VALUE.md - Pinpoint (worked example, fictional)

This is a complete VALUE.md for a fictional incident-cause-ranking tool, included as a reference example. Adopters can read it in 60 seconds and see the standard's full anatomy applied to a single concrete case.

Value Statement

Pinpoint gives the on-call engineer ranked likely causes within 90 seconds of the page; the engineer gives back the resolution data that sharpens the next page's ranking.

Q0: Grounding observation

On 2026-03-14, between 03:00 and 03:45 EST, I shadowed an on-call engineer who got paged at 03:12. She spent 22 minutes opening five separate dashboards, three log searches, and one tracing tool before identifying the root cause. The cause was visible in the third dashboard she opened, but she did not know which dashboard would contain it. The 22 minutes were not thinking; they were searching.

Q1: Who it's for?

The on-call engineer who gets paged at 3am and has to find the cause fast. Not their VP. Not "users."

Q2: What changes for them?

Before: they open the alert, navigate to whichever dashboard they most recently used for a similar-shaped alert, and start searching. Median time-to-cause is 22 minutes (n=14 observed pages).

After: the alert arrives carrying three ranked likely causes, each linked to the specific dashboard or log slice where the evidence lives. Median time-to-cause drops to under 5 minutes.

Q3: How will you know the change happened for them?

Each resolved incident records whether the real cause was in the top three at page time. The named recipient (the engineer who took the page) attests yes or no.

Breaks if: the real cause is NOT in the top three on more than 1 in 10 incidents over any rolling 30-day window.

Check: Field - per-incident yes/no, logged in audits/q3-acceptance/YYYY-Q.md.

Status: Proposed.

Last verified: not yet · Re-verify by: empty (status is not yet Proven)

Promises

P1 - Time-to-cause drops measurably

Status: Proposed · Recipient: the on-call engineer.

You spend fewer minutes per page on cause-search.

Breaks if: median time-to-cause across all pages in any rolling 30-day window is not lower than the pre-Pinpoint baseline.

Check: Field - median over rolling 30-day window, logged in the incident-tracking system.

Last verified: not yet · Re-verify by: empty (status is not yet Proven)

What we don't promise

• We don't promise the rankings are correct on novel-shape alerts. The system learns from past resolutions; first-of-kind incidents will still require manual search.
• We don't promise the ranking arrives faster than the page itself. The 90-second budget assumes ranking happens between page-fire and engineer-acknowledgment.

What we don't break

• We won't trade ranking quality for ranking speed. If the ranking-accuracy rate (cause in top 3) drops below 60% over any rolling 30-day window, the speed gain is no longer worth the cognitive cost of wrong rankings; we re-evaluate the model.