Project scoring is dead. Long live scoring!

For twenty years, demand management has rested on one ritual: scoring. You build a grid. Strategic alignment, estimated ROI, risk, urgency, identified sponsor. Every project goes through the mill. Out comes a score. And in theory, you sort by descending score, you draw the line at your capacity, and there's your portfolio.

In THEORY.

Scoring never decides anything (except when it bites)

The problem isn't the grid. The problem is that when a sponsor takes the time to fill in their request properly, the score is always good. Strategic alignment? Obviously, otherwise they wouldn't be asking. ROI? They've done the math, and it's positive. Urgency? Of course, otherwise they'd wait.

The result: you end up with 40 projects between 7 and 9 out of 10. A delta of 1.1 between the fifteenth and the thirtieth. And you can only do 15 of them. Scoring tells you nothing. It doesn't tell you which one to do first, which to push back, which to break down, which to drop. It just confirms what you already knew: they're all good projects.

Where scoring really works is on disqualification. The projects scoring 2 or 3, the ones where the sponsor didn't even bother to fill in the brief, where there's no quantified objective, no clear scope. Those, scoring eliminates. Fine. But that never accounts for more than 10 or 15% of the backlog. On the remaining 85%, you're still stuck.

There's one case where scoring can also work to qualify: when it's backed by real benefits tracking. At the framing stage, you don't just tick "high ROI" on a grid — you define precise types of benefits (financial, NPS, workload reduction, compliance), each with a target value, a measurement date, and an owner who'll go and check. And three, six months after delivery, you come back to measure whether what was promised actually materialized. That's what we're seeing take shape at several of our clients.

It changes everything. Because when a sponsor knows someone will come and track the benefits they announced, they stop inflating their project's score out of reflex. The score becomes binding — not decorative. And above all, it creates a learning loop: you compare what was promised against what was actually delivered, and that sharpens your ability to assess future projects.

But let's be honest: very few organizations actually run this process today. It takes discipline, a post-project follow-up that no one does naturally, and a culture willing to look at whether the promised benefits materialized — even when the answer is no. So in most cases, scoring remains an intake exercise that ranks 35 projects between 7 and 9, and the executive committee leaves without having decided anything.

What senior management really wants (WITHOUT ASKING FOR IT)

A CEO doesn't want a ranking. They want to understand the consequences of their choices.

What they're really asking, even if they don't always put it that way, is: "If we launch the AI program now, what happens to the ERP program? If we push the website redesign to Q3, do we free up enough bandwidth to fit the regulatory project into Q2? If we do all three, who goes to 160% capacity, and for how long?"

These aren't questions of score. They're questions of scenario. Questions of trade-offs. Questions of sequencing.

Scoring gives you a sorted list. A scenario gives you a conversation. And it's the conversation that produces the decision — not the list.

The executive committee that works well isn't the one that rubber-stamps a fixed top 10. It's the one that looks at three scenarios side by side and says: "We're going with scenario B, but we're watching the Q2 ERP deliverable — if it slips, we switch to scenario C." That's an act of steering. Not a stamp on an Excel spreadsheet.

Why we stayed stuck on scoring for so long

Because scenario planning was inhumanly expensive.

To produce a real capacity scenario, every project has to be broken down into deliverables. Every deliverable has to be assigned to a skill group. Every skill group has to have a known capacity per quarter. And you have to be able to move the pieces of the puzzle around to see what comes out.

On a big transformation program, we know how to do it. You bring in a consulting firm, two senior consultants, four months, 80,000 euros, and out come three scenarios in a 120-slide PowerPoint. The executive committee makes the call. Fine.

But on the other 130 projects in the portfolio — the mid-sized ones, the small ones, the ones that individually look manageable but collectively saturate the teams — no one does this work. It's too long. Too expensive. Too many projects. Not enough PMO.

So we do scoring. Because scoring scales. You fill in a form, you get a number. It's fast. It's clean. And it's useless for choosing between 35 good projects.

We were stuck in a trade-off: the only method that really worked — scenario planning — didn't scale. And the only method that scaled — scoring — decided nothing. So we stayed with scoring, hoping that adding one more criterion would eventually make it work.

