoke consulting services for the money-laden enterprises who have their own complex versions of it.
Not quite, the onprem version was older and didn't have Actions for ages, for example.
But to be fair, basically all enterprise deals involve services as you generally need to train people in how your thing works if you want to get enough adoption for when Finance come looking for operational efficiencies.
It's been a while, but I think that enterprise customers all had access to both enterprise cloud and on-prem. What I was really getting at though is that there was no custom code or consulting with the vast majority of GitHub's enterprise deals.
I think its less about custom code for Large Company Inc., but rather, when Large Company Inc. had a bug or an feature request that was prioritized and perhaps released to them first
When it comes to analytics or machine learning type products, the scaling without massive implementation effort problem seems to be pretty universally hard to solve. Of course the usual big names seem to manage this on some level but startups seem to get stuck at building a handful of great solutions for a handful of clients/problems but unable to increase the number without also increasing the effort proportionally.
Do you know of any examples of small scrappy places that have figured this out? Is there any good writing on what kind of organisational and engineering approaches are necessary to make this step?
This is because data is basically a service business. Like the value is the specific data and it's hard to build a product that works for customers without a lot of custom work.
If the service opportunity is to resell AI skill to the people who lack it, consider productizing and selling Education instead. Otherwise the clock is ticking for a new model to put you out of business. Most folks will need to learn “AI for ___” eventually, and product businesses are way better to own. Service businesses are a lot like being employed. (Having run both.)
On a really long timeline, it's probably high-end services businesses that have the most job security. I'm thinking of professions like Michelin-quality chefs, performance coaches for top CEOs and athletes, etc. These will likely never be replaced by any sort of automation, at least within our lifetimes.
I'm not sure about chefs (though I do have some opinions), but I'd be surprised to see much success for any CEO or athlete who would stick to a human performance coach once they can have "Her" level AI whispering expert-level advice into their ears at real time.
I don't think professional athletes are that dismissive of the collective wisdom of real-world experts, and are in fact much more likely to pay for the "human premium" and not a soulless chatbot, no matter how advanced its information.
But the LLM is almost literally "the collective wisdom of real-world experts" rather than a single one. Obviously it's not at that level yet, but I don't see a clear upper bound to it getting there. I'm not very familiar with athletes, but from what I know, regardless of the "human premium", if the AI coach helps their competitors perform better, they'll need to adapt or "perish".
Coaching, especially at elite levels, isn't about information. At that point the human connection to a coach with championship experience is better than pretty much any conceivable chatbot.
I'm confused, why would you insist on using an LLM for decision making? It's not
"the collective wisdom", it's a probabilistic language model. I imagine LLMs are better suited as a last step. I'd want a proper model to infer a decision, and an LLM to transform it to a human-friendly-formatted advice.
Any B2B vertical where customer ACV is less than $10k per year.
If the vertical has affiliated trade organizations when you go to their events, you’ll find either very few VC-backed startups or none at all.
The competition will be mainly service businesses without software chops. Sell cheap automation and price sensitive customers will move to your solution.
> Any B2B vertical where customer ACV is less than $10k per year.
I think any may be too broad. That encompasses things like:
- SMB accounting software
- Email marketing platforms
- Project management tools
- CRM software
- Website builders
- SMB HR and payroll services
- SMB IT and security solutions
- freelance/gig platforms
- SMB marketing services
- Legal/compliance for SMBs
And each of these seems like extremely saturated verticals with just a ton of offerings and competition.
This is where I get stuck. The idea of focusing on a small vertical that isn't attractive to venture capital funding appeals to me, but then I'm at a loss for ideas beyond the most obscure niches, like a bingo card creator :P
I think you're idea of a niche is a little too broad. SMB is not a niche, plumbers is a niche. Are there accounting software, or CRMs made specifically for plumbers? Probably. But you can make a better one. Don't like plumbers? Try electricians or roofers. How about solar panel installers? How about installers of a specific brand of solar panel? There may be many SMB softwares, but if you pick a niche, you will find less competition catering to that specific customer.
You're leaving out the big catch though, which is that if you don't have experience in these industries, the software you make will probably be terrible. The winning move is if you are a plumber that can learn software engineering, not the other way around.
Ah, that's helpful. You're right, what I listed are not niches. I'm wondering if having some knowledge about a particular niche is necessary to determine whether a tool for that niche, such as a CRM for plumbers, would be distinct enough from similar tools for other businesses. Or do you think you can generally assume that there will always be some way to tailor a tool to a specific niche?
Sorry for the off-topic, but reading this text I thought about myself.
I find myself at a critical point where I wonder what I should do next. I'm over 40 years old and have accumulated years of experience. Sometimes I feel like Neo when he controls the Matrix. Maybe I've been in the same role for too long, but it doesn't matter what project I'm assigned: I implement it with little difficulty (beyond the time cost, exhaustion, etc.).
On the other hand, despite the confidence they've always had in me, the place where I work is starting to feel hostile. Being an introverted person, I try to figure out if this is something that depends on me (and if I can redirect it) rather than looking for answers where I should. Being introverted, it seems unwise to start something that depends on my social skills. Conversely, every day it seems like new obstacles appear, as if going to a job that I liked 90% now feels like a problem.
I have to say, I work in a flat-structure company where the boss is a megalomaniac who wants to control everything. A good part of the new things we do are his impulsive ideas. The office is filled with figurines (dinosaurs, busts from 80s movies, ...) and motivational quotes. He gives talks about how well he does things and how great it is to work there, but he doesn't improve salaries or do anything to provide training, a better work environment, etc. He also plays dirty tricks on employees, like unexpectedly complicating pre-arranged vacations or trying to delay them without a real reason
While writing this, I realized that the 10% that made me not like the company has always been because of him. And if I don't like it, it's because of him.
The only thing that keeps me tied to this place (besides the money, since I don't live in the opulence of certain regions where software is very well paid) is the entrepreneurial spirit of having done so much in the product.
PS: If someone saw our product, the things we have done with just 4 people (on average), they would be amazed. And, with all humility, a large part of that has been driven by me, which ties me and kills me.
You deserve to be happy and you are not your job. I recommend not clinging to the work you've done. Treat all your work like a puff of smoke or a bubble on a stream.
Thank you for your comment. I'm working on it, to differentiate each aspect and to make it understood as well. It's hard for me because I have been badly accustomed to certain ideas. I do not rule out the idea of making a full reset.
very curious question to the author here: how is this different than the "gig" or "expert" economy thats offered from places like upwork and fiverr? these are technically all "services" offered by various people, but often feels like a race to the bottom (lowest price wins)
A good time to start a business is when you know a lot of people with some pain who could become your customers. Obviously, you should know how to solve their pain better (i.e. making a better design) and more effective than others.
There are lots of existing businesses that are not stupid. Of course a lot will try to integrate AI into their workflows. They have clients, you don't.
> How do you find that first customer? Talk to the people you would have recruited as your first users if you were starting a product company.
How to find customers? Go find them.
> Instead of convincing them to use your product to solve their problem, just solve their problem yourself. Charge more than you would’ve if you’d sold them a product!
They have some pain, and you are trying to convince them to pay more because you are spending your time to solve their problem yourself, but you will use AI.
There are existing businesses that already have pipelines and experience solving their problem, with or without AI.