Uncertainty weighted impact

The joys of the start-up journey have been neatly summarised a number of times. Some choice examples are “It’s like building a plane while it plummets back to earth” or “It’s like staring into the abyss and eating glass”

Part of that struggle is to understand what to build next, before customers are ripping product out of your hands (also known as product market fit). There is only 1 answer: talk to customers. And really we all know it intellectually, but most folks (myself included) struggle to put it to practice. To be fair, even talking to customers is not as straight forward as it might seem - there is an excellent little book about this - the MOM test.

So then, how do you choose what to do next, conditioned on some customer feedback and having a wide open space of potential directions given your domain?

So, let’s assume a few table stakes are in place:

• You know who the customer is! As specific and niche as you can get, so you can really target their problems. You might be an instance of this customer.
• You know one or more real problems. You actually understand what these people are trying to achieve, to get done or a pain they seek to relieve.

These table stakes should impose some constraints on your future roadmap. As you evaluate that roadmap, a super simple framework might help sharpen how you think about the things you might build next: uncertainty weighted impact.

The intent is straight forward, build an expectation of impact (over time) of what you might build next.

• Impact: attempt to be as explicit as you can about the impact you expect this new thing to have. Who will pay you for the feature? Are there contracts pending on what you will deliver? Who will you make happy because it’s out there? More than faux accuracy, what’s important here is that you posit some expected outcome that is as grounded as it can be. Consider what this impact should be over time.
• Uncertainty: how certain are you about the outcome from building this? Big bets are often quite uncertain - and that’s OK. A feature that your die hard users have explicitly asked you for, a little more certain.

Estimate this value, E(impact) = Impact (over time) * Uncertainty, for each item in your roadmap. Doing this should:

• force you to posit some impact expectations for the things you build - and make it clear when you don’t have them (time to talk to customers again)
• make salient when you are going for low hanging fruit and when you are making a bet

My take is, startups should be switching between these two modes all the time: making existing customers super happy and/or landing deals and making big bets about the future.