The guide
How to hear your customers.
Seven rules for turning what customers say into what you build. They need no tool, take about an hour a week, and work for a team of two or two hundred.
Most teams already talk to customers. The trouble is what happens next. What was said reaches the people deciding as a feature request, the least useful form it could take, and then the loudest opinion in the room decides anyway.
These rules fix the part between the call and the decision. Each one says what it costs you, because every rule here makes your numbers smaller and more honest.
1Listen for the struggle, not the request
A feature request is a customer guessing at your job. The useful part is what made them guess.
When someone asks for a filter on the reports page, the request tells you almost nothing. The reason tells you everything: they rebuild the same view by hand every Monday and no longer trust the numbers. So when you hear a request, ask one more question:
- "What were you trying to do when you noticed it was missing?"
- "What do you do today instead?"
- "How often does that happen?"
The answer to the second question is the most valuable thing a customer can tell you. A workaround someone built in a spreadsheet is proof the need is real, because they paid for it with their own time.
What it costs you: your list of requests gets shorter and your list of struggles gets longer. That is the point.
2Quote word for word, never paraphrase
Write down the sentence the customer said, exactly, with who said it and when. Never your summary of it. A paraphrase is the first place your own opinion sneaks in. "Customers want better reporting" could mean anything and convinces nobody. Compare it with this:
Every Monday someone exports it, fixes the columns by hand, and sends it round. If they're off, nobody gets it.
The quote carries a frequency, a workaround and a consequence, and none of it came from you. Keep it in the language it was said in, too: a Swedish sentence translated in a hurry loses the detail that made it worth keeping.
What it costs you: quotes are longer than summaries, and sometimes less flattering to your plan.
3Count organisations, not mentions
The number that matters is how many different organisations have the struggle, not how many times it came up. One customer who mentions something in every call is one customer. These are four different numbers, and most teams blur them into one:
| What you count | Example | What it tells you |
|---|---|---|
| Mentions | 14 | How often it came up. Mostly how chatty your customers are |
| Organisations | 9 | How widespread it is |
| Paying customers | 6 | Whether it touches revenue you already have |
| Organisations you want more of | 4 | Whether it matters to where you are going |
What it costs you: fourteen becomes nine on day one, and the loudest account loses its outsized vote.
4Unprompted beats prompted
Half of a good discovery call is the seller proposing a need and the buyer agreeing to it. Nine calls later you have nine organisations with a need your seller invented, and the better your sales team, the worse it gets. So note who raised it first. Then look at what came back:
- "Yeah, that would be good." Bare agreement. Worth almost nothing.
- "Yes, every Monday I copy these into a spreadsheet." Agreement with a detail only they could know. Strong, even though you asked.
- "No, we don't really need that." Disagreement. The strongest evidence of all, and the one most often left out.
What it costs you: nine often becomes four. That demos worse than nine, and it is the true number.
5Group by what sits underneath
A theme is something several customers struggle with, named from their side, in eight words or fewer: "No cross-team view of work in flight", never "Portfolio dashboard". If the name already describes the solution, you have stopped listening.
- Be suspicious of abstract names. If the only name that covers a group is something like "Visibility", it is probably two themes stuck together.
- Leave the leftovers alone. A sentence that fits nowhere waits until a second organisation says something like it.
- Show the negative space. If a theme has four quotes and no more, say so. What never comes up is information too.
What it costs you: fewer, smaller themes, and a pile of ungrouped sentences you have to look at every week.
6Write a brief with an appetite, not an estimate
When a theme is worth acting on, write one page that whoever builds can read without asking you anything. The first half is evidence, and nobody edits it to fit the plan. The second half is your decision.
| Part | What goes in it |
|---|---|
| The problem, in their words | Three quotes, and one sentence on the pattern behind them |
| Who has it | The organisations, and how many raised it unprompted |
| What we are doing | The change, in a few sentences |
| What we are not doing | The obvious extras, ruled out by name |
| Constraints | What must not change |
| Done looks like | What we would see that tells us it worked |
| Appetite | A day, a week, two weeks or six. A budget, not an estimate |
The appetite is borrowed from Shape Up: how much of ourselves this is worth. If the work does not fit, cut the scope, not the evidence.
What it costs you: you have to write down what you are not doing, which is where most of the arguments are.
7Close the loop with the people who raised it
When it ships, tell every organisation that raised it. You know exactly who they are, because you counted them in rule 3. A short note in their words works best: "In April you told us the Monday report never comes out right. It now does, and here is how."
Then keep listening to that theme. A theme is never done, because it is a fact about your customers. If it keeps collecting new quotes after you shipped, the fix did not land, and that is the most useful thing the whole method can tell you.
What it costs you: a few emails per release, and the occasional honest admission that it did not work.
Doing it by hand, starting this week
All seven rules work with a spreadsheet and an hour every Friday.
- Make one sheet with five columns: quote, organisation, date, who raised it first, theme.
- After each customer conversation, copy in every sentence about their work, a need or your product. Word for word.
- On Friday, give each new row a theme, or leave it blank if nothing fits yet.
- Count organisations per theme, and the unprompted ones separately.
- Bring the top three themes, with one quote each, to whatever meeting decides what gets built.
The loudest opinion can still win that meeting. It just has to win in front of the evidence.