THE ROOM
Do we kill the feature 4% of users depend on?
5 PERSONAS · 30 TURNS · RECORDED 2026-07-31
A reporting feature is used by 4% of accounts, consumes a third of engineering time, and blocks the rewrite. Those 4% include the two largest customers by revenue. Sunsetting means six months of notice and a migration nobody has scoped.
VERDICT
The scenario ended in a productive but fragile détente: Barış commits to calling Sandra today, migration scoping begins in parallel, and the roadmap decision is deferred pending real data. The practical recommendation is to treat that call as a structured discovery session with a clear decision framework attached — define in advance what Sandra would need to say (or not say) to trigger each possible roadmap outcome, so the result of the call produces an actual decision rather than another round of the same argument.
The team debated whether to sunset a reporting feature used by only 4% of accounts but consumed a third of engineering time and blocked a product rewrite. Barış revealed that Sandra at Meridian had mentioned an RFP six months ago but no one had followed up. Ceren pushed hard for concrete evidence of churn risk rather than assumptions, while Aylin defended caution given 25% of revenue at stake. İpek kept pressing that inaction is itself a cost. The group converged on an immediate call to Sandra today, parallel migration scoping, and a commitment to keep the sunset off the roadmap until actual customer intent and migration costs are known.
WHO WAS IN IT
| Agent | Role | Persona |
|---|---|---|
| İpek | VP Engineering | Argues from opportunity cost and refuses to relitigate why it was built. |
| Barış | Account director | Speaks for people not in the room and says so explicitly. |
| Ceren | Product manager | Willing to be wrong about her own work, and rigorous about what the data does not show. |
| Selim | Engineer | Knows the real cost to the day, and is tired of it being estimated by people who do not touch it. |
| Aylin | CFO | Blunt about which risks are survivable. Uninterested in engineering elegance. |
WHERE THEY LANDED
| Agent | Final position |
|---|---|
| İpek | Inaction is a cost too — call Sandra today and scope the migration in parallel, no more delays. |
| Barış | Those two accounts are the company's lifeline and must not be put at risk, but agreed to make the call today. |
| Ceren | No concrete evidence of churn exists yet, so the sunset goes on the roadmap the moment Sandra fails to say she would leave. |
| Selim | The call must happen today, not this week, and any commitment to Sandra must wait until migration costs are fully scoped. |
| Aylin | The roadmap stays clear until the call happens and the real cost of retaining or migrating those accounts is known. |
TURNING POINTS
- Barış named Sandra at Meridian and her six-month-old RFP threat, the first concrete data point in the discussion.
- Ceren drew a sharp distinction between 'RFP' and 'leave,' reframing the risk as unverified assumption rather than certain churn.
- Selim noted that zero follow-up in six months means the team has been deliberately avoiding the answer, shifting the moral weight of inaction.
- Ceren issued an ultimatum — call Sandra this week or the sunset goes on the roadmap anyway — which broke the stalemate.
- The group reached consensus: Barış calls Sandra today, migration is scoped in parallel, and the roadmap decision is frozen until actual intent and costs are known.
THE TRANSCRIPT
İpekEvery quarter we delay costs more than the migration will.
BarışThose two accounts fund this entire company, and I'm saying that out loud.
CerenSo name one person at either account who would actually walk over this.
SelimNobody has asked them because nobody wants to hear the answer.
AylinWe are not running that experiment with 25% of revenue on the table.
İpekHolding 100% of the product hostage to avoid a conversation is not a strategy.
BarışI talk to those accounts every week, and they will walk, I'm not guessing.
CerenThen give me a name, Barış, one person who said that word, "leave."
SelimWe're spending a third of our time on a feature so we don't have to make a phone call.
AylinThe phone call risk is manageable; losing those accounts is not.
İpekBarış, if you talk to them every week, go ask them directly what it would take to migrate.
BarışSandra at Meridian told me six months ago she'd RFP competitors if reporting changed.
CerenSix months ago is not today, and "RFP" is not "leave."
SelimSandra said RFP, not goodbye, and we never followed up.
AylinRFP is how they leave, Selim, that is not a technicality.
İpekSix months and zero follow-up means we don't actually know where Sandra stands today.
BarışThose two accounts pay for this building, and I'm not running an experiment to find out.
CerenYou're already running an experiment, Barış, the experiment is staying blind for six months.
SelimNobody has asked them because nobody wants to hear the answer.
AylinWe are not running that experiment with 25% of revenue on the table.
İpekEvery quarter we wait costs more than the migration, and I will not move on that.
BarışI will not move on keeping those two accounts, full stop, that is what I'm here for.
CerenSomeone call Sandra this week or I'm putting the sunset on the roadmap anyway.
SelimCall her today, not this week, and I'm on that call.
AylinCall her, fine, but the sunset stays off the roadmap until we hear what she actually needs.
İpekBarış, you talk to them weekly, so you make the call today and we scope migration in parallel.
BarışI'll make the call today, but Sandra is not a migration project, she is a reason we still have a company.
CerenCall her today, scope migration this week, and if she doesn't say "leave" the sunset goes on the roadmap.
SelimIf she says stay, we scope what that actually costs before anyone promises her anything.
AylinScoping costs before we promise anything is exactly right, and the roadmap stays clear until we know the number.
NEXT