I'll be interested to see a longitudinal study in 6-12 months.
FogBugz is a complex B2B product fueled by recurring revenue (either subscriptions or maintenance).
It's not diet pills or a self-help book.
Most (all?) companies find that the long-form, "Tim-Ferris-style" landing page converts better for impulse sales. But does that work well for non-consumer products?
My fear would be that if you 2x or 3x the top of the funnel with landing page optimizations, you run the risk of attracting the wrong kind of customers. Customers who initially signup and pay, but don't become long-term repeat customers at the same ratio as those who converted on the old landing page.
I would be curious, Patrick, if you're planning on looking at the lifetime value of these conversions, or if the scope of your project with Fog Creek was limited to lifting the number of initial installs.
(Oh, and like everyone else said: great job, great writeup, and thanks for sharing!)
according to the old school copywriters like Dan Kennedy who has also done lots of B2B work, "Tim-Ferris-Sytle" landing pages (also known as long form sales letters) work just as well in the B2B space (selling software, fax machines and high end office equiptment). The problem is that most companies selling who would benefit from the tried and true techniques from direct marketing are 'image conscious' and would rather have a feel good corporate speak landing page with a meaningless slogan like 'ideas for life' or some such!
It would seem pretty counterproductive for a web business that counts on repeat revenues--and even more so for a company with a monthly recurring revenue model--to not optimize CLV in favor of optimizing for short-term gains. Even if Patrick's work was a one-off deal, I'd hope FogCreek will be measuring lifetime value of these conversions vs. old customers.
Hideho everybody. I hope you like it. If you have any questions, I'm happy to answer anything that doesn't breach a confidence with them.
We're also going to be publishing a lot of good stuff in the coming months, so if there is a particular topic you'd love to see covered, we'd love your feedback.
The notes on the specific testing and implementation was great, but also the general tone of the piece. It can be hard to write about marketing to an engineering audience while honoring an engineering-driven culture.
For instance -
> FogBugz exists to help software developers write great software, but it also exists to pay the rent. This means that we have to actually consummate transactions on our website. This was previously delegated to the curiously named “Details” page. It had been designed primarily to explain to customers the difference between FBOD and the downloadable version of FogBugz. Sadly, this ended up being quite confusing.
Covers why marketing is needed in an extremely straightforward way (...it also exists to pay the rent). Then, it logically follows the page that converts needs to be easily accessible. And then, the notes here and elsewhere that non-technical people are often hitting these pages and getting lost trying to decide on the tradeoffs.
Great piece. Learned about the specific measures you took, but also a lot of meta-learning about how to present things well. Extremely straightforward, logical, acknowledging and honoring the engineer culture, while also noting that marketing needs to be part of the process to pay the bills and get the product in users' hands.
I respect patio11, but his bingo card earnings are far from impressive. What's impressive is the process he developed to optimize his earnings in a relatively small B2C market. That knowledge and expertise, along with the fame he earned by sharing them, is his most valuable asset. It just makes sense to capitalize on it with consulting gigs.
If you can earn 25k a year from a simple bingo app, put together a few more similar apps, then you have a decent salary. And so what if you think if his bingo earnings are far from impressive. That first sentence was unnecessary.
I understand where you are coming from, but I can assure you that my first sentence is not meant to be ill-spirited. Rather, it servers the purpose of setting up the second part of my comment which reinforces how much "patio11 is not just the bingo card guy".
I think that Patrick did an awesome job with Bingo Card Creator, but Joel didn't hire the guy because he manages to make $30K in software sales a year. That's not an impressive number for software sales in itself. Plain and simple. Patrick is however an impressive guy when it comes to the specific set of skills he has, so it just makes sense for him to capitalize (possibly far more than Bingo Card ever could) on that.
Any chance you'll be appearing on the stack exchange podcast anytime soon? I'd like to hear you guys talking and throwing ideas around.
I know you were involved in the Fog Creek side of things and not the Stack Exchange side... but that actually make it more interesting for me since I expect your ongoing series of blog posts will cover as much as possible of Fog Creek-related stuff.
What's your opinion on FogCreek's front page being left intact? Aren't a lot of marketing/sales opportunities being wasted by the huge "philosophy" block? Why isn't it more product and sales oriented. After all, I don't think customers are primarily interested in FogCreek's philosophy.
