Product wants a date the design cannot meet. How do you approach the conversation?

What is being tested

Whether you work from interests rather than positions, and whether you produce options rather than a verdict.

The first move: find the interest behind the date

"Six weeks" is a position. The interest is why: a customer commitment, a contractual date, a competitor launch, a conference, an expectation someone has already set internally.

Positions conflict; interests frequently do not. "We need it in six weeks" against "this needs ten" is a deadlock. "There is a pilot customer expecting access in six weeks" opens options that were not visible before.

Ask why before arguing about what. It is the single most useful move available and it is skipped almost every time.

Then produce options, not a verdict

"Full scope is ten weeks. In six we can deliver the pilot path with manual handling for exceptions — about ten a week — and complete the automation by week twelve. Or we can deliver the full scope for one customer segment in six weeks and extend afterwards. Which fits the commitment better?"

Three properties make that work: it takes the constraint seriously, it offers real choices with stated costs, and it leaves the decision with the person who owns the outcome.

The techniques that help

Separate reversible from irreversible. Concede on the reversible; hold on the one-way doors — a data model shortcut, a security boundary, a tenancy decision. Knowing which is which is what makes you credible when you do hold.

State constraints as constraints, sparingly. "This is not a preference — the residency obligation means this cannot go in that region." Reserve that framing for things that genuinely are constraints, or it stops working.

Make the debt explicit if you take it. "We can hit six weeks by skipping the reconciliation process. That means manual intervention for exceptions until month three, and here is the ticket." Deliberate debt recorded is legitimate; undocumented shortcuts are indistinguishable from mistakes six months later.

What to avoid

Agreeing to something you know will not work, which defers the conflict to a worse moment with less room to manoeuvre.

Winning, which produces a resented decision.

Escalating, which spends credibility and rarely produces a better answer than the conversation would have.