Stop building the loudest request. Build the right one.

Every feature board has the same problem: the loudest users win. A small number of vocal customers spam your upvote widget, and suddenly you are building for them instead of for the cohort that actually drives revenue.

Votes are a signal — not a decision. Here is how we built a scoring layer on top of raw vote counts that actually predicts impact.

The three signals that matter

We score every request against three dimensions before it touches the roadmap:

• Demand: How many unique accounts have requested or voted for this?
• Pain: How hard is the workaround? Are users churning over this?
• Willingness to pay: Are the requesting accounts on high-tier plans, or are they all free users?

Weighting by revenue exposure

We weight each vote by the MRR of the account behind it. A feature request from a €500/mo customer counts for more than ten requests from free-tier users. This sounds harsh — it is actually honest. You are building a business, not a democracy.

The loudest users are rarely your most valuable users. Weight by revenue and the roadmap reorders itself.

When to override the matrix

Sometimes the right call is to build something the matrix ranks low because it unblocks a whole category of users you want to acquire. Strategic bets require human judgment. The matrix tells you what is safe. Occasionally you want to make the less safe choice for good reasons.

Document the override. Future you will thank present you when you are trying to explain the decision a year later.