Decision record for the Weekly Signals Triage Ritual. Tracking issue: organvm-v-logos/.github#4.
Each cycle is labeled Wn and dated by the Monday the weekly-signals artifact
was published. Decisions use the taxonomy in the
ritual definition: Act / Watch / Ignore.
Raw signal data lives in
analytics-engine;
this file holds the decisions and the remediation they produced.
Publish-to-remediation latency = days from the publish Monday to the day the resulting fix shipped. Target: median ≤ 7 days (SLA).
| Cycle | Published | Signals (A/W/I) | Remediation shipped | Cycle latency (days) |
|---|---|---|---|---|
| W1 | 2026-03-16 | 1 / 1 / 1 | Reading-path CTA copy on landing surfaces | 7 |
| W2 | 2026-03-23 | 1 / 1 / 0 | weekly-signals.md narration legibility |
6 |
| W3 | 2026-03-30 | 1 / 0 / 1 | Referrer attribution fix in engagement export | 5 |
| W4 | 2026-04-06 | 1 / 2 / 0 | External-reader headline framing default | 8 |
| W5 | 2026-04-13 | 0 / 2 / 1 | — (watch-only cycle) | n/a |
| W6 | 2026-04-20 | 1 / 1 / 1 | Primary CTA moved to first screen | 6 |
| W7 | 2026-04-27 | 1 / 0 / 2 | Evidence block ordering on profile README | 4 |
| W8 | 2026-05-04 | 1 / 1 / 1 | external_feedback.yml issue template |
6 |
| W9 | 2026-05-11 | 1 / 1 / 0 | Quick Start “which repo do I clone” path | 7 |
| W10 | 2026-05-18 | 0 / 1 / 2 | — (watch/ignore cycle) | n/a |
| W11 | 2026-05-25 | 1 / 0 / 1 | Stale-link sweep across org READMEs | 5 |
| W12 | 2026-06-01 | 1 / 1 / 1 | Bounce-on-mobile landing fix | 6 |
| W13 | 2026-06-08 | 1 / 0 / 1 | Consult-path link surfaced on landing | 6 |
| W14 | 2026-06-15 | 1 / 1 / 1 | in progress — see cycle entry | open |
Rolling median latency (shipped cycles W1–W13): 6 days. Within SLA. This is the figure the June 5 outcome review cited when it decided to MAINTAIN the ritual.
Signals. First published signal set. Landing traffic up but the click-through to the essays (“qualified reader action”) sat near 1.1% — far below the share of visitors who scrolled past the fold.
| Signal | Decision | Notes |
|---|---|---|
| Reader-action rate ~1.1% despite healthy scroll depth | Act | Visitors read the identity copy but don’t reach a reading path. |
| Direct vs. referral split shifting toward GitHub profile links | Watch | One week of data; trigger: 3-week sustained shift. |
| Single 404 spike on a renamed essay slug | Ignore | Expected — slug was redirected the same day upstream. |
Remediation. Added an explicit “read the essays” call-to-action to the landing surfaces so the reading path is reachable without finishing the system explanation. Shipped 2026-03-23 (latency 7d). Confidence: directional — small sample, but the CTA was demonstrably absent before.
| Signal | Decision | Notes |
|---|---|---|
weekly-signals.md hard to act on — metrics without deltas or thresholds |
Act | The narration listed values but not week-over-week movement. |
| Reader-action rate ticked to ~1.4% after W1 CTA | Watch | Encouraging; needs more weeks before crediting the change. |
Remediation. Reformatted the human-readable signals narration to lead with deltas and threshold breaches (so triage classifies faster). Shipped 2026-03-29 (latency 6d). This change is the reason later cycles close faster.
| Signal | Decision | Notes |
|---|---|---|
| ~30% of sessions attributed to “direct” that were really GitHub referrals | Act | Referrer header dropped on one redirect hop, polluting attribution. |
| Weekend traffic dip | Ignore | Seasonal/weekly pattern, consistent across the whole window. |
Remediation. Fixed referrer preservation in the engagement export so the GitHub-profile channel is measured correctly. Shipped 2026-04-04 (latency 5d). Confidence: strong — verified against server logs after the fix.
| Signal | Decision | Notes |
|---|---|---|
| First-time visitors from profile links bounce before the reading path | Act | Identity-first headline front-loads internal language. |
| Reader-action rate now ~1.9%, holding the W1 gain | Watch | Promote to a conversion-experiment hypothesis (feeds #5). |
| Mobile share rising | Watch | Trigger: mobile bounce > desktop by 10pt. |
Remediation. Switched landing headline to external-reader framing (“Essays from building an eight-organ system in public”) ahead of the organ identity. Shipped 2026-04-14 (latency 8d — slowest cycle; copy needed sign-off). This decision seeded experiment DC-01 in distribution-conversion-experiments-2026.md.
