Whose feedback should shape the roadmap
A short test for weighing what people tell you, adapted from Superhuman’s approach to product-market fit.
On this page
Feedback never stops at a startup. It comes from customers, prospects, former users, investors, and friends. Most of it is sincere, and much of it points in different directions. Act on all of it, and the product drifts toward whoever spoke last.
So before I weigh what someone says, I weigh who’s saying it. The test I use is adapted from Rahul Vohra’s approach at Superhuman, and at Drafter it’s how we prioritize feedback today.
Start with the problem
The first question is whether this person has the problem our product solves, right now.
If they don’t, their feedback can still be interesting, but it shouldn’t steer the roadmap. Someone who had the problem in the past may have useful context. Someone who never had it is describing a different product.
Then ask how much they’d miss it
For current customers, the next question is Vohra’s. How disappointed would you be if you could no longer use the product?
- Tier 1: somewhat disappointed, and they value the same core thing as your best customers. This is the most useful feedback you’ll get.
- Tier 2: very disappointed. They already rely on the product, and they’re the group you want more of.
- Tier 3: not disappointed, or somewhat disappointed for a different reason. Their requests pull the product somewhere else.
- Tier 4: they have the problem but aren’t customers, or they had it in the past. Worth hearing, and low priority.
Everyone who never had the problem sits below all four.
Why “somewhat” beats “very”
The surprising part is that the most useful feedback comes from people who like the product but don’t love it yet. The people who love it tell you what to protect. The people who almost love it, for the right reason, tell you what’s in the way. Fix that, and they become the people who love it.
How we use it at Drafter
We built this test into the internal tool where our feedback lands. An agent takes notes on our customer calls, and the tool grades each piece of feedback by who it came from, so the right customers shape our priorities.
Blueprints came out of it. Our first workflow drew a complete drawing from the context an engineer gave it. The engineers whose feedback weighed most valued what Drafter did but didn’t rely on it yet, and what held them back was speed. So we built Blueprints, which does 50 to 70% of a drawing much faster and hands the rest to the engineer to finish their way. We had underestimated how much speed matters to engineers. In early testing, they used Drafter more with it.
Where the test stops
The test decides what we build next. Reports that something is broken, wrong, or unsafe skip it. They count no matter who sends them, even someone who has never used the product.
It also favors the customers we already have. Tier 4 is where we’d first hear about a market we’re missing, so we read it, even when we don’t act on it yet.