Can we check the evidence behind the customer slide before we use it to justify the roadmap item? I am not sure the quote supports the conclusion yet.
At the moment the slide says the customer wants it and that the customer would benefit from it. We have only confirmed the second part as a hypothesis.
The customer said the workflow was โinterestingโ and asked whether it was available. I would not describe that as a committed request yet.
We have three quotes and one interview. Two people described the current workflow as slow, but nobody asked us to build this exact solution.
I still want the customer context in the review. Let us show the observed problem, what we inferred, and what we have not validated.
I am uploading the interview notes and the usage data. The recommendation is weaker than the quote, but the evidence is more useful.
I will update the slide so the quote is evidence, not an approval. We can still make the case, but we should name the assumption.
There is a particular kind of product meeting where everyone says โthe customerโ and nobody can name what the customer did last Tuesday.
Our version began at 09:00 with a quarterly-planning deck titled Closer to Customers. Jules had brought a prospect quote. Leon had given it forty-eight points of elegant serif. Somebodyโnobody later admitted whoโhad placed a tiny golden halo above the customer icon.
Maya looked at the slide. The slide looked back with the serene confidence of a decision that had already been made.
โSo,โ Nora said, taking the first sip of an espresso that had correctly predicted the length of the meeting, โwhat changed because of this evidence?โ
The room performed the small choreography of people searching for the owner of a sentence. Jules adjusted the commercial context. Leon inspected a spacing issue. Victor turned a page very slowly, which at executive level can mean either profound attention or a desire not to become the next speaker.
Priya opened the research repository.
The customer had not, in fact, asked for the proposed workflow. One prospect had said the existing process was โinteresting,โ in the tone normally reserved for experimental theatre and unusually shaped hotel furniture. Two customers had requested better auditability. Three had built parallel spreadsheets because approval history was difficult to reconstruct. A fourth had no problem at all and would be irritated by the proposed change.
The slide said: Customers want a simpler, more flexible experience.
It was an impressive sentence. It included every positive product adjective and almost no information.
The customer enters wearing somebody elseโs strategy
Customer-centricity usually becomes theatre through compression, not conspiracy.
A researcher hears twelve complicated accounts. The readout becomes six patterns. The six patterns become three implications. The three implications become one planning bullet. By the time the bullet reaches the executive review, it has lost its segment, situation, contradiction, sample and uncertainty. What remains is a sentence polished enough to approve a budget.
Compression is necessary. Nobody can replay every interview before making every decision. But useful compression preserves the edges that let another person inspect the claim: who, in what situation, doing what, how often, and with what competing explanation.
โFour finance administrators created parallel spreadsheets because they could not audit approvalsโ is awkward copy. It is also usable evidence. A team can investigate it, define a target behaviour, test an intervention and discover that the original interpretation was wrong.
โCustomers want flexibilityโ is smoother. It can support almost anything, including two opposite decisions before lunch.
Teresa Torres describes continuous discovery as frequent customer contact by the team making product decisions. The operational value is not simply that more interviews occur. The distance between observation and decision becomes shorter. Fewer layers get an opportunity to turn uncertainty into decorative certainty.
Priya explained this without using the phrase โdecorative certainty,โ because she wanted to remain employable. She opened the raw notes and showed the team where the wording had changed.
Observation: administrators cannot reconstruct who approved a change.
Interpretation: they need stronger audit visibility.
Proposed solution: a flexible workflow builder.
Somewhere between the second and third line, a product idea had borrowed the customerโs coat and walked into planning as a fact.
Three acts of customer theatre
The first act is the decorative quote. A vivid sentence opens the presentation, usually beside a large pair of quotation marks. The quote creates emotional permission, but the decision does not depend on it. If the quote disappeared, the roadmap would remain untouched.
The second is executive ventriloquism. A leaderโs preference is retold as customer demand because โthe customer needs thisโ sounds less political than โI believe this is strategically important.โ The preference may be sensible. Disguising it makes disagreement feel anti-customer.
The third is the research shield. Someone says โresearch validated itโ without explaining the question, sample, alternatives or contradictory evidence. Research becomes a badge that ends discussion rather than material that improves it.
