There’s a specific, uncomfortable moment that comes up in most holds at some point: the CEO has a growth story they genuinely believe, and the numbers, looked at carefully, don’t actually support it. 

Handling this well is one of the harder skills in the operating partner toolkit, because getting it wrong in either direction, too soft or too blunt, tends to cost you something you need for the rest of the relationship. 

The mistake on the soft end is raising the concern so gently, hedged with enough qualifiers, that the CEO doesn’t actually register it as a serious challenge to their thesis. 

This feels kind in the moment and is genuinely unhelpful, because the CEO walks away from the conversation still believing the story holds, and the actual disagreement resurfaces later, in a worse form, once the numbers have made the gap undeniable to everyone. 

The mistake on the blunt end is presenting the disagreement as a verdict rather than an open question, “the numbers say your story is wrong,” rather than “here’s what I’m seeing in the numbers that I can’t reconcile with the story, help me understand it.” The second framing invites the CEO into genuinely examining the gap with you. The first framing invites defensiveness, because it positions you as having already decided they’re wrong, and very few people update their view constructively from that starting position, however sound the underlying analysis actually is. 

The version that tends to work best treats the conversation as a joint investigation rather than a challenge. Bring the specific numbers, be precise about exactly what doesn’t reconcile, and frame it explicitly as something you want to understand rather than something you’ve already concluded. 

Sometimes the CEO has context that resolves the apparent gap and the growth story was fine all along. Sometimes the conversation reveals the story genuinely doesn’t hold, and having got there together rather than through a delivered verdict, the CEO is far more likely to engage constructively with revising it.