Watch-only cycle — no remediation shipped, by design.
| Signal | Decision | Notes |
|---|---|---|
| Conversion variance high week-to-week | Watch | Controlled experiment (#5) is the right instrument, not ad-hoc fixes. |
| Returning-visitor rate flat | Watch | Trigger: 4-week decline. |
| Bot-filtered traffic noise | Ignore | Already excluded upstream; restating for the record. |
Note. Deliberately shipped nothing. Acting on noisy week-to-week conversion swings would have pre-empted the controlled experiment series and produced false attributions.
| Signal | Decision | Notes |
|---|---|---|
| Motivated visitors scroll past the fold before reaching any CTA | Act | CTA still lived below the identity block. |
| Experiment DC-02 reading positive | Watch | Let the experiment window close before generalizing. |
| Duplicate pageview events on reload | Ignore | Deduped in export; immaterial to decisions. |
Remediation. Moved the primary essay CTA into the first screen, repeated after repository context. Shipped 2026-04-26 (latency 6d). Consistent with the DC-02 result later canonicalized as a default operating rule.
| Signal | Decision | Notes |
|---|---|---|
| Visitors arriving from GitHub profile distrust philosophy-first framing | Act | Proof-of-activity buried below conceptual rationale. |
| Low-engagement traffic from an unrelated aggregator | Ignore | Not the target audience; no action. |
| Search-referral keywords off-topic | Ignore | Below materiality; note for a future SEO pass. |
Remediation. Reordered the profile/landing to lead with concrete evidence (repository count, essay count, live destination) before the conceptual explanation. Shipped 2026-05-01 (latency 4d — fastest cycle). Mirrors experiment DC-04.
| Signal | Decision | Notes |
|---|---|---|
| Engagement present, but no path for a reader to respond | Act | Stars/traffic rising; zero structured feedback channel. |
| Conversion now ~2.6%, sustained | Watch | Trending toward the review’s 3.5% figure. |
| One spammy issue | Ignore | Closed; no pattern. |
Remediation. Added the
external_feedback.yml
issue template so engagement can convert into the substantive feedback that
#7 requires. Shipped
2026-05-10 (latency 6d). This is the engagement-vs-feedback boundary the
feedback ledger enforces.
| Signal | Decision | Notes |
|---|---|---|
| New visitors land on the org but can’t tell which repo to clone first | Act | High org-page exits with no repo click-through. |
| Mobile bounce climbing toward the W4 trigger | Watch | Trigger nearly met; carry forward. |
Remediation. Shipped the Quick Start guide answering “which repository do I clone?” and “how do I run it locally?”. Shipped 2026-05-18 (latency 7d). Tracks org issue #25.
Watch/ignore cycle — no remediation shipped.
| Signal | Decision | Notes |
|---|---|---|
| Post-Quick-Start repo click-through rising | Watch | Let it accumulate before crediting the guide. |
| Referral burst from a single external link | Ignore | One-off; no durable channel. |
| Slight conversion dip | Ignore | Inside noise band established in W5. |
| Signal | Decision | Notes |
|---|---|---|
| Several 404s from inbound links to renamed/old org READMEs | Act | Quick Start raised inbound link volume, exposing stale links. |
| Steady weekend pattern | Ignore | Known seasonal shape. |
Remediation. Swept and corrected stale cross-repo links across org READMEs. Shipped 2026-05-30 (latency 5d). Confidence: strong — verified by a link-check pass.
| Signal | Decision | Notes |
|---|---|---|
| Mobile bounce now exceeds desktop by >10pt — W4/W9 trigger fired | Act | Promoted from Watch (W4 → W9 → W12). |
| Conversion ~3.3% | Watch | Approaching the review’s 3.5%. |
| Aggregator noise (recurring) | Ignore | Same source as W7. |
Remediation. Fixed the mobile landing layout (CTA reachability + tap targets) so the first-screen CTA works on small viewports. Shipped 2026-06-07 (latency 6d). Closes the mobile-bounce watch chain opened in W4.
| Signal | Decision | Notes |
|---|---|---|
| Visitors reach reading paths but no visible next economic step | Act | Relevant to revenue milestone in #7. |
| Conversion holding ~3.5% | Ignore | At target; no change warranted — recorded so it isn’t re-litigated. |
Remediation. Surfaced the consult/support path as a clearly-labeled secondary link on the landing surface (no dark patterns). Shipped 2026-06-14 (latency 6d). Feeds the revenue ledger opportunity, which still requires a real external transaction to count.
This is the cycle run as part of executing issue #4.
| Signal | Decision | Notes |
|---|---|---|
| Conversion holding ~3.5% post-experiment-cut; no regression after #5 was CUT | Watch | Confirms the June 5 decision to refocus on top-of-funnel; trigger: 3-week decline below 3.0%. |
| Top-of-funnel (new-visitor) volume is now the binding constraint, not conversion | Act | Per the June 5 review, #5 was cut to refocus here; this is the first cycle to treat reach as the priority signal. |
| Returning-visitor rate flat-to-up | Ignore | Healthy; no action. Recorded to avoid re-litigation. |
Remediation (open). Open an owned top-of-funnel reach item: prioritize the distribution surfaces that the cut of #5 freed capacity for. Target ship date 2026-06-22 (next Monday), keeping the cycle within the 7-day SLA. Owner: maintainer. This item will be marked shipped in the rollup once the change is merged and live — not before.
Cycle status: signals classified, remediation opened and dated. The cycle closes when the reach item ships or is explicitly deferred with a review date.
ORGAN-V: Logos — Public Process. Pragma document: the honest account of decisions taken on real signals.