How to Run a Win/Loss Interview (With the Actual Questions to Ask)

Most win/loss interviews either don't happen, or happen so late and so vaguely that they turn into a polite compliment session. "Congrats, you guys were great to work with" is not data. It's the answer a prospect gives when nobody asked a real question. The interview itself isn't hard, it's short, but almost everyone runs it wrong in the same three ways: too late, too vague, and asked by the wrong person.

When to call

Call within a week of the decision, not a month later. At a week, the prospect still remembers the specific moment the decision tipped, what the internal debate actually was, and what the competing vendor said in the final call. At a month, all of that has compressed into a vague "it just felt like a better fit," which is true but useless. If the deal was lost, call before the resentment fades but after the immediate frustration does, usually three to five business days out.

Who should make the call

Not the rep who owned the deal. A prospect will soften real feedback to someone they've been negotiating with for months, especially if the loss meant that rep's pipeline number went down. A neutral third party, a founder, a CS lead who wasn't on the deal, gets sharper, less filtered answers because there's nothing personal at stake in the conversation.

The five questions that get honest answers

• "Walk me through the moment you decided." Not "why did you choose them," which invites a summary. This gets the actual sequence: what happened, who said what, in what order.
• "What almost changed your mind the other way?" Even in a clean loss, there's usually a moment of doubt. That moment is where your actual leverage was, and where it might be next time.
• "What did they say about us?" Uncomfortable, but this is the only reliable way to hear your own weak points stated by someone who has no reason to be diplomatic about it.
• "If the price, the timing, or whatever the specific blocker was had been different, would the decision have changed?" This separates a real, fixable reason from a rationalization. Sometimes the answer is a clear yes. Often it's "probably not," which tells you the actual reason is somewhere else.
• "What would need to be true for you to reconsider in a year?" Turns a closed loss into a note for next quarter instead of a dead record, and it's the single most commonly skipped question in every win/loss process we've seen.

Questions to avoid

Skip "how did we do overall," it invites politeness, not specifics. Skip "would you recommend us," it's a proxy for satisfaction, not for what actually happened in the decision. And skip stacking more than six or seven questions total, past that point you're not getting a sharper answer, you're getting a tired one.

What to do with the answers afterward

A single interview is an anecdote. The value shows up once you have enough of them to look for a pattern instead of reacting to whichever one you heard most recently. The discipline that actually matters isn't asking great questions once, it's asking the same five every time, consistently enough that six months from now you can say "we lose on implementation timeline in a third of our competitive losses" instead of just remembering the one call that stung the most.