RICE and WSJF Don't Work. And I Used to Swear By Them.

I have sat in rooms where grown professionals debated whether a feature's "Confidence" score should be 80% or 90% and treated the outcome like it settled something. I have colour-coded RICE spreadsheets. I have explained WSJF to stakeholders with a straight face, as if dividing one estimate by another estimate produces a decision rather than a ratio of two guesses.

I believed in it. I genuinely thought the frameworks were protecting me from politics, from HiPPOs, from the chaos of competing opinions. What I didn't see, what took me years to see, is that they weren't replacing the chaos. They were just giving it better formatting.

The Maths Feels Like Safety. It Isn't.

Here's the thing about RICE: Reach × Impact × Confidence ÷ Effort looks rigorous. It has variables. It has division. Your brain pattern-matches it to science.

But every single input is a number someone made up. And not just any someone, the person with the most conviction in the room, or the most seniority, or the most time to prep the spreadsheet before the meeting.

I've seen this exact scenario play out, not once, not twice. A team sits down to score three features. The Reach numbers come from whoever proposed each item. Impact gets rounded up from "probably 1.5." Confidence of 90% is, frankly, vibes. Nobody pushes back because the numbers look clean. The highest score wins. Everyone nods. The spreadsheet closes.

The framework didn't produce a decision. The framework laundered a decision.

A number built on estimates is still an estimate. Dividing it by another estimate doesn't make it data.

WSJF Has the Same Problem, But Fancier Vocabulary

Weighted Shortest Job First comes from SAFe, and I want to be fair to it, the underlying logic is sound. Prioritise things with high cost of delay relative to their duration. That's a reasonable heuristic.

But in practice, Cost of Delay becomes a political arena. Who decides if the cost of delaying a feature is "high"? The person whose roadmap depends on it. User-Business Value, Time Criticality, Risk Reduction, these aren't measurements. They're opinions wearing a Likert scale as a costume.

I've watched teams spend 45 minutes arguing about whether something is an 8 or a 9 on Time Criticality, a scale invented for the purposes of that meeting, never calibrated against reality, forgotten six weeks later when the next planning cycle resets everything anyway.

That's not prioritisation. That's prioritisation theatre.

The Deeper Problem: It Ignores Power

Here's what no framework documentation tells you: organisations are not merit systems. Features don't win because they score highest. They win because the right executive mentioned them in a meeting, or because a key customer threatened to churn, or because the sales team has been pestering the PM for three quarters and everyone just wants the noise to stop.

RICE and WSJF don't neutralise politics. They give politics a place to hide. The VP who wants her feature built doesn't override the score, she influences the inputs. She asks leading questions during the estimation session. She frames the use case in a way that inflates Reach. Nobody calls it manipulation because it doesn't look like manipulation. It looks like collaboration.

The frameworks create the illusion of objectivity while the actual decision-making — the power, the negotiation, the relationship dynamics — happens in the gaps between the cells.

So What Do We Actually Do?

I want to be honest here: I don't have a clean replacement. Anyone selling you a clean replacement should make you suspicious.

But I've started asking different questions. Instead of "what's the score?" I ask: what would have to be true for this to be the right bet? Instead of ranking everything, I ask: what are we willing to not do? Instead of pretending the numbers are neutral, I ask: whose estimate is this, and what do they stand to gain?

Maybe the answer is simpler than a framework, more honest conversation, more explicit trade-offs, more willingness to say "we're choosing this because the CEO wants it" rather than reverse-engineering a score that happens to justify the decision we'd already made.

Prioritisation is a political act. The sooner we stop pretending it's a mathematical one, the sooner we can have the conversations that actually matter.