Voice of Customer research: how PMMs build messaging that actually converts

"The product was fine."

That's what the PMO director said three weeks after she went dark. She'd sat through the walkthrough, the QBR, and a personalized trial, and the AE had the deal in commit.

Then she explained.

"Every time your team talked about it, they talked about the platform. The integrations. The configurability. What I needed to hear was whether you understood what was happening in my org. My VP is asking me to do a lot more with the same headcount."

She got a perfect demo. She never got one sentence telling her we knew why she was under pressure.

Most B2B messaging is written from the inside out. It starts with the product and works outward toward an assumed buyer, and it reads like it was written by the people who built the thing, because it was.

Buyers understand that copy fine. It just doesn't move them.

We can't write our way out of this. The words we need are already being said, in other people's sentences, on calls most of us don't listen to.

---

Conversation, then patterns

Voice of Customer research is the practice of capturing how buyers describe their own problems and decisions in their own words, and then using those words.

A survey won't get us there. Neither will an NPS score.

VoC is conversation followed by pattern recognition. We talk to wins, losses, churned accounts, and the near-misses who almost bought and didn't.

We listen for the specific language. Then we file it somewhere we can find it under deadline.

It answers a question the roadmap can't. The roadmap says what the product does. It never says why anyone cares, or which words make a buyer feel seen.

---

Where the language lives

Win/loss interviews are the highest-signal source available. If we do nothing else, we do this.

A win/loss interview is a structured conversation about one decision and the moments that shaped it, which makes it a different animal from a success check-in.

Four questions earn their place in a win interview.

Walk me through what was happening in your business right before you started looking. What changed?

Who else were you evaluating? Was there a moment when you felt we understood your problem?

The loss interview is harder and worth more.

Where did we lose you? What did the winner say that we didn't? What could we have shown you that would have changed the outcome?

Most PMMs never ask, because this needs a process. Someone recruits, someone schedules, someone keeps the AE from having a defensive conversation with the buyer first.

Build it anyway. Five wins and five losses a quarter per major segment, and the patterns surface within two or three cycles.

Everything else is supplementary and nearly free.

Support tickets are sentences about what a buyer expected versus what they got, written by someone who has stopped being polite. "I thought this was going to" is where the gap between our promise and their hearing sits.

Public reviews are miniature win/loss interviews. The headline on a five-star review is often the most distilled version of our value proposition, written by somebody who doesn't work here. The complaint in a three-star review is the objection our reps are handling live, or failing to.

Recorded sales calls are VoC research the company already generates and mostly ignores. Four to six a month, mined for buyer language instead of rep performance, will change how we write.

And an advisory board is worth one question above all the others. How did you make the internal business case for us?

That answer is how our customers already sell us inside their own companies. If that language isn't on our site, we're making their job harder.

---

The words before the category

Jobs-to-be-Done, Clayton Christensen's reframe, asks what someone is trying to accomplish rather than what they want. People buy a hole, not a drill. More precisely, they buy a way to hang a picture so the house feels like theirs.

So run the interview along the timeline instead of the feature list.

What was going on before you started looking? Was there a moment you decided you had to do something? What did you search for first, and what were you calling the problem then?

That last one is where the money is.

Buyers rarely use our category words when they start looking. They search for the problem the way they'd describe it to a colleague.

Not "project management software." How to stop the team from missing deadlines.

That's a headline, and nobody inside the building would have written it.

---

The language bank

The bank is the deliverable. Quotes, organized so we can pull from them while writing a homepage, a battle card, or a deck.

Four columns do it.

The goal is specificity. "Customers care about ease of use" is a conclusion, and conclusions flatten exactly the language we went looking for.

Keep the sentence that produced the conclusion. "I'm not going to pay for a tool my team won't use because it takes an hour to learn."

A hundred quotes filed this way is a messaging brief written by our customers.

---

The highlighter test

This is where most VoC work dies. The research happens, the quotes pile up, and the copy doesn't change, because nobody can tell whether the new draft is any better than the old one.

So try this instead. Print the page.

Highlight every phrase that came verbatim out of a buyer's mouth, from an interview or a ticket or a review. Then look at how much yellow is on the page.

Under half, and we wrote it from the inside. That's the highlighter test, and it's the rare messaging review a room full of opinions can't argue with.

Spend the borrowed language in two places first. The trigger goes in demand gen, because it meets people at the week they realized they had the problem.

"After we missed a launch date." "When the board asked for a number we couldn't produce."

The outcome goes in the hero section. Not "real-time visibility into project status." Whatever she actually said when we asked what life looks like now, which was probably closer to "I stopped getting the 10pm messages asking where things stood."

---

The comfortable interview

Two habits undo most of this work.

The first is comfort. We interview happy customers because they're easy to book and they say kind things, and happy customers confirm what we already believe. The buyers who left, and the ones who nearly bought, are the ones who move the messaging.

The second is cleanup. We hear a real sentence and we tidy it. "We were drowning in spreadsheets" gets filed as "manual process inefficiency," and with that one edit the language walks straight back into the building it came from.

Keep the raw sentence. It's the whole asset.

Buyers can't usually explain why one company's copy lands and another's doesn't. They feel it, and the feeling is "these people get it."

The PMO director wasn't holding out for a better demo. She was waiting to hear her own sentence come back to her.

Say it back to her.

---

What to do next

If the copy is accurate, the demo is good, and buyers still go quiet, that's a messaging sprint. We run the interviews, build the bank, and rewrite the pages out of the buyer's sentences instead of the product's.

If that's where you are, start here. The first conversation is free.

---