Beyond RICE: Strategic Prioritisation That Actually Works

Why RICE breaks hardest at planning time

Annual product planning is where this gets expensive. You have 30 or 40 initiatives. You run them all through the scoring model. You get a ranked list. Two people adjust their effort estimates and the list reshuffles. Someone senior questions the reach assumption on item three and suddenly everything moves.

The problem isn't that people estimated wrong. The problem is that you're asking a formula to absorb decisions that require human judgment and strategic clarity. The list feels like output. It's actually just noise with numbers attached.

What the list doesn't show: which of these initiatives maps to the company bet that matters most right now? Which ones are we doing because someone asked loudly enough? Which ones are dependencies in disguise? Scoring doesn't answer any of those questions. It just sorts.

From scoring to structured reasoning

The teams I've seen make genuinely better roadmap decisions aren't using better formulas. They've changed the unit of analysis.

Teresa Torres has written about Opportunity Solution Trees as a way to structure product thinking, and the core idea is useful here. Instead of asking "which feature scores highest?", you ask "which opportunity is most critical to the outcome we're pursuing?" The starting point is the outcome, not the backlog. Every candidate initiative has to connect to a specific opportunity before it's even eligible for prioritisation.

That shift sounds small. It isn't. When you structure it this way, a lot of initiatives reveal themselves as answers to questions nobody actually asked. They exist because they were requested, not because they serve an outcome. The tree makes that visible.

This connects directly to how OKRs should work in practice, something I've written about before in the context of why OKRs fail. The same gap shows up: when outcomes aren't concrete enough, any activity looks like progress.

Bets, not features

The other change that makes a real difference: stop prioritising features. Start prioritising bets.

A bet is time-boxed, hypothesis-driven, and has explicit success criteria before you start. It has a review gate, a point in time where you evaluate whether the hypothesis held and decide what to do next. That's a different kind of commitment than "we'll build this feature in Q2".

When you're running a planning process around bets, the conversation changes. Instead of "should we build feature A or feature B?", you're asking "do we believe this investment will move this metric, and how confident are we in that belief?". That's a harder conversation. It's also the right one.

The format forces trade-off thinking. If you're placing three bets this quarter, you're explicitly choosing not to place six others. That choice is visible. You can defend it. You can revisit it at the review gate.

How to run a prioritisation process that produces decisions

Start with the strategic question. Not "what should we build?" but "what is the one outcome that would make this quarter genuinely successful?" If that question doesn't have a clear answer, stop. The prioritisation problem is actually a strategy problem in disguise, and no framework will fix it.

Once the outcome is clear, map candidate initiatives to it. Anything that doesn't connect to the outcome goes to a separate list. It's not killed, it's just not on the table for this cycle. Then have the trade-off conversation explicitly: given our outcome, which of these bets do we believe in most, and why?

The formula doesn't have that conversation. You do.

RICE is a useful starting point for very early-stage triage. As a primary prioritisation method for a product organisation that has a strategy, it's not enough. Stop trying to improve the formula. Start making the strategic reasoning explicit, and then trust it.