Product & DesignFree
Stakeholder Requirements and Conflict Finder
Every ask is reduced to the need behind it, then checked against every other ask for the pairs that cannot both be satisfied.
River takes the pile as it is. Forwarded threads, a ticket queue, three sets of meeting notes and the message somebody sent at eleven at night all go in together. Each request is reduced to the need behind it and recorded with who asked, where they asked, and what they actually said. Then every ask is checked against every other ask, and the pairs that cannot both be satisfied come out as their own list with a name against each.
Reducing an ask to a need is the step that makes the rest work, and it is older than software. NASA's requirements guidance tells authors to state the problem rather than the solution, and to ask why the requirement is needed, because the answer may point to the real requirement. Two people asking for different buttons often want the same thing. The templates ranking for this query give you a spreadsheet with a priority column, which sorts requests without ever comparing them.
Built for the product manager who has been asked four times this week when their request is happening, and who does not yet know that two of those requests contradict each other. Reach for it before the spec, not after. What comes out feeds the PRD template, the effort side of any conflict gets priced by the feasibility brief, and customer requests arriving through the same pile belong in feedback triage first.
The conversation that waits until build time
Halvard spent a quarter rebuilding onboarding and collected forty-seven requests along the way: nineteen in email threads, sixteen in tickets, twelve in meeting notes. Sixteen were restatements of something already captured, which leaves thirty-one distinct asks. Those thirty-one reduce to nineteen underlying needs, and ten of those nineteen have more than one ask attached, which is the first useful number. Ten needs are already contested before anyone has written a requirement, and nobody in the company knows that yet.
Thirty-one asks make four hundred and sixty-five pairs, and every pair gets checked for whether both could hold at once. Seven cannot. Sales wants a trial workspace with no payment method; the controller wants a card on file before any import runs, and the import is the first thing a trial does. Security wants single sign-on enforced from account creation; sales wants a solo trial user with an email address and no identity provider to enforce against. Four of the seven have sales on one side.
Fourteen of the thirty-one asks become the shape of a requirement. Seventeen get a written decline, and ten of those seventeen are not refusals at all: the need is being met, in somebody else's version. That distinction is most of the political work. NASA's guidance on stakeholder expectations says to know who has the decision authority to help resolve conflicts, so every conflict here carries a named person rather than a note saying it needs discussion.
How it works
Hand over the pile
Forwarded threads, exported tickets, meeting notes, and the requests that only ever existed in conversation.
Reduce to needs
Every ask restated as the problem behind it, so different solutions to one problem merge.
Scan every pair
Each ask checked against each other ask for the pairs that cannot both be satisfied.
Write the answers
Requirements, the conflict register with a named decider, and a decline for every ask not shipping.
What you get
- Every raw request kept with who asked, where they asked, and their words
- Each ask reduced to the underlying need, so competing solutions to one problem collapse
- Every pair of asks checked, and the ones that cannot both hold listed
- Each conflict carrying the two people in it and the person who decides
- Ordering dependencies separated from real conflicts, because one is a schedule and one is a choice
- A written decline per unshipped ask, separating a refusal from a need already covered
- A per-person answer, so each requester learns where their asks landed without asking again
Common questions
How is this different from a request log with a priority column?
A log ranks requests against a scale. This compares them against each other, which is a different operation and the one that finds contradictions. Ranking thirty-one asks tells you which is most important. Scanning all four hundred and sixty-five pairs tells you that seven of them cannot both be built, which ranking never surfaces.
Does it really check every pair?
Yes, and that is cheap at this size. Thirty-one asks make four hundred and sixty-five pairs, which is a few minutes of reading. Past roughly eighty asks the pairwise scan is restricted to asks that touch the same surface or the same data, because two requests about unrelated parts of the product cannot contradict each other.
Will it decide the conflicts for me?
No. A conflict between the general counsel and the implementation lead is not a question with a correct answer, it is a question with an owner. Each row states both positions in their own terms, why they cannot both hold, what each choice costs, and the person whose call it is.
What happens to the requests that do not make it?
Each gets a written decline naming the requester and the reason. The split matters more than the count. In the worked example seventeen asks do not ship: ten are covered by somebody else's version of the same need, and seven attach to a need deferred to a dated slot on the roadmap.
Some of my requests are from customers, not colleagues.
Then triage those separately first, because customer volume means something and a colleague's insistence does not. Run feedback triage over the customer pile, bring its output in here as one input among the internal asks, and the conflict scan will happily find contradictions between a customer theme and an internal ask.
What is an ordering dependency, and why is it separate?
Two asks that can both happen, but only in one sequence. The signed agreement has to exist before real customer data enters a mapping step. That is a schedule problem with one answer. A conflict is a choice with two answers and a loser, and mixing the two makes both look negotiable.
Can I show this to the stakeholders themselves?
It is built to be shown. The per-person view gives each requester their own asks and where each one landed, which answers the question they were going to ask you anyway. In the worked example one stakeholder has nothing shaped this cycle, and that is the one conversation to have in person.
Stakeholder Requirements and Conflict Finder
Fill in the form and your workspace opens with the work already underway.