Customers for specific products don't usually end up on the home page. Fans of Fog Creek / Joel's blog / etc do. Slotting them straight into a FogBugz trial is not necessarily our #1 interest -- for example, we want to continue being known as a great place for engineers to work, because that is core to the brand and to hiring efforts. Also, one of my major areas of focus is making fogcreek.com valuable to people who have no interest in a commercial relationship with FC. For example, if y'all were to appreciate and link to stuff we write about running a software company, that would be a Very Good Thing for our ability to sell people other than y'all on bug tracking software.
Luckily, Google does a pretty good job these days of sending you to internal pages for entrances when appropriate to demonstrated user intent.
I'm not patio11, but what I can tell you as a customer is that your first page (from when you launched) did a better job at selling your product to me.
What I still remember from that page was :
1) Hand-made
2) Scarity (in a good way)
I just checked your website and 30 seconds after I don't remember anything about it.
David thanks for the feedback. You raise some excellent points. Designing a simple and direct site is easy when you only sell one product.
The challenge is how do you add more products while keeping complexity to a minimum. We still want to have a clear message to someone that lands on our page but now they have more options as to what they can buy on our site. I guess it is a bit of a trade off.
I understand what you mean but I think it's still possible (ex. Apple). When I want to do a quick evaluation of a website I usually ask a random person to load the website and tell me the first 3 things they see and to rank them in their perceived order of priority.
When I do the test myself on your website :
1) REX RAY DODOCASE
2) DODO Case - Protects From Extinction
3) Couldn't tell in less then 5-10 seconds
Is that really what you want me to remember ? Also, my comment apply to "Acquisition", repeat customers is another matter.
PS : Sorry for my poor grammar/syntax, english isn't my first language.
"You can also see that the See Pricing link attracts more people than the free trial button, possibly because the free trial isn’t identified as being free."
Or possibly because the free trial is pointless if you can't afford to pay for the real thing after it ends.
I absolutely will not start a free trial for a product I don't know the price of.
I'll add the data point that I think this way too; the pricing is more important than the free trial, to me, as part of the process for acquiring software.
This may be one of those things though where normal people see the word "free", go "oooh lollipops!" and click.
This is so true, software sites that invite you to call, just so you can get some idiot to tell you about their "value proposition" and discuss your "licensing estate" really get my goat.
Well. Goat or no goat, for many software companies, this is the right decision; a salesperson will do a better job of closing an enterprise sale than a web page will, especially when the web page invites prospects to disqualify themselves because of their immediate reaction to pricing.
People investigating bug-tracking software may well not be "normal people" in the relevant sense. (Count me into the "will not try until I know the cost of saying yes" pigeonhole, too.)
Definitely. And I'm surprised how many companies don't get this.
I recently had to decline a free beta account for a cool, pre-launch web service because the company behind it had no idea how much it was going to cost once it launched.
Do you generally think --- gut take --- that a video of the founder on the front page (kind of costly to make, timewise) will perform well? Or is Spolsky just that kind of magnetic?
Gut take? If you had told me "We're an enterprise software company and a 3 minute video on our front page did really well, outperforming text and static images.", I'd say "I'll buy that." The hour long video killing it? That surprises me. This was one of a couple results where data said everything I know is wrong.
We'll iterate further on it. One of the challenges is that you can swap out text as quickly as I can write it and you can get an image crafted in a day, but doing a new video is a Project with a capital P.
My understanding: we measure it (Michael Pryor posted a link below), at least some people really do stay glued to the tube for a long time, and then they convert like crazy.
I'd be interested in hearing what you mean here by the video performed better, in terms of conversion to clicking the trial button, conversion to signup for the trial, conversion to becoming a paid user? Are you tracking results all the way down the funnel or just to the next step you have chosen?
Typically the team either tracks clicking the trial or actually starting the trial. We track stuff all the way through. Over the last couple of months of A/B testing, they've noted that the conversion rate of the trial page is not extraordinarily sensitive to A/B tests in progress before that page, so often we use "opened the page" as a proxy for "started a trial" because it gets results weeks faster. (BCC uses trials and not sales as the conversion of interest for much the same reason.)
"It ROFLstomps everything else we’ve tried there. We tried short videos, we tried screencasts, we tried images — in A/B test after A/B test, our users said Give Us More Joel."
