(It's a baby! Just like I'm a baby... when it comes to running IntroCave. We made it to 6 months, though!)

Six Months, Not Dead

Sometime last week, we passed the 6-month point from when I purchased IntroCave. And the site's not dead! I'll post some more interesting stats in my next post, where I'll have six full calendar months worth of data (I took over in the middle of June). Purchasing an online business is one of the most stressful things I've ever done, but it's turned out to be stressful in many different ways than I thought.

I read an article on a similar deal (CodeTree) a couple of years back on Hacker News and immediately signed up for FE International's deal list. I've always bounced between web development and game development in my day job, but my recent gigs at digital agencies had left me wanting a little more ownership. You tend to build a lot of MVPs and 1.0s in work-for-hire, and I was really missing the fun of operations—analtyics, A/B tests, and optimization are my favorite parts of the product lifecycle.

Over the next year and change, I requested deal sheets on over 30 online businesses. I have no idea if that's low or high for deals of this type, but I'm pretty sure the folks at FE assumed I was just a looky-loo. I started building spreadsheet models on each business that would come through, trying to read between the lines on each prospectus. It became a hobby of sorts, and I'd like to think I got pretty good at that sort of analysis. I came close to putting in offers on a couple of sites, but never quite found the perfect fit. My wishlist, it turned out, was pretty rare: SAAS, ruby backend, in an interesting niche.

After looking at a bunch of SAAS deals, though, I realized the biggest problem with purchasing a SAAS site in the $50k-$100k range is a lack of data. If a hypothetical site sells for around 3x EBITDA, your broadly talking about sites that do $1,500 to $3,000 a month in revenue. That number can go up or down based on age/growth/platform dependencies/etc, but it's a good ballpark. If you're charging $30/month for your service, that's only 50 to 100 subscribers. From a data perspective, that's not much. Somebody with more of a sales background might see that as enough initial proof to dive in and start pounding the pavement to drum up more subscribers. I thrive more on the tech/data side of things, though, so until I starting poking through IntroCave's stats I didn't realize the thing I was really looking for was a volume of incoming data that would support meaningful optimization.

IntroCave had exactly what I wanted... so what was I worried about?

The Pre-IntroCave Worries

Unfamiliar Tech

IntroCave's backend is PHP/Laravel and some .Net. I've never worked professionally with either of those, but I've gotten pretty close. Laravel is the PHP equivalent of Rails, something I've used to operate sites with millions of daily active users. The architecture is super familiar, and PHP isn't THAT bad anymore. On the .Net front, I've shipped multiple games using C# with Unity. I think it's a pretty great language, so it didn't seem like too much of a hurdle to pick up a more enterprise-y app framework. Still, IntroCave's backend is not simple. Rather than a single serverside application with a single database, IntroCave runs multiple databases (one for data, one for queuing) and multiple server types that need to talk to each other (the web server talks to the render server).

When the original owner offered an earnout clause to the sales contract, I jumped at it. I was pretty worried the whole thing would be too complicated to grasp within the 30-day "support" window. The idea of keeping him engaged and interested for 12 months seemed like a pretty good way to de-risk the tech side of the acquisition.

Time to Recoup

The fact that I financed the site and agreed to an earnout makes pegging down an exact time to recoup a little tricky. If revenue stayed perfectly flat and missed all of the earnouts, I was looking at roughly a 32-month repayment time before figuring in interest payments. Call it 4 years with interest. That's on the lower side of multiples I'd seen, but IntroCave didn't have a ton of operational history to go on yet. If I nailed the earnout (which called for a 60% growth rate over the first year), I'd end up shaving a few months off that even with a higher total purchase price. I ran projections. A lot of projections. What if the optimizations I thought would juice the numbers did nothing? What if search algorithms changed? What if traffic dropped 50% per year for three years straight? 75%? I tried to tell myself that the earnout would ensure ongoing support and keep things from dropping off a cliff, but there's just no getting around the fact that 3-4 years is a really really long time on the internet. This worry isn't unique to IntroCave, though, and it's just something you have to get over if you're going to buy an online business.

The Real Stress

Brittleness

Before purchasing IntroCave, I worried that I'd suck at Laravel or not be able to wrap my head around how the render servers operate. The problem now is that I understand how everything works... and I'm not happy with it. The site is pretty well architected, but there is a level of brittleness that is nerve-wracking. Some of these issues are just inherent to how the render servers interact with 3rd-party software, and there's not a whole lot I can do about it in the short term. I'm hoping the backend rewrite will shore up some of that instability, but the site can feel a little bit like Russian Roulette at times. I have to check on it every few hours just to make sure everything is still humming along, and it feels like every time I don't pay close attention I'll randomly wake up to find that the servers have been down for hours and I have 25 emails in the support queue. I don't have a ton of time to spend on the site, so I want every minute of that time to go towards improvements instead of just putting out fires all the time.

Speed

The thing that keeps me up at night is that I'm just not moving very fast. Besides dealing with outages, I spend a considerable amount of time on normal support requests, blog posts, accounting, and writing emails. In my pre-acquisition fantasy, I was just going to spend all my time making cool videos and working on new features. There's a ton of less exciting operational tasks that still need to get done.

The Email. The Email.

I thought email was going to be a low hanging fruit. The site was getting a couple thousand user signups per month and not doing anything with the emails. Surely regular email blasts to those users would drive a lot of traffic and sales! Not only has email not been a super effective sales channel (so far!), the process of writing and sending a monthly newsletter has been way more stressful than I anticipated. Even ignoring the fact that I screwed up the December newsletter, IntroCave is in a pretty weird place when it comes to email. There are about 32,000 emails on my list (up from around 20k when I bought the site). The open rates on my emails are only about 7.5% and they generate around 200 clicks apiece.

I'm really bad at email, it turns out.

I have some ideas on how to improve those numbers. One of the easier ones is switching away from a DIY newsletter solution (I send them transactionally for now) to a real email marketing service that will make it easier to do visual layouts and A/B test the subjects and content. That would also allow me to do some basic best-practice stuff like sending a welcome sequence when new users sign up for an account or a post-sale sequence to make sure users are happy with their videos. I'm currently paying about $10/month for transactional email, and I'm pretty sure I'm getting $10/month in value of out of the newsletter. Most of the marketing sites charge based on the total number of subscribers, so 32k puts me up around $250-$300/month at most of the sites I've evaluated. I'm pretty sure I'm not getting $300/month of value out of email any time soon. Even doing all those best-practice-y things, I don't see improving the email ROI by a factor of 20-30 to justify the expense. The ROI is super low right now compared to the amount of stress and dollars I'm throwing at the problem, but I'm not ready to declare email bankruptcy just yet. It just pushes it to the bottom of the queue—something that should be driving growth and revenue but currently isn't worth investing more time to make better with so many other areas to work on.

How's It Going?

Going back to those pre-purchase jitters, how are things going? The tech stack isn't intimidating anymore. I'm six months in and the site hasn't imploded. There have been lots of bumps and bruises, but these are more related to overloading the current architecture than any bad code I've pushed out. Traffic and revenue have grown. I worry a little about if the site will stay healthy and keep growing for the 3-4 years needed to pay for itself, but way less than I thought I would. Some of my optimizations are working and some aren't. I'm feeling optimistic going into 2019!