Skip to content

Instantly share code, notes, and snippets.

@gsingal
Created June 8, 2026 12:30
Show Gist options
  • Select an option

  • Save gsingal/4c3cea6c63bcbf42f8f3b2af11f315c1 to your computer and use it in GitHub Desktop.

Select an option

Save gsingal/4c3cea6c63bcbf42f8f3b2af11f315c1 to your computer and use it in GitHub Desktop.
Weekly Launch Review — 2026-06-08

Weekly Launch Review — Week of 2026-06-08

93 monitoring launches reviewed. 10 deep-dived (Tier-1), 83 orchestrator-assessed.

TL;DR

  1. Pro Plan experiment (abc-pro-plan-pricing-v3) is structurally unresolvable — control deprecated Day 6, only 47 exposures. Pro adoption only 3.5-7.9% of purchasers. Needs human decision: graduate best variant or re-run.
  2. Blog migration working — impressions +49%, clicks +53% in recent 14d vs pre-migration. But KPI pipeline is broken (collecting garbage numbers). Fix metric collection, set real targets.
  3. Internal linking SEO shows WORSE rankings on 7/8 city pages — opposite of hypothesis. Needs human review before continuing.
  4. 5 launches ready to graduate — map-markers, address-autocomplete, mobile-address-input, email-pii-removal, free-listing-deferred-activation. All 30-75 days monitoring with zero regressions.
  5. Reactivation win-back email for expired free listings is dead — queries wrong status (STATUS_INACTIVE vs STATUS_EXPIRED). Revenue leak.

Top Actions (7)

  1. [DECISION] Pro plan pricing v3: decide experiment fate (graduate best variant vs re-run with control) — @gsingal
  2. [DECISION] Internal linking SEO: 7/8 city pages have worse rankings. Continue investing or pivot approach? — @gsingal
  3. [BUG P2] Free listing reactivation email queries STATUS_INACTIVE instead of STATUS_EXPIRED — never sent — @ApxSnowflake
  4. [BUG P2] Quarterly billing variant has 3x payment failure rate (6.6% vs 2.1%) — investigate decline codes — @ApxSnowflake
  5. [CHORE] Graduate 5 launches: map-markers-grey-blocks, address-autocomplete-mobile-styling, mobile-address-input-styling, email-pii-removal, free-listing-deferred-activation — @gsingal
  6. [SEO] Optimize FF vs RR blog post SERP snippet — 19k impressions at 0.3% CTR, biggest quick win — @gsingal
  7. [FEAT] Registration wall loses 52% of inquiry users — highest-leverage conversion optimization — @ApxSnowflake

Graduation Summary

Launch Days Recommendation Confidence
map-markers-grey-blocks 55 graduate 0.90
address-autocomplete-mobile-styling 55 graduate 0.85
mobile-address-input-styling 55 graduate 0.85
email-pii-removal 75 graduate 0.92
free-listing-deferred-activation 61 graduate 0.95
free-listing-expiration-fix 61 graduate 0.85
posthog-messaging-instrumentation 61 graduate 0.95

Tier-1 Deep Dives

1. abc-pro-plan-pricing-v3 — needs-human-input (confidence: 0.35)

Experiment structurally unresolvable. Control deprecated Day 6 with only 47 exposures (PostHog flags not-enough-exposures). All stats null. Descriptive data: Pro adoption 3.5-7.9%, Standard dominates at ~45%. Quarterly billing captures 14.1% when offered. pro-new-design underperforms old layouts by ~6pp on Plan Page -> Checkout. Payment failure rate 3x higher in quarterly variant.

2. blog-migration — change-fix-kpi-collection (confidence: 0.80)

Product is working (impressions +49%, clicks +53% recent 14d). KPI pipeline severely broken: GSC metric swings 16k-149k due to collection method; target of 100 is a smoke placeholder. blog-visitor-conversion mislabeled as rate but stores raw counts. FF vs RR blog post has 19k impressions at 0.3% CTR (biggest optimization opportunity).

