Failing the Mom Test

For the last couple months, I’ve been including a link to a short survey in the welcome email that follows every TraitLab signup. I figured that by asking every interested person some questions about their background and interests, I could narrow the scope of what I’m building and only build the essential pieces first. I thought it was pretty smart, but in hindsight, it was probably a waste of time. Why?

  • Most new signups don’t open the survey.
  • Several questions in the survey are almost impossible for anyone to answer accurately, e.g., “How much would you pay for the best personality test?”, “How much would you be interested in feature X or Y?”, etc.
  • Many people who complete the survey aren’t potential customers, but they were just signing up to follow along. I can definitely get good feedback from this group, but they don’t have the problem that TraitLab is trying to solve.

Just when my survey failure started to sink in, I heard an interview with Rob Fitzpatrick about customer conversations and his book, The Mom Test. It sounded like the perfect thing to read at the moment, so I immediately bought the book and burned through it. The so-called “Mom Test” is a set of rules for asking questions that “even your Mom couldn’t lie to you about”, and my hypothetical would you ever? questions were clearly failing the test.

Stepping back and reading The Mom Test gave me some perspsective on what I was trying to accomplish with the survey in the first place: avoiding building something that nobody wants. That approach makes sense when there’s a lot of market risk (i.e., Is there a market for the product? Do enough people actually care about this problem?), I’m not sure that’s where the risk is with TraitLab.

There’s already an existing market for similar products, and there are thousands of people searching for and buying online personality tests every day. Because there’s already a market, there’s not as much to learn from customer conversations, compared to the case of a brand new, one-of-a-kind SaSS that solves a problem I’m not sure anyone actually has.

The bigger risks are in the product itself (Can I build what I’m envisioning?) and in the marketing (Can I reach people when there’s already few dominant products in this space?). To learn about those, I need to just start building the damn thing. So I’m dropping the survey and focusing on grinding out the first version.

Back to the code!