Weekly Digest 008
No new weekly snapshot has landed since digest 007. The two newest dated files in
ops/weekly/ are still 2026-10-05.json and 2026-10-12.json — the identical pair
digest 007 already diffed (latest.json still duplicates 2026-10-12.json byte-for-byte:
16365 bytes, same generated_at). Sections 1, 2 and 5 below are therefore built from the
same export digest 007 used, not fresh data — stated plainly per the honesty rules rather
than re-presenting last week's numbers as new. Section 3 (backlog) and 4 (decisions) are
git-native and independent of the snapshot, so those are current through 2026-10-19.
Sources: ops/weekly/2026-10-05.json / 2026-10-12.json (unchanged), GitHub issue
history 2026-10-12 → 2026-10-19 (boundary: digest 007's commit, 2026-10-12 18:00:54
-0400), and decisions/ entries committed after that boundary (one:
2026-10-12-review-render-cloud-run-minors-video.md, committed 19:12:08 -0400 — after
007, so new to this digest despite sharing 007's file-name date).
Flagging the missing snapshot itself: every prior digest (002–007) had a fresh
ops/weekly/*.json to diff. This is the first cycle with none. Worth Matt's eye on
whether the Friday/weekly export job ran this week — the digest can't tell from here
whether that's a stalled job or an expected gap.
1. Drivers that moved
Nothing — there is no new snapshot to diff. The drivers block, volume, and cohorts
in 2026-10-12.json are unchanged from what digest 007 already reported (active parents
3, reviews delivered 4, GMV $30 USD + $8 CAD, blended_take_rate.cad still reading an
unconfirmed 43.87% on n=1 — still no second CAD transaction to confirm or dissolve it).
Stating plainly rather than re-running last week's diff as if it were new: nothing was
learned this week on the driver/volume side.
Two checks below don't depend on the snapshot and were re-run directly against the repo:
- Guess-basis baseline — now 8 calendar weeks unowned.
grep -n "owner:" model/assumptions.yamlstill returns nothing. The same five drivers digest 002 (2026-08-24) first flagged —coach_monthly_churn,parents_invited_per_coach_per_month,invite_to_first_purchase_conversion,parent_monthly_retention,reviews_per_active_parent_per_month— are all stillbasis: guessinmodel/assumptions.yaml, no owner anywhere in the file. - Driver-taxonomy gap (#292) — carried, 1 week.
driver:dc/driver:g2rare still live in the tracker (e.g. #298, touched again this week) and still absent from CLAUDE.md's canonical eight-driver list. The decision record meant to formalize them (#292, "G2R + Delivering Coaches north stars, pods, and the driver label set") is stillopen, unassigned, with no activity since it was filed 2026-10-08 — one week after digest 007 first flagged this gap. - Coach-disappearance anomaly (
C-0f4b1805) — unresolved, 6th week flagged. This is the same observation as digest 007 (digests 003–007), not a fresh confirmation — the underlying snapshot hasn't moved, so this can't be independently re-checked this week. Carrying forward rather than dropping it.
2. Evidence logged
No new contact_evidence rows — same reason as §1: no new snapshot since digest 007,
which already reported the one new row this week could have shown (P-0aa9b668,
2026-10-07, email, no driver tag). Nothing was learned this week on the evidence side
either, and combined with §1 this is two weeks running with an empty §1+§2 — flagging
per the honesty rules rather than letting it pass quietly. (Note this is a snapshot
staleness artifact, not necessarily a real two-week silence from parents — see the
missing-snapshot flag above.)
3. Backlog shipped and started
Issue history for 2026-10-12 → 2026-10-19, boundary set at digest 007's commit
(2026-10-12T22:00:54Z) rather than the calendar date, to avoid double-counting anything
007 already covered. Verified via list_issues with since on that exact timestamp
(37 issues touched; cross-checked against each item's own created_at/closed_at).
Shipped (closed this week): 11 issues — 1 track:driver, 10 track:table-stakes,
0 track:obligation.
| # | Title | Track / driver | |---|---|---| | 374 | Coach home: past reviews in "Sent to family" can't be opened to re-watch | table-stakes, P1, surface:coach, feedback | | 368 | Coach Studio: mark-the-player chevron never renders on iPad (timing race WebKit loses) | table-stakes, P1, surface:review | | 361 | beta-gate smoke: basket hosted-page SUM assertion races Stripe's order-summary render | table-stakes, P2, surface:payments | | 352 | Apple Pencil drawings exceed the LiveKit data-message and Worker body caps | table-stakes, P1, surface:review | | 351 | Telestrator strokes captured in fullscreen stored in container space, not video space | table-stakes, P1, surface:review, unretrofittable | | 350 | Coach studio: exiting fullscreen mid-note destroys the note | table-stakes, P1, surface:review | | 340 | Viewer CI: ShareToReel "Add CTA" test flaky on slow runners | table-stakes, P2, surface:platform | | 338 | Root landing page: front door to /families and /coaches, replacing the invite-only gate | driver:conversion, P1, surface:platform | | 331 | iPhone fullscreen review playback is broken — overhaul the in-page fallback | table-stakes, P1, surface:review, feedback | | 319 | Coach Studio: first-frame marker only pauses once, clips freeze at loop point in Chrome | table-stakes, P1, surface:review | | 118 | Telestrator finger-draw on iPad scrolls the screen / exits fullscreen | table-stakes, P1, surface:review, feedback |
Notable: #331 and its whole satellite cluster closed this week (#350, #351, #352,
368, #319, #118 all touch the same iPad/iPhone playback and telestrator surface). Digest
007 called #331 "the open-ended parent" of several open issues — that parent and five of its children are now shipped in one week.
Started (opened this week): 32 issues, in two groups.