Fog Creek's marketing has always seemed to me to center on Joelonsoftware.com, and then Stack Overflow, and also posts like this one. Unless I'm missing something we are part of the target market. Now, I don't have a problem with that, but it makes me read posts like this one (and this one: http://www.joelonsoftware.com/items/2007/11/16.html) a bit differently.
If you're reading this, you probably will never give FogCreek a dollar. That's the hard truth of the matter. We literally give the product away to startups (Student and Startup Edition, free forever for 2 users) because a) in aggregate startups demonstrably don't spend appreciable amounts of money on this, often preferring OSS or "we can totally duplicate that 100 man-years of development in a weekend" and b) dedicated effort at making sales to you doesn't make any sense relative to dedicated effort at selling 20 seats to an insurance company in Idaho.
But do we care about the community liking us? Oh heck yes. The average reader of HN is very link rich relative to e.g. the project manager for the IT team at an insurance company in Idaho. Producing stuff which you like enough to link to gets us higher on the Googles and thus into the line of vision of people who can spend more money on bug tracking than many startups spend on everything. Similarly, if you feel good things about us, you might recommend us internally to someone who has ability to make decisions about which product to purchase.
Giving it free to startups is not just "good faith" or "do good", "hire better" etc. it's getting user hooked up to the software, long term investment and almost free word-of-mouth marketing.
FC knows that FB is a tool that will show itself after a long time (hence 45 day) trials. When we start using FB our company was only 2 people now we are 9 and also we are a paid customer of FB.
For these reasons I don't want to believe that FC is not using channels like SO,Joel's blog as a marketing tool. That would be a waste of awesome free marketing.
Wait what? There's a free Startup Edition?!? I've been looking all over for free issue management software that doesn't suck and here Fogbugz and Kiln are free for little 1-man software shops.
What great news! I've been using the EpicWin iPhone app for issue management and DropBox for source control up to now. Every time I thought about it for any length of time, I'd get really nervous. :)
Signing up right now.
Edit: Also, this made me so happy, I signed up for HN, after lurking for over a year, just so I could say so. Now that's great marketing.
Jira with Greenhopper is almost free: $20 for up to 10 users. It beats the pants off of any open source solution. For us it won out over Fogbugz in the trials. When we pass 10 users (there's currently 5 of us), we'll pay their regular price.
Anecdotal evidence against the assertion that people reading this will never give FogCreek/Atlassian a dollar: we will.
The funny thing is that they just ripped off 37 signals' marketing page. Not to say that this strategy won't work - just that the "redesign" involved some heavy borrowing. I recognized it because I just used 37 signals' page as the basis for a marketing page on my own site.
I think Patrick is an amazing argument for bootstrapped entrepreneurship in a way that I haven't really heard mentioned enough... The freshness of perspective that comes from an incredibly bright person of one field (in this case development) thrust into other roles (marketing, sales, web production) seems awesome fuel for innovation and new thinking.
Thrilled with the article, and I think like many on HN thrilled to see patio11 killing it. Cheers, Patrick.
I know that you are a rails guy and they use an MS stack. Any friction there? Were they able to use your A/B framework or did they have something on their own stack?
We don't use Google Website Optimizer, which I won't recommend to anyone. We're using a .NET framework that I won't name because we're probably going to transition off of it.
I enjoy Rails a lot more than I enjoy .NET, but give me some credit: looking at new languages doesn't cause my brain to throw a WTFCodeIsCamelCasedException.
By the way, Fog Creek isn't strictly a MS shop. There is a lot of "best tool for the job" going on, including e.g. Python being used as utility glue.
Wow. So much practical advice in there. I'd heard much of it before, of course, but seeing it in action really helps cement it. I may have to read it again to make sure I caught everything.
The cause was... wait for it... Apache KeepAlive. For more details, searchyc.com for [patio11 keepalive]. It kills more blogs than cancer.
I should start a How To Make Sure Your Blog Does Not Die Horribly consulting business. And it will consist of disabling Apache KeepAlive, and then chanting arcane incantations in Latin about caching for the next hour to make people feel like they were really getting their money's worth.
...or just don't use apache unless you have a specific need otherwise. NginX handles tons and tons of alive but idle connections gracefully, and will still be snappier for users that want additional pages.
I first learned to disable Apache KeepAlives in 1998. Yes, 1998. It's disheartening that Apache still ships with it enabled by default. It has always allowed a relatively small number of lingering clients to completely DoS your server.
