A UVP Sprint session starts by naming four roles out loud before anyone starts talking about the business.
One person facilitates. That person doesn’t decide what the company is; they rule on whether an answer holds up. If someone claims a differentiator, they get asked how a customer would see it, taste it, or check it for themselves. A claim that survives that question stays. A claim that doesn’t gets set aside, out loud, and the group keeps going. Nobody wins a weak claim by repeating it louder.
The CEO is the only person who makes the calls on the business itself. Everyone else can argue and push back all day, but when evidence alone cannot settle the choice, that decision belongs to the CEO, and it gets asked for directly: here’s the choice, which one do you pick.
The rest of the leadership team argues their part of the business, and they’re also the evidence for it. Outside research matters, but it does not replace what the team knows firsthand. A claim about the sales process gets tested against what the head of sales actually knows, not against a market report nobody there wrote.
A fourth role manages the mechanics: the clock, the record, who speaks next.
All four roles get set at the very start, and everyone is asked, out loud, whether they agree to work this way. That agreement makes it possible to spend the rest of the day on the business instead of on who gets to talk. The hard part happens once, before anyone has said anything substantive yet.
From there, the day uses one repeating move. Someone makes a claim. It gets compressed down to what it’s really saying. Then it gets pushed: how would you show that, what’s the simplest version of it, how would a stranger know it’s true. If the answer holds, it stays. If it doesn’t, the facilitator does not force it through. There’s a specific point where the facilitator stops pushing and says plainly that a claim didn’t prove out today. That’s a real outcome, not a failure of the day. A UVP built on five true things beats one built on six, where the sixth is decoration.
One thing is off the table entirely: naming, taglines, branding, identity. Those disciplines matter. They use a different standard of proof. Mixing them into this session makes it harder to tell whether a claim was proven or merely liked. If the conversation drifts there, it gets redirected back, every time. The process stays credible by staying inside its own boundary.
By the end of the day, the company has a position every person was actually asked about, on the record, in some form. A UVP that sounds unanimous only because someone stayed quiet isn’t one, so nobody gets to stay quiet.