All three acts can happen with good intentions. Organisations reward coherent stories. Planning time is short. Confidence travels faster than caveats. The person presenting wants to protect months of work from being reopened by somebody who joined the story on slide thirty-seven.
The problem is not theatricality itself. Every strategy needs a narrative. The problem begins when the narrative conceals which parts are observed, inferred, preferred and still unknown.
Before a customer quote is allowed onto the roadmap
- Name the segment and the moment in their workflow.
- Separate what they did from what the team thinks it means.
- Show the strongest contradiction, not only the cleanest quote.
- State which decision could genuinely change because of this evidence.
Evidence has texture
Nora asked whether they could simply โuse more data,โ a sentence she immediately regretted because Beaโs analytics tab opened at the speed of a summoned spirit.
Data did not settle the matter. It gave the disagreement better furniture.
A support ticket proved that one person encountered something painful enough to report. It did not prove prevalence. A survey could estimate prevalence among respondents, but flatten context. Usage data showed behaviour at scale, but not intent. Interviews provided context, while memory and politeness distorted what people said. Sales calls exposed urgency, but overrepresented prospects currently close enough to be heard.
The answer was not to crown one method. It was to combine evidence whose weaknesses differed.
The team traced the approval problem across support tickets, event data and interviews. They found that customers were not asking for unlimited flexibility. They wanted to understand who had changed what and recover safely when something went wrong. The flexible workflow builder might still help, but it was now competing with smaller, clearer interventions.
Intercomโs product principle โStart with the problemโ is useful because it gives problem definition actual space before solution commitment. This does not mean spending months admiring a problem from increasingly sophisticated workshop templates. It means resisting the expensive habit of committing most of the organisationโs effort while understanding least about what deserves to change.
Maya replaced the customer slogan with an evidence trail. It was less inspiring. It contained a table. Nobody would print it on a tote bag.
It also produced a better decision.
The team kept the broader workflow concept as an option, shipped a smaller audit-history improvement first, and agreed on the behaviour that would justify further investment. Jules could still tell a commercial story. Priya could recognise the customer inside it. Tom could estimate something narrower than โflexibility.โ Leon removed three screens and, in a private act of celebration, deleted the halo layer permanently.
The one-page evidence trail
You do not need a new research bureaucracy. For a meaningful decision, keep a short note with five headings:
- Who and when: the segment and situation represented.
- Observed behaviour: what people did, not only what they requested.
- Interpretation: what the team thinks the behaviour means.
- Counter-evidence: what does not fit the preferred story.
- Decision consequence: what changes because of this evidence.
The last heading is the uncomfortable one. Research that cannot affect a decision may still be valuable learning. It should not be presented as justification for a choice already made.
This note also protects teams from a different failure: treating every customer request as a requirement. โPlease add an export buttonโ contains information. It does not yet reveal whether the person needs portability, auditability, offline work, easier reporting or reassurance that their data is not trapped. Building the noun in the request can satisfy the sentence while missing the job.
By 10:12, the planning review moved on. The customer slide no longer opened with a moral. It opened with a behaviour, a gap in confidence and the next decision. Nobody applauded. Good governance rarely gets applause. It mostly produces fewer exciting misunderstandings six months later.
My take
Customer-centricity should make decisions more inspectable, not more untouchable. When โthe customerโ ends the conversation, the customer has become a prop.
Ask for the smallest concrete evidence: what did someone do, in what context, what happened next, and what would change our interpretation? If the answer is only a quote, keep asking. Quotes are clues. They are not tiny customer-authored strategy documents.
Questions to take back to your team
- When your team says โthe customer,โ can everyone name the segment and situation being discussed?
- Which recent decision changed because of customer evidence rather than merely citing it?
- What counter-evidence disappeared from the last readout to make the story cleaner?
- Are you investigating what people do, or collecting quotes that approve an existing solution?
- If the preferred solution vanished tomorrow, would the evidence still point to the same problem?
Pick one live roadmap item and trace the customer from observation to interpretation to decision. The interesting moment is where the person turns into a slogan. That is usually where the costume department begins.