Demand management becomes scenario management

What's changing today isn't that we have better scoring. It's that we can finally do scenario planning at scale.

Every project in the portfolio — even a request that landed this morning as a three-line email — can be briefed, broken down into deliverables by team, and estimated in effort within minutes. Not in four months. Not by a consulting firm. By a chain of AI agents that knows your organization, your skill groups, your estimation baselines, your delivery history.

What used to be reserved for two-million-euro programs becomes feasible on every line of the backlog.

And that changes the very nature of the conversation about demand. You no longer ask "does this project deserve to be done?" — the answer is almost always yes. You ask "under what conditions can we do it, and when?"

A business unit shows up with a need. In five minutes, you know it takes 15 days of the data team and 8 days of integration. You drop that onto the quarter's capacity view. And you see immediately: it fits if we push deliverable 3 of the supply chain project. Or it fits in Q3 without moving anything. Or it doesn't fit at all before September because the data team is locked up by the regulatory program.

That's an answer. Not a score. Not a "we'll put it in the backlog." Not a "we'll look into it." An answer with a scenario, options, and a visible opportunity cost.

And when you put three scenarios in front of an executive committee — "here's what we can do if we keep the current plan, here's what changes if we add the new program, here's the trade-off if we break it down differently" — the conversation is radically different from "here's the top 40 sorted by score."

The collaborative promise scoring never kept (and the two things that really work)

Scoring was never sold as an authoritarian tool. Quite the opposite. It was the "objective" and "collective" answer to the prioritization problem. You build the grid together in the executive committee. You define the criteria together. You weight them together. And everyone scores their projects within the same framework. On paper, it's a model of shared governance.

Except the result decides nothing. Everyone played along, everyone filled in the grid, and you end up with 35 projects between 7 and 9. The collaborative formula produced a soft consensus. No one cheated — it's just that the mechanism can't discriminate between good projects. And so you leave the executive committee without having decided, telling yourselves you'll "refine the criteria next time."

What we see working at our clients is two very different things — and they complement each other.

The first is scenario planning. It's no longer "everyone scores on their own and we compare the numbers." It's "we look together at what happens if we do A rather than B this quarter." The same numbers, the same constraints, the same trade-offs — in front of everyone, in real time. A business director who sees that their project pushes back a peer's doesn't react the same way as when they see a score of 7.2 versus 7.8. A score is abstract — you can dispute the weighting, the criterion, the mark. A scenario is factual: if your project comes in at Q2, here's the team going to 150%, here's the deliverable that slips. We're no longer in a debate about method. We're in a concrete choice.

The second is benefits tracking. And here's where it makes scoring genuinely useful. The principle: at project framing, you no longer tick "high ROI" on an abstract grid. You commit to specific benefits — type (financial, NPS, workload reduction, compliance), target value, measurement date, and a tracking owner. And three, six months after delivery, someone comes back to check whether they've materialized. That completely changes the dynamic. A sponsor who knows their promises will be measured stops inflating their project's score out of reflex. The score no longer decorates — it commits. And above all, after a few quarters, you accumulate real data on the gap between promised benefits and delivered benefits. You start to know which types of projects keep their promises and which ones systematically bluff. Scoring becomes a learning tool, not just an intake filter.

The two together are formidable. Scenario planning tells you when and how to fit a project in. Benefits tracking tells you whether it was worth doing — and it makes the next decision cycle better than the last.

And paradoxically, you end up getting more projects through. Because when you see capacity and dependencies clearly, you find smart sequencing. You trim a scope. You push a non-critical deliverable by three weeks to free up a team. You turn a "no" into a "yes, but in Q3" or a "yes, if we cut down batch 2." And because you're tracking the benefits, you know which projects are worth fighting to get through.

That's what we're discovering with our clients right now, and honestly, it's wild. Organizations that had been stuck for years in the same priority war are starting to move — not because they have more resources, but because they can finally see the same things at the same time. And because they learn from what they've delivered, not just from what they promised.

The problem was never sorting out the bad projects. It's choosing among the good ones. And for that, you have to stop scoring into the void — and start scenario planning and tracking.

This article first appeared in my LinkedIn newsletter Il était une fois un CODIR, where every two weeks I share real situations from the boardroom.