War-game a beta invite list before selection rules, NDA language, and feedback expectations harden into who sees the unfinished product.
Beta invite lists fail when friends-and-family crowd out target ICPs, when NDAs are optional in practice, when feedback channels drown product with noise, and when support capacity cannot clear the claimed white-glove promise. A spreadsheet of emails is not evidence.
Freeze first
State the ICP you claim to learn from, the mix of friendly and adversarial seats, the NDA and disclosure rules, the feedback cadence you will honor, and the success metric after the beta window. Attach support capacity, known defect severity, legal review of invite copy, and the kill criteria if the cohort skews useless. If product, sales, and legal disagree on who may talk publicly, reconcile before seating.
Name the decision you will make if the war game finds nothing new, and the hold criteria if any invite lacks an owner for support or feedback.
Attack surfaces
- Selection bias: lists that only confirm what the team already believes.
- NDA theater: agreements that exist on paper and fail under social posts.
- Support fiction: white-glove promises the team cannot clear at volume.
- Feedback noise: channels that drown signal without a triage owner.
- Sales collision: invites that become unpaid pilots with enterprise expectations.
Optional counsel seat if regulated data or export controls bind the cohort. Optional CS seat if the same owners will carry the accounts after GA.
How to run it
Feed Pingpong the draft list criteria, invite copy, and open risk list. Early passes steelman the cohort. Later passes attack from target buyer, skeptic, support, legal, and sales seats. End with a pass that turns surviving objections into narrower eligibility, clearer NDA rules, or a hold. Delete invented "representative" mixes and dual-counted friend accounts.
Ask product and sales seats to price the expectation the invite will create. If the email reads like a private GA, buyers will demand SLAs the beta cannot meet. Write the intended promise, the escape language, and the upgrades you will refuse, then attack whether learning still happens under that discipline.
Force a day-after narrative: what happens if a beta user posts screenshots, if a friendly account never files feedback, or if support volume spikes on one unfinished flow. If those stories are stronger than your mitigation plan, fix the list before you send it. Related: war-game a beta invite, war-game a launch checklist, pretend you are the first customer, war-game a product launch, and the war-game decisions hub. Process: how to run a Pingpong.