Erik Wolf7 min read
Someone at a talk last year asked me the question people assume I’d rather not answer.
“Do I actually need you for this? Or can I just call the people who told us no?”
You can just call them. You should. There’s a chapter in my book about exactly how, and I’ve spent more time teaching people to run this themselves than I’ve spent selling it — because a business that talks to its customers badly is still miles ahead of one that doesn’t talk to them at all.
So here’s how to do it. Then I’ll tell you honestly which parts are hard, because some of them are.
1. Aim for eight to twelve conversations — and plan to ask a lot more people
Go back two quarters, take every deal that closed, and set a target of eight to twelve interviews.
Now the part I watch happen almost every time. Companies talk to the customers who chose them. Those calls are pleasant — the person is glad to hear from you, the conversation is warm, and everyone comes away feeling good about it. Then the list of people who said no sits there and nobody calls it, because ringing up someone who rejected you feels awkward and there’s always something more urgent to do.
That’s the half with the information in it.
The buyer who chose you can tell you what tipped it, and that’s worth knowing. The buyer who didn’t can tell you what’s broken — and they’ll usually say it plainly, because they have no relationship to protect and no reason to be careful with you. The discomfort and the value point in the same direction. The call you least want to make is the one worth making.
Include the deals that went quiet, too. A deal that faded without a decision is telling you something specific, and it usually isn’t price.
One more thing nobody warns you about: to land ten interviews you’ll need to approach twenty, thirty, sometimes more. People are busy, some never reply, some agree and don’t show. Budget for that up front or you’ll get four conversations, decide it isn’t working, and stop.
2. The rep who lost the deal cannot make the call
This matters more than everything else on this list.
It isn’t about trust. It’s that the buyer will keep being polite. They rejected this person a few weeks ago; they aren’t going to open up about why the proposal felt thin. They’ll say something kind and vague, your rep will write “price” in the notes, and you’ll have spent the effort and learned nothing.
Use anyone else. Someone from operations, someone from finance, your executive assistant if they’re curious and good with people. Distance from the deal matters more than research experience.
3. Ask for fifteen minutes, and don’t call it feedback
“Can I get your feedback?” sounds like a survey and gets survey answers.
Try: “You looked at us earlier this year and went a different direction. I’m trying to understand our own business better and I’d rather hear it straight than guess. Fifteen minutes, and I’m not going to pitch you anything.”
Then don’t pitch them anything. Word gets around.
4. Do it on video, and record it
Run these on Zoom or Teams — somewhere recording is one click and the file lands somewhere you can find it. Ask permission at the top; almost nobody says no.
Avoid the ordinary phone call. Recording is awkward, quality is worse, and you lose the face, which is where half the signal lives. Someone saying “no, it was fine” while their expression says otherwise is a thing you want to have seen.
And record it even if you’re sure you’ll remember. You won’t. More importantly, the exact words matter — “we needed someone who’d been through an FDA audit” is usable; your paraphrase three days later is not.
5. Write your questions, then arrange them into a conversation
Write the questions out. Don’t wing it, and don’t hand someone a blank page and hope.
But the order is the craft. Open somewhere easy and factual — how the search started, who was involved. Get the decision story before you get anywhere near your own performance, because once someone starts assessing you they start being kind. Save the pointed question for after they’ve warmed up. Close somewhere generous.
Then teach whoever’s running it to hold the list loosely:
- When to go deeper. If an answer surprises you, stop and stay there. The question you planned will keep.
- When to skip. If they’ve already answered something further down the list, don’t ask it again. Nothing kills a conversation faster than being asked something you just answered.
- When to move. If you’re at minute twelve and haven’t got to why they chose the other vendor, leave the texture behind and get there.
A list of questions is a safety net, not a script. Read it out in order and you’ve built a survey with a human voice attached.
6. Ask why twice, then chase the moment
The first answer is the stated reason. The second is usually the real one.
When someone says “it came down to price,” don’t accept it and don’t argue. Ask what the other vendor would have had to get wrong for price not to have mattered. That question produces a pause, and then something true.
Then get concrete. “They were more responsive” is a summary you can’t act on. “I called Tuesday and heard back Thursday” is something you can fix on Monday.
7. Write down what you believed before you start
Before the first call, get your sales team’s theory on paper. Why do they think you lose? Why do they think you win?
Seal it. Compare it at the end.
That gap is the whole exercise. It’s also the only version of this that changes anyone’s mind — a finding lands differently when it’s sitting next to the confident prediction it just contradicted.
8. Read across the interviews, not inside them
The temptation after ten conversations is to fixate on the most dramatic one. Resist it. One furious buyer is a story; three buyers mentioning your quote turnaround in passing is a finding.
Read them together, and pay particular attention to anything that shows up in your losses and never in your wins. That’s your gap.
The hard parts
I said I’d be straight about this.
Knowing what this kind of research is for. Ten conversations tell you why, in your buyers’ own words, with enough specificity to act on. They don’t tell you how many. Those are different jobs, and the most common mistake is running a small qualitative study and then reporting it as though it were a survey — “sixty percent of buyers said.” Ten people didn’t say sixty percent of anything. Use this to understand the reasons; use a survey if you need to size them.
It stops. Almost everyone who runs a DIY win/loss study runs exactly one. It’s useful, everyone agrees it should be regular, and then Q3 happens. A single snapshot is worth having. The compounding value only shows up when you can see the picture change.
Consistency. You’ll interview better on call nine than on call one — which sounds good and isn’t, because it means your last three conversations aren’t comparable to your first three. Consistency across interviews is unglamorous and it’s most of what makes the analysis trustworthy.
Hunting for what you expect. This is the one that decides whether the exercise is worth anything. Done properly, this research almost always turns up something a leadership team genuinely did not know about their own business — that’s the entire point, and it’s the part that pays for everything. Done while quietly looking for confirmation, it returns your own assumptions with citations. Ten transcripts, a theory already in your head, and no outside perspective is a hard place to be surprised. Go in trying to be wrong.
The lesson
The worst version of this research is still better than the confident story your company is currently telling itself.
Start with ten calls. Have somebody else make them. Ask why twice. Go in expecting to be surprised.
And if you’d rather hand it over: we’ll run it faster, and the analysis goes deeper than any of us can get reading ten transcripts by hand. A full study is three to five weeks and starts at $5,000. If you’d rather it ran continuously instead of once, we do that too.