How to Tell a CEO Their Signature Initiative Should Be Shut Down
Twelve million dollars. Eighteen months. Zero revenue-generating products. Two shipped pilots abandoned by business units. Sixty percent staff turnover. The data is sitting in your slide deck, and in twenty minutes you're walking into Robert Townsend's corner office to tell him his innovation lab — the initiative he launched in a keynote speech titled "Banking Beyond Tomorrow," the project he defended to the board three times, the vision he staked his reputation on — needs to be shut down. He hired you to "optimize the innovation pipeline." Not kill it. And your firm has a $1.2M annual retainer with his bank. If you sugarcoat this, your credibility dies when the board finds out. If you deliver it wrong, Robert might kill the engagement before you finish the sentence.
Why This Conversation Goes Wrong
You bury the conclusion in qualifications. "While there are some promising signals in the innovation pipeline, our analysis suggests that certain structural adjustments might be worth considering..." Robert is too smart for hedging. He will hear the doubt in your voice, extract the real conclusion himself, and lose respect for you because you couldn't say it directly.
You present the data without a path forward. "The lab has consumed $12M and produced zero revenue." True. Also useless. Robert didn't pay you to be an expensive mirror. He needs someone who can tell him what's broken AND show him how to fix it without looking like a fool to his board.
You frame it as his failure. Any phrasing that implies Robert made a bad decision — "The original thesis didn't hold up" — triggers defensiveness. He will spend the next 20 minutes defending the lab instead of discussing the future. The lab failed. That's different from saying Robert failed.
You offer only "shut it down" with no face-saving alternative. Robert already suspects the lab isn't working. What he can't do is stand in front of his board and say "I was wrong." If you offer only termination, you're asking him to publicly humiliate himself. He will fight that recommendation regardless of the data.
The Data Shield
When delivering unwelcome findings to powerful executives, the data isn't your weapon — it's your shield. The Data Shield framework positions you as a translator of reality, not a judge of decisions. You don't tell Robert he was wrong. You show him what the numbers say, acknowledge the vision behind the initiative, and offer a path forward that lets him evolve the narrative without abandoning it.
Anchor in the original vision
"The reason you launched this lab was to position the bank for a digital future — that instinct was right, and it still is." Start with what he got right. The market need is real. The competitive pressure is real. The original vision wasn't wrong. The execution vehicle was wrong. This distinction is everything.
Lead with the one number that can't be argued
"In 18 months, the lab spent $12M and generated zero revenue. Turnover hit 60%." Don't present 47 data points. Present the one or two that are incontrovertible. The CEO who argues with 47 numbers finds one weak data point to discredit the whole analysis. The CEO who faces two undeniable numbers shifts to "What do we do about it?"
Acknowledge the personal stakes
"I know this initiative has your name on it. I'm not here to second-guess the decision to launch it — I'm here to give you an honest assessment of where it stands today." This is not flattery. It's recognition that recommendations exist in a political context, and ignoring that context is naive.
Offer the evolution, not the obituary
"Rather than shutting down the lab, consider restructuring it as an embedded innovation function inside your core business units. Same vision, different vehicle." This gives Robert a narrative: "We learned, we evolved, we embedded innovation into the company's DNA." It's not retreat. It's strategic maturation.
Hold your ground when tested
Robert will probe your analysis. "Are we measuring the right things?" "Did we account for the longer innovation cycle in banking?" These are legitimate questions, not attacks. Answer with data, not defensiveness. "We looked at that. Even on extended timelines, the unit economics don't support the current structure." The CEO who sees you hold your ground with evidence gains trust, even when the message is hard.
The moment that changes everything
He already knows it's failing. He needs a way to say it out loud.
Robert Townsend is not surprised by your findings. He's seen the turnover numbers. He knows the pilots were abandoned. He watched his best people leave the lab within six months. What Robert needs is not information — he needs permission and a vehicle. Permission to admit the lab isn't working without it being a referendum on his leadership. And a vehicle — a restructuring narrative that lets him stand in front of the board and say "We're evolving our approach" rather than "I was wrong." The consultant who walks in with "shut it down" creates an adversary. The consultant who walks in with "here's how we evolve your vision" creates a partner. Robert's deepest fear isn't the data. It's that at 62, staking his reputation on this initiative, there's no second act if it fails publicly.
What to Say (and What Not To)
Instead of
"Our analysis suggests the innovation lab has significant challenges..."
Try this
"In 18 months, the lab spent $12M and generated zero revenue. I owe you an honest read on that."
Instead of
"The initiative failed to meet its objectives."
Try this
"The vision behind the lab was right. The execution vehicle needs to change."
Instead of
"We recommend shutting down the innovation lab."
Try this
"Rather than shutting it down, consider restructuring it as an embedded innovation function."
Instead of
"With all due respect, the numbers speak for themselves."
Try this
"I know this has your name on it. That's exactly why I'm giving you the unvarnished version."
Instead of
"We should discuss the implications for the budget."
Try this
"The $8M gap in your core banking modernization? This restructuring frees up the runway for that."
The Bigger Picture
McKinsey's 2024 study on corporate innovation found that 72% of standalone innovation labs launched by large enterprises are either shut down or restructured within 3 years. The failure mode is almost always the same: labs operate in isolation, disconnected from core business units that refuse to adopt their output. The companies that succeed embed innovation inside existing teams rather than creating separate entities.
A Bain & Company survey of 300 CEOs revealed that the #1 quality they value in their advisors is "willingness to deliver uncomfortable truths with viable alternatives." Not intelligence. Not industry expertise. Honest counsel paired with constructive solutions. The consultant who only diagnoses is a physician who never treats. The one who diagnoses and prescribes earns a lifetime relationship.
The political dimension of consulting is not cynicism — it's professionalism. A recommendation that is technically correct but politically impossible is not a good recommendation. Robert will not stand in front of his board and say "My consultant told me I was wrong." He will stand up and say "Based on our analysis, we're evolving our innovation strategy to embed it directly in business units." Same outcome. Different story. The story is what makes it executable.
Practice This Conversation
12 minutes · AI voice roleplay with Robert Townsend
Reading about this is step one. Practicing it changes everything. Sonitura lets you rehearse this exact conversation with Robert Townsend, a realistic AI ceo of a regional banking group who reacts to your words in real time. It takes 12 minutes. The next time you face a client who built something that isn't working, you'll know how to save the vision while fixing the execution.
Practice This Scenario Free →