Have you looked into W3 Total Cache? I've heard it's quite a bit better at handling heavy traffic. I've never gotten enough hits to see, but it's worth testing.
There are a couple of specific results mentioned in that post. If you're wondering about specific "Did it make them money and, if so, how much?", the answers are "Yes" and "Fog Creek is owned by the Creekers, and like most privately held companies they don't share revenue numbers."
But if you're trying to guess how sore the marketing team's hands are from high-fives at the moment, here's an easy math exercise: make a guess as to how many digits of sales FogBugz makes in a year. Pick a number for a single successful A/B test -- I suggest 5% for this one. Multiply.
Longer sessions and less bounces don't really matter to us. (They strike me as easy to measure but of limited business utility. I can do a lot of things to make your session longer -- re-bork navigation, for example. That won't make folks terribly happy, either our customers or other Creekers.)
Conversions are much more interesting. Unfortunately, that gets into the sensitive territory. There's at least three levels, right: how many more trials are we getting? How many more trials are we converting into paying customers at the end of day 45? How many users do those customers add (FB is priced on a per-user basis, so an account with 50 users is worth 10 accounts with 5)? Then of course there is the churn/LTV question.
We're obviously quite interested in these numbers. I don't know how much I can tell you about them specifically. How's about this: I wanted to increase trials by 90%. FC would have been happy with 10%. Actual results as of now are somewhere between those two numbers. Please don't binary search me in that range. The subsequent actions are also up, and we won't know about LTV/churn for quite a while yet, but early indications are good.
I am interested in lesser popular stats like session time and bounces because it sheds some light on conversion differences. If you increase conversion and decrease session time that means you have succeeded in getting out of the way of people who want to give you money. If you increase conversion and increase session time it's for a different reason (maybe you weren't providing enough info before). Same for bounces--is the new higher conversion figure due to enticing people who previously bounced or are you doing a better job at hooking in people who were already more interested? I could go on, but I'm already rambling.
Regardless, great write up. I found it interesting and useful.
In my opinion software homepages should look somewhat like wikipedia pages. Any software that tells me how awesome it is before explaining what it is quickly loses my interest. And using a grid with images in place of a list really annoys me.
I've used fogbugz, can't say I was very impressed with the product. Maybe they need to spend more time on that and the product will market itself. Seriously, this is a great example of the wrong way to view a product.
It's amazing that a one hour video performed best. I get that people want to click on it more, and it IS obvious that putting a big play button would lead to more views... but how far do they get and how does it impact conversion etc?
I saw that same graph on Khan's FAQ today. He was explaining that if there's a lull in the video, then he works to make it more interesting at that point, until the lull goes away.
We're working on a product that will be able to give you this data regardless of number of views. Ping me if you'd like more info or an early access account!
Looking at the heatmap, "Pricing" and "Features" are your most clickable menu items. For some reason you completely ignored it and put "Pricing" all the way to the right, where 10% of the visitors can't even see it.
You don't need a "home" link in the menu, that's what the logo is for.
Your menu should look like this:
Pricing | Features | Plugins | Support | other crap nobody ever sees
Then your most important money-making link-button "Try FogBugs" is below the fold for at least 30% of your visitors.
I think you really need to look at this and understand what it means:
A very quick, no-research, gut reaction I just had to your comment was that the numbers you throw out -- "30% of your visitors" -- assume that the demographics of their visitors are inline with the demographics of the internet as a whole.
They're selling developer tools, that changes things. And all that is part of your suggestion of "understanding what it means."
FogBugz is a complex B2B product fueled by recurring revenue (either subscriptions or maintenance).
It's not diet pills or a self-help book.
Most (all?) companies find that the long-form, "Tim-Ferris-style" landing page converts better for impulse sales. But does that work well for non-consumer products?
My fear would be that if you 2x or 3x the top of the funnel with landing page optimizations, you run the risk of attracting the wrong kind of customers. Customers who initially signup and pay, but don't become long-term repeat customers at the same ratio as those who converted on the old landing page.
I would be curious, Patrick, if you're planning on looking at the lifetime value of these conversions, or if the scope of your project with Fog Creek was limited to lifting the number of initial installs.
(Oh, and like everyone else said: great job, great writeup, and thanks for sharing!)