3. internal-linking-seo — needs-human-input (confidence: 0.45)

7/8 city pages have WORSE rankings post-launch (opposite of hypothesis). Only Austin improved. Average position degraded from 14.3 to 19.2 across targeted city pages. "Completion rate" metrics are actually raw 7-day rolling counts mislabeled as percentages.

4. posthog-messaging-instrumentation — graduate (confidence: 0.95)

All 10 events firing reliably. Modal-to-submit: 81.1%. Registration wall loses 52% of users (biggest optimization target). Similar-listing cross-sell converts at 14%. Smoke target (1) exceeded at 7,000/day. KPI pipeline has intermittent spike anomaly.

5. map-markers-grey-blocks — graduate (confidence: 0.90)

Fix verified working. No recurrences in 55 days. Representatives deployment-2026-04-14 cluster (24 launches). Completion-rate metrics are mislabeled (counts, not percentages).

6. address-autocomplete-mobile-styling — graduate (confidence: 0.85)

Mobile styling fix working. Playwright visual tests confirm layout. No regressions in 55 days.

7. mobile-address-input-styling — graduate (confidence: 0.85)

Mobile address input fix working. Visual regression tests pass. No issues in 55 days.

8. email-pii-removal — graduate (confidence: 0.92)

PII absent from all email templates. Three defense layers active. 75 days, zero regressions. Test fixtures broken (stdClass vs Eloquent) but templates are correct. email-pii-audit-pass metric never operationalized.

9. free-listing-deferred-activation — graduate (confidence: 0.95)

Three defense layers, 27 tests, 61 consecutive days at 0.0 premature activations. Sibling issue #3556 (paid-plan variant) still open and unmonitored.

10. free-listing-expiration-fix — graduate (confidence: 0.85)

Core fix solid. 11 tests, 61 days no recurrences. Found: reactivation win-back email queries wrong status (STATUS_INACTIVE vs STATUS_EXPIRED), never sent.

First-Appraisal (14 launches, too early to evaluate)

admin-downgrade-pro-premium, lcp-optimization, travel-nurse-landing, prebook-duplicate-charges-fix, pii-leak-listing-pages-fix, null-location-fix, admin-user-edit-fix, register-ip-guard, subscription-type-guards, 404-page-fixes, 4259-messaging-logged-out, 4338-systemic-checkbox-trait, 4316-room-under-review-uncheck, 4319-labs-nav-kpi-smoke — all have null launch dates or launched within last 7 days. No data to evaluate. Will review next week.

Anomaly (2 launches)

  • 4648-keyword-fraud-detector (launched May 6) — fraud keyword detection on active listing edits. No metrics registered. Working as designed (RiskReview rows being created) but invisible to KPI system.
  • 4554-off-platform-contact-detection (launched May 6) — chat + listing edit fraud detection. Same: working but unmonitored by KPI.

Off-track (18 launches)

All 18 have null objectives AND null KPI metrics. These are registered features that were never properly instrumented. Most are bugfixes or infrastructure from April/May 2026. Notable: pro-video-tour, attribution-phase-a, posthog-session-recording-re-enable are feature launches that should have metrics. The off-track bucket is dominated by KPI registration debt, not feature failures.

Stable-green (11 launches)

All 11 have null metrics and are showing no regressions. Continue monitoring: fraud-queue-v1.5, churnkey-hmac-null-guard, geo-radius-listing-price-stats, 4317-breadcrumb-fallback, listing-active-regression, stripe-link-dual-email-review, mobile-safari-419-handler, messaging-dashboard-timezone, rocket-loader-compat, saveform-ownership-check, edit-listing-500-fix.

Other (3 launches)

fraud-post-payment-phone-suppress, 4617-riskreview-v1-production, blog-content-sitemap — specialized launches, no standard KPI pattern applies.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment