Stop Scoring Your Backlog

Every prioritization framework I’ve ever used was a hedge against an expensive mistake. RICE, scoring, the quarterly roadmap argued over for two days. The whole point was to be sure before you spent a sprint, because the sprint was the thing you couldn’t get back. Deliberation was cheap relative to building, so you deliberated.

That trade just flipped.

Building got cheap. With AI I ship in an afternoon what used to take a week, which means a wrong feature now costs an afternoon. I can rebuild it tomorrow. What I can’t get back is the week I spent scoring a backlog to avoid the wrong feature in the first place. The expensive thing used to be the build. Now the expensive thing is the deliberation.

So I stopped scoring. A scored backlog is a tool for deciding slowly and carefully, and slow and careful is exactly the wrong setting when reversal is free. It feels rigorous. It produces a tidy spreadsheet everyone can point at. And it’s optimizing a constraint that left the building two years ago.

What I keep instead is one question. What’s the goal right now?

Not the mission. Not the five-year vision. The goal for this phase, the one that’s actually true today. Launch by July. First 50 beta users in the room. Profitability before the runway gets tight. I pick the one that’s real and say it out loud, and then every feature sorts itself into must-have, can-wait, or no-go against it. The sort takes a minute, not a planning offsite, because I’m not trying to be certain. I’m trying to be moving.

On one client product the goal is a July launch with real beta users testing it. So the feature people show up for gets all the oxygen, group chat waits, and a photo restyle I’d genuinely enjoy building is a no-go until after launch. On another the call was to launch free in July and turn paid on in August. That one decision pulled three weeks of billing work out of the critical path without losing anything. Neither of those needed a framework. They needed a goal stated plainly and a willingness to decide in the room.

The same filter runs on feedback, which is where it earns its keep. Every stakeholder has a list. Every customer wants the thing that would make their exact case perfect. Every prospect swears there’s one feature standing between you and the deal. You cannot build all of it, and trying is how a product turns to mush. The goal tells you which requests are a yes today and which are a yes later, and it lets you say no to something three people asked for without flinching. Saying no there isn’t neglect. It’s knowing what the phase is for.

Here’s the part that catches good people. Frameworks feel responsible. Scoring feels like diligence. So they survive long past the world that made them worth the time, and you end up applying sprint-era rigor to a decision that costs an afternoon to reverse. That’s not careful. That’s just slow. The diligence moved. It used to live in picking the right thing to build. Now it lives in being honest about what the goal actually is, because that’s the only input the whole system runs on.

When you don’t have the goal, everything looks reasonable. Every idea has a case, every request feels urgent, and you ship a pile of defensible features that don’t add up to the number you needed to hit. The north star was never a poster on the wall. It’s the thing you say no with, fast, while the cost of being wrong is still small enough not to matter.