How many interviews confirm a sales process pattern?
A sales process pattern typically needs to repeat across roughly ten or more unrelated deals — different buyers, different reps, different deal sizes — before it counts as a finding rather than a single deal’s story. One buyer complaining about slow follow-up is an anecdote; the same complaint surfacing independently across a quarter of losses is a pattern.
This isn’t a separate threshold from the interview counts that govern win/loss research generally. A full program typically runs a minimum of 20 interviews, weighted toward losses, before findings of any kind are considered defensible — the same logic applies here, just applied to a specific friction point rather than the program as a whole. Across a full quarter of loss interviews for one program, the same gap surfaced repeatedly: buyers waiting on answers to technical questions that took too long to come back, regardless of which rep or account team was involved. No single deal proved that on its own. Twelve independent instances did.
The reason the threshold sits in double digits, not lower, is that any single deal’s story could plausibly be explained by something specific to that account — a busy week, an unusual internal process, a personality mismatch. Repetition is what rules out those one-off explanations. A friction point that shows up in one interview and nowhere else stays an anecdote no matter how vivid the account is. The same friction point surfacing across buyers who never spoke to each other, worked with different reps, and closed deals of very different sizes is what makes it a process gap worth fixing rather than a story worth retelling.
This threshold matters practically because it changes how findings get reviewed. A program that reads deal summaries one at a time, as each deal closes, will surface the same friction point repeatedly without ever recognizing it as one finding — because each instance reads as specific to its own account until someone looks at a full set of interviews together. Running interviews in a batch, over a bounded window such as a single quarter, rather than continuously and reviewing them one at a time, is what makes the pattern visible in the first place.
There’s also a lower bound worth naming explicitly: a friction point mentioned by only two or three buyers out of a full interview set is still worth noting, but it shouldn’t anchor a strategic conclusion the way a pattern present in a third or more of the set would. The gap between “worth watching” and “worth fixing now” is exactly the gap between an emerging signal and a confirmed pattern.
Treat an early signal as a hypothesis to test in the next batch of interviews, not as a finding to act on yet. If it repeats, it graduates. If it doesn’t, it was never more than one deal’s story.