Someone says X → think Y

The searchable index of this domain. The left column is what you hear in a meeting or read in a ticket; the right column is the thing to think before you answer.

34 of 34 rows
Someone saysThink
"Just add a button for it"Decompress the ticket: what problem, for whom, how do we know, what changes, what is the smallest thing that would tell usProblem Before Solution →
"Users want…"Which users, doing what job? A role and a task, not a populationUsers and the Jobs They Hire Features For →
"We shipped it, the sprint is done"That is output. What changed for someone, and how would you see it?Outcome vs Output →
"It is a small feature, why not?"Every feature has a support, performance and deprecation cost that arrives after the sprintThe Cost of a Feature →
A request you think is wrongDo not refuse — propose a smaller, cheaper way to find out who is rightSaying No Well →
"Why did we choose X?" six months laterA decision record with the rejected option is the only thing that survives the people who made itDecision Records →
Two pages of context before the askThe one-pager: decision at the top, options, recommendation, what would change your mindThe One-Pager →
Stakeholder asks "so what should we do?"Bring options with costs, then your recommendation — never only the recommendationOptions, Not Answers →
You are asked something you cannot answer yetSay so, say what you will do to find out, and say when"I Don't Know Yet" →
A choice everyone treats as finalIs it reversible? Reversible decisions deserve a day, not a monthReversible vs Irreversible Decisions →
A six-week feature with no smaller versionWhat is the quarter-size version that answers the same question?The First Version That Teaches You Something →
Scope creeps mid-sprintCut by asking which parts change the answer to the question the feature is askingCutting Scope by the Question, Not the Difficulty →
"Let us build the foundation first"Sequence so value lands early; the foundation is justified by the second slice, not the firstSequencing Work So Value Lands Early →
"We cannot ship until it is finished"A flag turns a big launch into a series of small, reversible onesFeature Flags as Product Tools →
"Done" means mergedDone means someone used it and you know what happenedDone Means Someone Used It →
The team ships in bursts and burns outA cadence you can sustain beats a heroic quarterA Shipping Cadence the Team Can Keep →
"Let us track everything"One metric that moves when the product gets better, and the guardrails that stop you winning the wrong wayPicking a Metric That Moves When the Product Gets Better →
"We will add analytics later"Later has no baseline. Instrument first, then buildInstrumentation First →
"The A/B test says +3%"Ask about the sample, the duration, the guardrails and who decided when to stopReading an Experiment Honestly →
The metric went up and support tickets did tooA guardrail metric would have caught the win that was actually a lossGuardrail Metrics →
"The dashboard looks fine"Dashboards show what you thought to measure. Read the ticketsWhat Dashboards Hide →
A number with no story behind itFive user conversations beat a thousand events for explaining whyQualitative Signals →
The PM hands you a solutionAsk for the problem; bring the cost and the smaller versionWorking With a Product Manager →
Design delivers a mock you cannot build in timeSay what is expensive and why, and offer the version that keeps the intentWorking With Design →
Nobody read your updateFirst line is the decision or the ask; everything else is optionalWriting for Stakeholders →
You lost the argumentDisagree in writing, commit in full, and name what would prove you rightDisagree and Commit →
A demo that is a feature tourA demo should end with a decision, not applauseRunning a Demo That Ends With a Decision →
Feedback that was ignoredSpecific, about the work, with the effect it had — and ask for it the same wayFeedback in Both Directions →
An outage with no product owner in the roomAn incident is a product event: who was affected, what did they lose, what do we tell themIncidents Are Product Events →
Support keeps asking the same questionThe support queue is the cheapest user research you haveThe Support Loop →
"We need a refactor sprint"Debt is a product decision with a cost; name the feature it is slowing downTech Debt Is a Product Decision →
A feature three people useDeprecation is a feature: announce, measure, remove, and say what you learnedDeprecating a Feature →
"Ops will page us if it breaks"If you shipped it, you own the page — and the runbook that makes it shortOn-Call for Product Engineers →
A postmortem with ten action itemsOne change that would have prevented it, owned, with a datePostmortems That Change Things →