BendTech

Mint.com is a personal finance management website.   Last month they were acquired by Intuit for $170M.  This month their CEO, Aaron Patzer, gave a presentation called “Accounting for Startups“, in which he describes the basic blueprint for a startup, backed by hard data from the mint.com experience.  He discussses the various stages to expect, what the goals and finances should be at each stage, and what the tangible and intangible assumptions are.  This is pure gold for aspiring entrepreneurs because it provides real, concrete data against which you can compare your own efforts.

Are you paying your first engineers more or less than Aaron did?  Are your per-user revenue projections comparable? Are you pitching your company using a similar business plan?

There is something in this for everyone.  I’ve been fortunate to have been through more than one startup, and a lot of what Aaron says is right on the mark.  Not only that, it’s in depth enough to get even a reasonably seasoned entrepreneur thinking about some of the assumptions they take for granted.  For example, here are a few of the takeaways I had while watching the video …

Interesting #’s for what percentage of the company you should expect to give a CEO, a VP of engineering, cofounder, “first hires”, etc.  And at which point you should expect to hire those people.

I really like his job titles for “Business Generalist” and “Front End Guy”.  They perfectly describe real people very early stage startups have to have, and that are never adequately captured by more traditional titles.

The two bullets that made the biggest impact on me, that I’ll be continuing to mull over in the months ahead, are these:

Your absolute revenue projections will be bullshit.

Per use/Per client revenue is more important.

This isn’t news to me, but hearing these points so simply stated gives one pause.  The revenue models I periodically throw together for my latest starry-eyed, startup idea always contain one or two variables that have a ridiculous amount of uncertainty; they could be off by 5%, or 500%.  I simply don’t know.  Multiplied together (and they always are) they result in an error best measured not in “percent”, but in “orders of magnitude”.  Fortunately I’m pretty good at admitting how big a pile of manure it all is but, still, I do love having those numbers to look at.

My revenue in in 2012 could be as much as $5M!  What? Oh yeah, it might just as easily be $5K … but, dude, it could be $5M!!!”

Aaron says in no uncertain terms that this sort of prognostication is pointless, and investors know it. Having a slide in your business plan that says, “$50M/year in three years,” doesn’t convince people to invest, it has them running for the exits (or, if it does convince anyone, they’re probably not the kind of investors you should be taking money from.)  It shows that you’re naive about your ignorance.  Not only will these forecasts be ignored, but you’ll find yourself having to defend them. Worse, you may actually start to believe them and make plans that rely on them being right.  Mr. Patzer recommends that you instead focus on the details of how you will monetize each user.  Let the VC’s and investors use your per-user projections to draw their own conclusions about total market size and revenue.

Speaking of per-user revenue, I found Aaron’s data for mint.com – $30 per user per year – to be a good gut check.  I’ve been using a rule-of-thumb of $30-$50 for free or “freemium” services.  But that is just conjecture based on a couple of articles I read earlier this year.  What makes Aaron’s number especially useful is that he talks a bit about how they achieve that.  That $30 is the result of a robust business model with a lot more opportunities for monetization than most of the hair-brained ideas I have  (which is why they won the 2007 TechCrunch40 event, of course).  So in the future, I’ll go with a more conservative number, probably in the $10-$30/year range.

As you can see, this was an educational video for me.  I hope you’ll find it similarly worth your time.

P.S. TechCrunch has the slide deck from the presentation.