template · signal reading

Signal reading template

Copy-paste skeleton. Five lines plus outcome plus voice-of-customer. Written on the calendar-committed check date. Lives next to the brief.

How to use

The five lines are non-negotiable. Replace every [...] with concrete content. Write it before the retro, not after. The gap line is the discipline — "better than predicted" alone is not a gap line.

# Signal reading — [cycle name] · [check date YYYY-MM-DD]

Prediction:  [Quote the prediction from the Feature Brief
              verbatim.]

Baseline:    [Number with sample size and date, captured how.]

Target:      [Specific number / range / threshold from the
              brief.]

Measured:    [Number observed on the check date, with sample
              size, range of dates, and method-of-measurement.]

Gap:         [Named anomaly. What the data says, what the
              variance shows, what next-cycle work this
              implies. Specific. Not "better than predicted".]

Outcome:     Met | Missed honestly | Refuted | Not checked

## Voice-of-customer   (added by CS within 24h)
[CS Lead's paragraph: what tickets witnessed during the
 window. Moment, not feature request. Named themes, not
 ticket counts.]

## Linked artefacts
- Feature Brief:           [link]
- Technical Design Brief:  [link]
- First 48h watch note:    [link]
- Cycle's retro (when written): [link]
- Cycle's model update (when written): [link]

## Sign-off
PO: [Name] · [date]

Worked example — Gal's grading session

# Signal reading — Grading shortcut · 2026-06-25

Prediction:  "Gal completes the grading cycle in under 15 minutes."
             — Feature Brief 2026-05-04, signed by Alex (PO),
             Esti (TL), Maya (Designer).

Baseline:    47 minutes (mean, n=12, captured 2026-04-22 to
             2026-04-26 by sitting next to Gal in three
             grading sessions across three Fridays).

Target:      Under 15 minutes.

Measured:    11 minutes 20 seconds (mean, n=8, captured weeks
             1-3 post-flag from product analytics + observed
             on 2026-06-23 in a follow-up session with Gal).

Gap:         Better than predicted. Investigate why — likely
             the new keyboard shortcut absorbed more time than
             the deep-link navigation we built for. The deep
             link is barely used (1 use in 8 sessions). Consider
             retiring the deep-link path next cycle; sharpen
             the shortcut's UX for the cases where it stalls
             (Hebrew language switching mid-session — see CS
             voice below).

Outcome:     Met

## Voice-of-customer (added by CS within 24h)
Three Hebrew-default schools reported the shortcut "freezes"
at submit. Pattern: language-pack load delay creates a 2-3s
unresponsive feel. Eight tickets over four weeks; pattern
named in helpdesk-reading 2026-05-16. Brief input for next
cycle.

## Linked artefacts
- Feature Brief:           /briefs/grading-shortcut.md
- Technical Design Brief:  /briefs/grading-shortcut-tdb.md
- First 48h watch note:    /watches/2026-05-25-grading.md
- Helpdesk reading:        /helpdesk/2026-05-16.md

## Sign-off
PO: Alex · 2026-06-25

Where this lives in your project

The signal reading lives next to the Feature Brief — same location, same folder. It is the brief's answer. Linked from the retro and the model update. Never deleted; subsequent cycles reference it.

Outcomes — pick one

Outcome	When	What happens next
Met	Measured ≥ target	Retro + model update; scale the bet or close the initiative
Missed honestly	Measured < target; the gap is named honestly	Retro + model update; the cycle was not waste
Refuted	The mechanism the bet relied on did not hold	Retro + model update; consider killing the initiative
Not checked	The date passed without measurement	The worthless outcome. Retro names it as a chain-level signal.

What to do if a section resists

Resistance	What it means	Where to go
Check date drifted	The PO did not hold the calendar	Mark not checked honestly; chain-level signal at retro
Method was reused from baseline	The check is recursive	Clinic — A brief that didn't witness
Gap line is one sentence	The PO did not look at the variance	Re-read the data; the gap is in the variance, not the mean
No CS voice	The reading was written in a vacuum	Skill · CS Lead foundations · Step 8
Outcome line is "kinda met"	The team is hedging	Pick one of the four. Hedging is not checked in disguise.

See also

• Practice — Signal reading
• Practice — Retrospective · Model update (the two that read this template's output)
• Canon — After We Build · Signal & The Prediction
• Principle — Not checked is the only worthless outcome