Strategy
Product launch marketing questions nobody wants to ask out loud
The launch marketing questions teams avoid asking out loud, with the answer, the customer evidence that settles it and the trigger that reopens it.
What to take away
These are the product launch marketing questions teams avoid asking out loud, with the answer, the evidence that settles each one, and the trigger that reopens it.
- Write the launch decision record before the campaign. It carries the decision, the owner and the population, then the evidence period, the action and the limit.
- Ask the questions below in a room with sales, support and finance present, not in a marketing review.
- Set the stop rule and the review date at the same meeting, so evidence can overturn the answer later.
- Treat launch day as the start of the evidence period, not the end of the work.
- Escalate anything touching law, safety, contracts or privacy to the named authority, not to the launch team.
The questions nobody asks out loud
Most launch post-mortems fail for the same reason: the hard questions were never asked before release, only after the numbers disappointed. The ten below carry the most weight. Write an answer under each one before the campaign ships.
Who is the buyer when the champion cannot sign?
What does the product replace?
Which claim would we retract first if a customer challenged it?
What breaks in week one of real use?
Zendesk and Intercom both meter help desk cost by agent seat, so a ticket spike is a budget line. Entry plans typically run $55 to $150 per agent per month as of 2026, and Intercom adds a per-resolution charge on its newer plans.
What would make us stop?
Who owns the correction if a claim proves wrong?
Who signs a claim that touches a regulated category?
What did we promise in contracts marketing has not read?
Which number will the board ask for, and can we produce it?
Which customer are we excluding on purpose?
The launch decision record
Every answer above belongs in one file, dated and versioned. Two teams answering the same question differently usually differ on definitions, not facts. The architecture decision record pattern, published by Michael Nygard in 2011, works as a template: context, decision, consequences.
Launch decision record fields
- Decisioncommitment, named owner and deadline
- Populationwho claim covers and excludes
- Evidence perioddates and source date
- Actionwhat changes, owner and budget
- Limitwhat evidence cannot support, reopen trigger
| Field | What to record | Acceptance test |
|---|---|---|
| Decision | What the launch commits the company to | Named owner and deadline |
| Population | Who the claim covers, and who it excludes | Eligibility written down |
| Evidence period | Dates the supporting data comes from | Source date beside each figure |
| Action | What changes if the answer holds | Owner and budget line |
| Limit | What the evidence cannot support | Reopen trigger and date |
| Claim ranking | Weakest claim first, with its source | Retraction owner and channel |
| Week-one load | Ticket volume, staffing and seat cost | Support lead sign-off |
For interoperability and supplier dependence, the GOV.UK open standards guidance sets out how open standards reduce lock-in. It governs UK public services, so read it as a portability prompt rather than a rule for your market.
Evidence that survives a challenge
Test the launch result on a defined population before widening spend. Track qualified adoption after release alongside cost and quality, so a rising number that costs more per account does not read as progress.
Observe one normal path and one failure path. Preserve inputs, versions and rejected explanations. Amplitude and Mixpanel both report by cohort, which separates new accounts from imported ones.
The GAO evaluation design guide links evaluation questions to evidence needs and design choices. Its context is federal programs, so borrow the discipline and avoid causal claims your own method cannot carry.
Where a workflow uses AI, the NIST AI RMF Playbook lists voluntary actions under govern, map, measure and manage. It is not a certification and not a product ranking.
Privacy claims need the same care. The W3C Privacy Principles warn against pushing privacy work onto individuals. Apply that to the actual data flow, then confirm the governing law with counsel.
PIPEDA, CASL and Quebec's Bill 96 each set requirements that vary by activity and province, so name the regulator and take qualified Canadian advice.
Worked example: reopening a claim
A team launches with the claim that setup takes one afternoon. Support logs show a median of three days for accounts importing historical data. The decision record said the claim covered new accounts only, and that was never printed.
The correction is not a new campaign. It is a narrower claim, a segment note on the pricing page, and a script change for sales. The reopen trigger fires, the owner named in the record issues the fix, and the next review reads adoption by segment rather than in total.
Common questions
What is the first decision in a launch?
Name the decision the launch commits the company to, and the owner who can reverse it. Everything else, including budget and channel, follows from that sentence. For teams that want the fuller working method behind this, how to improve product launch marketing for business teams covers the same ground in more detail.
How long should the evidence period run?
Long enough to cover one full buying cycle for the target segment, and no longer than the point where the stop rule can still save spend. Write both dates in the record.
When should a launch claim be pulled?
When the evidence behind it weakens, not when a competitor points it out. Pull it early, in the channel where it ran, with the correction owned by a named person.
What should stay out of the launch team's hands?
Legal, tax, safety and certification questions. Name the standards body, regulator or counsel, and let the launch record show who confirmed the answer and when.
