We moved to a capacity-constrained model. Instead of ranking projects in the abstract, we ask: given our actual delivery capacity, which combination of projects delivers the most strategic value? That reframes the conversation from "my project is important" to "what can we realistically execute well."

The constraint makes the trade-offs visible in a way that scoring alone doesn't. When people see that saying yes to project X means saying no to project Y, the conversations get more honest.

The scoring model gets gamed because the inputs are subjective and the people entering them have skin in the game. We solved this by separating the people who score from the people who benefit.

A small, neutral prioritization panel, finance, strategy, and the PMO, applies the criteria independently. Business units present but don't score. It's not perfect but it reduces the worst of the lobbying.

I'm going to be the contrarian here: scoring models are mostly theatre. In my experience they don't reduce politics, they just give politics a spreadsheet to hide in. The real decisions are made before the session starts, and the scoring reflects those decisions rather than informing them.

That's not a criticism of the people involved; it's how organisations work. The question is whether the model is adding clarity or just adding paperwork to a decision that's already been made.

Joost, I've heard that argument before and I think it's half right. Yes, scoring models get gamed. But the alternative, pure gut feel from whoever shouts loudest, is worse. A flawed model at least forces people to articulate criteria, which creates some accountability even if the scores get massaged.

The problem isn't the model, it's who controls the inputs. Fix that and scoring becomes genuinely useful. Prism PPM have a decent write-up on separating scoring from advocacy that's worth looking at.

We moved away from traditional scoring models towards OKR-linked prioritisation about 18 months ago. Every project bid now has to explicitly state which OKR it contributes to and how. Projects that can't make a clear connection don't get scored; they get sent back.

It hasn't eliminated politics but it has shifted the conversation from "this project is important" to "this project contributes to objective X." That's a more tractable debate.

Bas, how do you handle projects that are operationally necessary but don't neatly map to any strategic OKR? Compliance work, infrastructure maintenance, replacing a system that's falling over. These need doing regardless of which OKR is in fashion this quarter.

We tried an OKR-linked model and ended up with a shadow category called "keep the lights on" that everyone just dumped their non-strategic work into. Killed the clarity we were hoping for.

Thomas, fair challenge. We handle it with a separate "run the business" bucket that's ring-fenced at about 30% of capacity. Nothing in that bucket gets scored, it just gets prioritised against other run-the-business items. The strategic OKR model only applies to the 70% that's genuinely discretionary. Messy? Yes. But it reflects reality better than pretending everything is strategic.

The 70/30 split is practical but I'd be nervous about the ring-fencing calcifying over time. In my last organisation that 30% "operational" bucket crept to 60% over three years as people learned that operational projects had less scrutiny. By the time anyone noticed, the strategic portfolio was effectively a rounding error.

Might be worth building in an annual review of what's in each bucket, with a genuine challenge process for anything that's been sitting in "operational" for more than a year.

I recognise this pattern very well. In my organisation we solved part of the problem by separating the scoring conversation from the resource conversation. When they happen at the same time, every business unit scores their own project high because they know a low score means loosing their budget slot. When you score first and allocate second, the incentive structure changes slightly.

We also introduced what we call a shadow portfolio, which is a list of projects that scored well but have no resource attached to them yet. When something from the active portfolio is descoped or delayed, the shadow portfolio feeds in automatically rather then going through a whole new prioritisation round. This reduced the quarterly horse-trading significantly.

It is not perfect and we still get gaming on the scoring inputs, but it has made the process more predictable, which stakeholders actually value even if they complain about it.