Shutting down a project: how to navigate the politics

Short answer: shutting down a project is rarely a technical decision. It’s a people decision. The work is managing the sponsor’s ego and reputation, naming the biases that keep everyone committed, having the hard conversations in the right order, and giving people a way to say yes without losing face. Bring an alternative, let evidence decide, and protect the team.

I’ve written before about the signs a transformation project is worth stopping. Knowing a project should stop is the easy half. The hard half is getting a room full of invested, intelligent people to agree, without anyone losing face or the business losing the team.

At a global consumer brand I was the business owner of a US$1M+ digital program that wasn’t going to deliver. I recommended we stop it, proposed an alternative and asked for due diligence before anything was decided. The business adopted the alternative and avoided more than US$1M in capital spend and around US$600K a year in operating cost (the case study is here). The numbers were the easy part to explain. Most of the work was in the conversations.

Here’s what I’ve learned about navigating them.

Why smart people keep failing projects alive

Before you have any conversation, understand what you’re up against. It’s rarely stupidity or bad faith. It’s a handful of very human biases, and everyone in the room has them, including you.

• Sunk cost. “We’ve already spent $800,000.” Money already spent is gone whatever you decide next, but it feels like a reason to keep going.
• Escalation of commitment. The organisational researcher Barry Staw described how decision-makers tend to invest more in a failing course of action they personally chose, to justify the original decision. The more public the commitment, the stronger the pull.
• Optimism bias. Every status report says the next phase will fix it. Teams genuinely believe it, because they’re close to the work and can see the progress, not the gap.
• Loss aversion. Stopping feels like a certain loss. Continuing feels like a chance to avoid one, even when the expected cost of continuing is higher.
• Identity. For a sponsor, a flagship program can become part of how they’re seen. Stopping it can feel like a verdict on them, not the project.

Naming these biases out loud, without pointing at anyone, does a surprising amount of work. “We’re all at risk of sunk-cost thinking here, so let’s look only at the cost and value from today” gives people permission to change their minds.

Map the people before you make the case

A stop recommendation fails more often because of who heard it first than because of the evidence. Before you write a slide, work out:

Person	What they stand to lose	What they need to hear
The sponsor who approved it	Reputation, credibility with the board	That the decision was reasonable at the time, and that they can lead the change of course
The delivery team	Months of work, possibly their roles	That this isn’t a judgement on them, and what happens to them next
The vendor or integrator	Revenue, a reference client	A clear, professional exit and what’s expected of them
Finance	A write-off they’ll have to explain	The cost of continuing compared with stopping, in their terms
The board	Confidence in management	Evidence, an alternative and a plan

Then decide the order. As a rule, the sponsor hears it first, privately, and never in a meeting full of their peers. Nobody should be surprised in a steering committee.

The hard conversations, in order

1. The sponsor, one on one

This is the conversation that decides everything. Go in with three things: the evidence, an alternative and a way for the sponsor to own the decision.

Start with the business need, not the project. “We both want X. I don’t think this program gets us there, and here’s why.” Acknowledge that the original decision made sense with what was known at the time. It usually did. Then offer them the lead role in changing course: a sponsor who announces a better plan looks decisive, while a sponsor whose project was killed by someone else looks exposed.

Expect pushback. Don’t argue every point in the first meeting. Ask what evidence would change their mind, and agree to go and get it.

2. The people who did the work

When the decision is made, tell the delivery team early and directly, before rumours do it for you. Be specific about what changed: the business need moved, the design couldn’t scale, a cheaper option emerged. Credit the work. Most stopped programs leave behind something valuable, such as requirements, data, process maps or a better understanding of the problem. Say what you’re keeping.

Then tell them what happens next for them. Uncertainty about their own roles is what people remember, long after the project is forgotten.

3. The vendor

Be professional and quick. Check the contract for termination terms and notice periods, give a clear date, and agree what’s handed over. Vendors have their own internal politics too, and a clean exit protects the relationship for the next piece of work.

4. The board

By the time it reaches the board, the recommendation should already have the sponsor’s support. Present it as a decision between two options, not a confession: the cost and value of continuing, the cost and value of the alternative, the risks of each, and what you’ll learn. Boards don’t punish well-reasoned stop decisions. They punish surprises.

Let evidence decide, not seniority

The most useful thing I did with that US$1M program was not to argue for stopping it. It was to propose a short, structured due diligence on the alternative, and to agree up front what would count as success.

That changes the politics completely:

• It lowers the stakes. Nobody has to concede today. They only have to agree to look.
• It moves the argument from opinion to evidence. People find it much easier to change their minds in response to new information than in response to a colleague.
• It gives the sponsor a dignified path. “We tested the alternative and it’s better” is a far easier sentence to say than “I was wrong”.

Agree the criteria before the review starts: cost, time to value, risk and fit with the business need. If the alternative doesn’t stack up, you continue with more confidence than before. Either way, the business wins.

Separate the decision from the people

The single biggest mistake in stopping a project is letting it become personal. A few practices help:

• Use neutral language. “The program” and “the design”, not “your project” or “their mistake”.
• Talk about the future, not the past. Spend most of the time on what you’ll do next, not on how you got here.
• Never blame in public. If there are performance issues, deal with them separately and privately.
• Thank people specifically. “The data model you built is what makes the alternative possible” means more than a general thank-you.

Check your own ego too

If you’re the one recommending the stop, you have biases too. You may be keen to be proved right, or attached to the alternative you’ve proposed. Hold your recommendation loosely. Invite people to challenge it, and be the first to change your view if the evidence does.

It’s also worth asking honestly whether you’re the right person to make the call. If you approved the program yourself, or your team has spent a year on it, you may be too close to it. An outside view, from a peer, an adviser or a fractional executive with no stake in past decisions, can say what insiders can’t.

Make it easier next time

The best way to handle the politics of stopping a project is to settle them before it starts. When a board or leadership team approves a significant program, agree:

1. Stop criteria. What would have to be true for us to stop? Write it down.
2. Review points. Fixed dates when the program is assessed against those criteria, not just against its own plan.
3. An independent voice. Someone at each review who isn’t accountable for delivery.
4. Permission to stop. An explicit statement from the CEO and board that stopping a program on good evidence is good management, not failure.

That last point matters more than any process. In organisations where stopping is treated as failure, people hide problems until they can’t. Where it’s treated as good judgement, problems surface while they’re still cheap to fix.

The bottom line

Stopping a project well is one of the most valuable things a leader can do, and one of the least rewarded. Do the people work first: understand the biases, talk to the sponsor privately, bring an alternative, let evidence decide and protect the team.

If you have a program you’re quietly worried about, and the politics are what’s stopping anyone from saying so, I’m happy to take an independent look. I work as a fractional COO or Chief AI & Transformation Officer, and a two-week diagnostic is often the fastest way to get an evidence-based answer. Book a 30-minute call.