Alpha safety measurement
- Status: needs operator input, opened 2026-08-24.
- Issue:
#867: feat(safety): how we measure whether Alpha participants felt safe, without averaging groups together - Launch plan: ALPHA.md Phase 4 (Run), "Safety-comfort measurement" row. (Was Phase 2: Prove before the 2026-08-31 timeline shift moved the row beside the launch events.)
What is this?
A survey portal to measure whether Alpha participants felt safe, without averaging groups.
Alpha exists to prove two things: that anonymous strangers approach each other, and that people felt safe doing it. The second one has no design yet. The risk is that an overall average hides exactly the signal the pilot exists to find. If women's comfort differs from the overall mean, that difference is invisible in the numbers.
The design is a random portal, separate from the app.
Survey takers get there via a QR code at the event or a pop-up prompt in-app later. Merchants offer coupon incentives. The submission is never tied to their account, and all fields are optional so respondents control their own re-identification risk.
Four design decisions have been made.
- The coupon is a code word told to staff (settled 2026-08-16)
- The pop-up opens the portal as a plain link with no token or uid (settled)
- Granularity is kept and disclosed plainly so respondents can choose (settled)
- Responses live in their own isolated store, decoupled from the main app (decision pending: same GCP project or separate)
What is the current state?
The spec is written. Open risks are named as the remaining work.
The design spec lives at docs/planning/specs/2026-08-16-alpha-safety-measurement-design.md.
The four risks and their status are documented in issue #867. Remaining questions: where to store responses (same project or separate), and what the severity levels (red/orange/yellow) trigger in terms of stopping events, launches, or product.