Big rocks first, then the pebbles:

How to make prioritisation easier when your executive committee has a new idea every day

Everyone knows the jar metaphor. If you pour in the sand first, the big rocks no longer fit. If you place the big rocks first, the sand slips into the gaps.

In project portfolio management, the mechanics are exactly the same - except almost no one actually applies them. Because the problem is never the bad ideas. Bad ideas, everyone knows how to filter out. The problem is the good ideas - the ones that land in March when the plan was signed off in January, championed by a CEO or a shareholder who met someone, read an article, saw a competitor make a move.

The idea is relevant. Often even urgent. And it lands with an implicit label: "do it right now". Except that "right now" means: pushing back what we promised three months ago. Clawing back bandwidth from teams already committed. Breaking a sequencing that had been thought through, arbitrated and signed off.

The CIO or the PMO ends up in an impossible position: saying no to someone with authority, or saying yes and knowing that two other projects will quietly slip. Most people choose the second option. And six months later, no one understands why the transformation programme is a year behind.

The real problem: we never show what it costs. Not in euros - in capacity. In projects that won't get done. In promises we're going to break.

Here's how we do it in practice.

1 - Place the big rocks across the year

We start with the strategic programmes, the ones that come from the transformation roadmap, signed off by the executive committee, committed to over 12 to 18 months. We break them down into deliverables per quarter and assign them to the company's skill groups: development, data, integration, architecture, business steering.

Then we map that onto an annual view, team by team. Each quarter, we see the workload these big programmes consume. And above all, we see the delta what's left as available capacity once the rocks are in place.

That delta is the truth. It's what says: this quarter, there's 20% of bandwidth left on the data team. Or 5%. Or zero. As long as you haven't laid it out visually, everyone thinks there's room. Once you show it, the conversation changes.

2 - Fit the pebbles into what's left

Once the big rocks are ring-fenced, we take the business requests, the shorter projects, less strategic but important to the departments. Each business unit has its top 1. Sometimes its top 3.

We place them on the capacity view. And then we look: do all the top 1s fit into the remaining delta? Can we only do half of them? Does the sales department's top 1 compete with supply chain's on the same team, in the same quarter?

That's the powerful moment. We no longer say "your project isn't a priority" we show that both projects want the same team at the same time, and that a choice has to be made. It's no longer the CIO arbitrating alone in their corner. It's the executive committee that sees the jar, sees that the rocks are already in it, and decides together which pebbles can still fit.

The business units are no longer in "I'm pushing my project" mode they're in "I'm defending my top 1 against the others" mode. It's a completely different dynamic. And a much healthier one.

3 - At the end of each quarter, we re-lay the big rocks

The plan isn't set in stone. At the end of each quarter, we put the big rocks back on the table. The ERP programme has fallen behind three unfinished deliverables are spilling over into Q3. The data workstream went faster than expected capacity frees up.

We update the programme plan, recalculate what's left to do, reassign it across the following quarters. And mechanically, that changes the game for the pebbles. The one we'd signed off in Q2 may no longer fit in Q3, because a strategic programme slipped and is taking back the bandwidth.

It's hard to hear for the business unit that was waiting for its project. But it's defensible because the rule has been clear from the start: the big rocks are ring-fenced. When they move, it's the pebbles that adapt, not the other way around. And we can show it. We can point to exactly what slipped, why, and what it implies. No surprises. Traceability.

4 - The executive committee's new idea: 10 minutes to show the impact

This is where it all plays out. A shareholder or a CEO shows up with a new strategic priority. A programme that didn't exist two weeks ago. Classically, it takes weeks to assess it, cost it out, understand what it implies. Meanwhile, impatience builds, and often the teams start working on it "while waiting for the scoping" the worst-case scenario.

With AI, we change the game. In ten minutes per project, we generate a structured brief, break it down into deliverables tailored to the company's teams, and produce a workload estimate. Not a perfect costing a rough order of magnitude, enough to make a decision.

And above all: we put it on the capacity view. Immediately. We show what this new programme does to the big rocks already in place. "If we launch this in Q2, here's what happens to the digital transformation programme. Here are the two projects that have to be pushed back. Here are the teams that go to 140%."

That's what changes the conversation with an executive committee. We don't say "it's complicated" or "we'll look into it". We show, in real time, the price of the decision not in budget, but in impact on what they themselves signed off three months earlier. And oddly enough, when a CEO sees that their new idea pushes back their own flagship programme by a quarter, they put things into perspective.

5 - Every small project is scoped in a few minutes, not stacked in a backlog

The last change happens upstream. Traditionally, small projects pile up in a backlog. No one really sizes them. We vaguely accept them, stack them up, and one day someone realises there are 47 of them waiting and no one knows which are feasible this quarter.

With AI, every small project is briefed and estimated in a few minutes. Not in six weeks when the PMO has the time now. The brief is structured, the breakdown by team is done, the workload estimate exists.

What changes: when a business unit raises a need, we know immediately whether it fits into the available delta or not. We no longer say "we'll put it in the backlog" that phrase that means "we'll never talk about it again". We say "this project takes 15 days of the data team, we have 12 days left this quarter, so either we push back X, or we schedule it for Q3". That's an answer. Not a flat rejection. Not a vague promise. An answer with numbers, in five minutes.

What this changes for the CIO

The CIO who operates like this is no longer the one who says "we're doing our best" or "we'll see". They're the one who shows a screen and says: "here's what we're carrying, here's what's left, here's what your request implies".

It's no longer a political negotiation. It's visual management. And paradoxically, it's what makes it possible to say yes more often because when you clearly see what's possible, you find solutions. You sequence differently. You reduce a scope. You push back a non-critical deliverable to make room for an urgent one.

Power isn't about saying no. It's about knowing how to keep packing the jar, continuously.

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