Very average analytics

"...and all the children are above average." - Prairie Home Companion

Venture capital pitch meetings are full of average numbers. This can lead to sloppy thinking.

What's the average order value? What's your customer acquisition cost? Who's your typical customer? What lifetime value are you calculating with? What are your unit economics?

The first thing to note is that these are all close-ended questions. You might as well be asking a spreadsheet. Conversations with founders are more interesting if you ask open questions. (Conversations with VCs who ask open questions are also more interesting. Raise from those.)

But the obsession with averages is more pernicious. Especially if you use them to actually run your company. Averages obscure the distribution. And it is in distributions that you can find a lot of truth about what's working in the business and what isn't.

I have a company in the portfolio that sells to businesses. The potential customer first has a chat with the sales team and then does a paid pilot. In the early days of that company, most pilots didn't convert. For a while, churn was above 70%.

Sales: it's clear the product isn't working. People hate it. Also, the leads are weak.
Marketing: it's clear sales is not explaining the product right. Also, coffee is for closers. 
Product: Look at my roadmap! I can fix this!
Finance: We are in trouble.
CEO: ...give me that spreadsheet again.

As the CEO found out, a good number of people loved the product. Except those customers had a specific profile. They were large employers and had a big and recurring problem that our product solved. These customers did not churn after the pilot. They had much higher order values and bought more often.

We stopped selling to the smaller clients and focused all our efforts on the right customers.

Whenever I see an average, I reflexively think about what the distribution looks like. Are there several businesses hiding in one? Is there stuff that's just not working? What's working phenomenally well? What predicts a happy customer? I like when founders pull a histogram out of their back pocket and start explaining that because they've been thinking about it all along.

One other caveat: at the stage we invest in - Seed and Series A - sample size is usually so low that the mean can wander a lot over time. Interestingly, as your business evolves, the distribution can wander, too.

It took me a few years to always go at least that one question deeper on everything that's presented as "average." Maybe I save you some of those years.

P.S. Another hard-earned lesson: always look at the actual cash-flow. Always. But that's for another post.