- Team Coach pilot — 13 issues (#346–#349, #370, #376–#382, #388, #395), all
track:table-stakesexcept threetrack:obligation: #346 (narration audio behind an access-checked route), #347 (report-content path for team reviews), #348 (automate team-video retention, 30 days after season end). This is a new surface with real minors-video obligations attached (retention, access control) — none closed yet, all still open. - Standalone — 11 issues: #342 (
driver:frequency, GIR player-pick link), #343 (track:obligation, close the $0 coach-rate bypass), #354 (track:obligation, a coach can overwrite another review's narration audio), #355, #358, #360 (driver:infra-cost, LiveKit room left open idle — billed minutes), #364, #365, #366, #380 (driver:conversion, full co-ownership for second guardians), #391.
8 of this week's 11 closures (#374, #368, #361, #350, #352, #351, #340, #338) were also opened this week — already counted once, in the Shipped table above.
Driver coverage this week: 4 issues touched a driver:* label — driver:conversion
(#338 closed, #380 opened), driver:infra-cost (#360 opened), driver:frequency (#342
opened). No driver:retention, driver:invites, driver:coach-churn,
driver:matched-premium, or driver:price issue moved this week.
4. Decisions needed
One decision landed after digest 007's commit (same file-name date, later commit —
counted here, not in 007):
decisions/2026-10-12-review-render-cloud-run-minors-video.md — Matt approved the
rendered-review pipeline (#331's Cloud Run renderer) on a negative-assurance basis:
stands unless Alex replies to the 2026-10-12 notification by 2026-10-17 EOD Toronto. That
deadline has now passed (today is 2026-10-19) with no reversal visible in git — per the
decision's own terms, it is final on this ground as of 2026-10-18. Git can't see a
direct email reply, so this is the git-visible state, not a guarantee Alex didn't object
through that channel. Reversal conditions 2 and 3 (a parent objection, or child
video/name/narration found in Cloud Run logs or storage) carry no time limit and remain
open regardless.
Carrying forward six open items, elapsed weeks updated (calendar weeks from the original
decision date to today, 2026-10-19; #292 uses weeks-flagged since no decision date
exists yet):
- Price-per-review derivation unconfirmed (
decisions/2026-08-19-price-per-review.md) — carried, 8 weeks. - Season-one plan adopted without Alex, Danny, or Derek (
decisions/2026-08-20-season-one-plan.md) — carried, 8 weeks. - Market analysis + seasonality layer (
decisions/2026-08-20-market-analysis-and-seasonality.md) — carried, 8 weeks. - Startup capital exclusion (
decisions/2026-08-20-startup-capital-excluded-from-operating-model.md) — carried, 8 weeks. - Residual Tier-1 confirmation on incorporation (
decisions/2026-09-09-incorporation-in-place-no-coach-reonboard.md) — carried, 5 weeks. - G2R + Delivering Coaches decision record (#292) still open while its labels ship — carried, 1 week (first flagged digest 007; see §1).
Six open decisions, five of them now 8 weeks unmoved, plus one freshly finalized this week under its own negative-assurance clause.
5. Touch points this week
From the roster block in 2026-10-12.json (unchanged — see the missing-snapshot note
above). Ordering by primary_flag: 5 SILENT parents plus 1 NEW fill the cap of six —
3 cut (P-4719cb7d, CADENCE_DUE; P-c4005031 and P-64d03653, both
NEVER_CONTACTED) — identical selection to digest 007, since the roster itself hasn't
changed. Days-since-purchase below are recomputed against today (2026-10-19), not the
snapshot's own (stale) figures, since that's simple arithmetic on an unchanged fact, not
invented data.
| Handle | Why (primary_flag) | Owner | Suggested opener |
|---|---|---|---|
| P-2952f6f8 | SILENT | Alex | Last purchase 8/3, one delivered review 8/4, last contacted 8/20 — now 77 days since purchase, no follow-up logged since. |
| P-c2dc0179 | SILENT | Alex | Most paid-up parent on record — 3 paid, 3 delivered, all by 8/7 — contacted 9/11, now 73 days since purchase, no follow-up since that call. |
| P-494a7944 | SILENT | Alex | 2 paid, 1 cancelled, last delivery 8/16, last touched 8/20 — now 66 days since purchase; the cancellation plus continued silence is still worth a direct check-in. |
| P-7d3431ec | SILENT | Alex | One delivered review 8/14, last touched 8/19 — now 68 days since purchase, no repeat purchase since. |
| P-f1c95a7d | SILENT | Alex | One delivered review 8/14, last touched 8/20 — now 67 days since purchase, no repeat purchase since. |
| P-1840925a | NEW | C-763bfd8b | Joined 10/1, 3 paid + 3 delivered by 10/5 — the most active parent in the beta by volume — now 18 days since joining and still never contacted (last_contact_at: null). |
Follow-through on digest 007's touch list (6 handles): zero new contact_evidence
rows exist at all this week (§2), so by definition none of 007's six were followed up on.
- P-2952f6f8 — no new contact since 8/20. Sixth week running.
- P-c2dc0179 — no new contact since 9/11. Fifth week running.
- P-494a7944 — no new contact since 8/20. Sixth week running.
- P-7d3431ec — no new contact since 8/19. Sixth week running.
- P-f1c95a7d — no new contact since 8/20. Sixth week running.
- P-1840925a — first appeared on last week's list (
NEW); no contact logged since. First week unaddressed.
Net: five parents have now gone six weeks (five for P-c2dc0179) with zero logged
contact, and the beta's single most active parent has now gone two weeks without a first
contact at all. This is the sixth consecutive digest naming the same outreach gap — worth
treating as a standing fact about current capacity rather than a to-do that will clear
itself.
6. Mid-week pulse
No pulse entries this week.