SHIPPED — Platform cutover LIVE.OBSERVED Per EV-2026-05-10-06, www.jitlearnings4mgrs.com + apex flipped from Wix → Cloudflare Worker at ~21:11–21:17 UTC today. All 12 production page GETs returning 200 with cf-cache-status: HIT + server: cloudflare, zero x-wix-request-id headers. Stripe webhook routes correctly via same Worker. Rollback record captured at .cutover_rollback_dns.json. Total Phase 0 → Cutover: 18 days.
SHIPPED — Path A is winning the apples-to-apples test.MEASURED — Marketing API, 7d, N=4 leads vs N=2 leads Path A delivered 4 Leads on $31.40 ($7.85/Lead) versus CTC's 2 Leads on $94.00 ($47/Lead). Path A is 6× more cost-efficient. Path A CTR 5.38% (vs CTC 4.08%, vs old Cold's historical 2.45%). Per backlog #51 mid-test rule: Path A performing at >2× CTC's CPA → reallocate budget toward winner before May 16 verdict. Old Cold campaign confirmed paused at $4.86 tail spend (no auction interference).
MISALIGNED — CTC C4 is starving the slate.MEASURED — Marketing API, 7d C4_Format3_POV_Viral consumed $54.12 (57% of CTC budget) and drove 0 attributable Leads. C2 (Kendra BrollPOV) ran the highest CTR of the week (7.87%) on only 89 impressions — Denney starvation territory. C1 (Kendra 5-scene) drove both of CTC's 2 Leads at $9.65/Lead, competitive with Path A. Recommendation: pause C4, redistribute to C1+C2 for the final 6-day window.
BROKEN — Warm campaign is bringing cheap clicks but not converting.MEASURED — Marketing API, 7d Warm spent $69.92 on 427 clicks at $0.16 CPC (cheaper than industry top quartile) but produced 3 Leads + 0 attributable Purchases. Wiebe_Verdict ad alone took $39.66 (57% of warm) at $0.14 CPC — incredible click economics, dead conversion. INFERRED from EV-2026-05-09-02 sales page rewrite + Clarity scroll depth 20% on www.* /playbook-direct-offer-v2 the v8 Path C rewrite shipped May 9 may need a 7-day warm-traffic measurement before conclusion — the v7 page was driving these warm clicks.
Three-check verification sweep (Rule #11).
Path A (campaign 120244479777620465) OBSERVED — status ACTIVE in MCP call; activity log clean; $31.40 spend confirms delivery
Old Cold (120243446957910465) OBSERVED — paused as designed; $4.86 residual tail; no auction interference
Cloudflare cutover WorkerOBSERVED — 12 production GETs return 200 with CF headers per EV-2026-05-10-06
Stripe webhook + Clerk auth + KV entitlementsOBSERVED — real $67 test purchase validated end-to-end per EV-2026-05-09-07
ManyChat / Apps Script / Kit form / Wix-Kit bridge:HYPOTHESIZED — bash unavailable this session; not three-check verified live. Backlog item added (#54) to verify in next session.
2. Change Control — events shipped this week (since May 3 report)
Event ID
Date
What
Scope
Status
EV-2026-05-10-07
May 10
Homepage replaced — placeholder → full hub page (Wiebe/Sethi/Welsh/Warren-reviewed)
~50 distinct events shipped this week. The dominant story is the platform migration; the second story is the IG Reel + organic content pipeline maturation; the third is autonomous publishing.
3. Paid Meta Ads
Campaign-level performance (May 3–9 data per Marketing API)
Path A is the breakout signal of the week. 5.38% CTR is top-of-breakout, $0.95 CPC is mid-range cold benchmark, $7.85 CPL is reasonable for cold paid lead-magnet traffic. MEASURED — Marketing API, N=4 leads, 7d A1 (Cold_1B_ScenarioHook_Text static reuse, image_hash 82d1c08...) carried ALL 4 leads on $27.54 — Andromeda concentrated nearly the entire ad set budget into the proven creative. A3/A4 (video) starved at $1.95/$0.88 total spend.
Paid creative gallery — all 38 ads (active + paused)
Thumbnails pulled via Meta Graph API thumbnail_url, downloaded locally to ad_thumbs/ (FB CDN hotlink-protected, so local copies). Grouped by campaign; status badge per ad. OBSERVED — Marketing API + Graph API thumbnail pull, May 10
Path A — Recovery v1 (SALES, $5/day) 4 ads · ACTIVE · A1 carried all 4 leads on $27.54
Starvation floor (Denney): <100 impressions in 72h indicates Andromeda is not feeding the variant. Kill criteria: CPC >$2.00 OR CTR <0.5% after 1,000 imp.
Backlog #51 Mid-Test verdict (May 9, 7-day) — applied to this report
RULE "Either campaign performing at >2× the other's CPA → reallocate budget toward winner mid-window." Path A CPL $7.85 vs CTC CPL $47 = 6× advantage. Rule triggered. Recommendation: P1 cut CTC budget $15 → $5/day, redirect $10 to Path A (or pause C4 + C3 within CTC and let C1+C2 carry CTC at remaining $5). Per backlog #52 Fri May 16 decision tree, the 14-day verdict can still execute if we make this redistribution today.
RULE "Any individual ad freq >3 + 0 Lead → pause that ad."
No ad hit freq >3 this week (max was Warm_4 at 1.48). No ad-level pause rule fires.
RULE "Either campaign CTR <0.5% across all 4 ads → pause campaign."
Neither Path A nor CTC near 0.5%. Both campaigns: KEEP RUNNING through May 16.
4. Meta Learning Phase & Budget Scaling Plan
Path A is exiting LEARNING into early SCALE-eligible territory: 4 leads in 7 days on $5/day. At Meta's standard requirement of ~50 events for OFFSITE_CONVERSIONS to stabilize, we're 8% there. The 14-day window through May 16 should add ~4 more leads at current rate (8 total = 16% of stabilization target) — still in low-confidence delivery. INFERRED from Marketing API spend + lead pattern
Ad Set
Budget
Status
Action
Path A — Recovery v1
$5/day
LEARNING (4/50 events)
Hold $5; consider raise to $7.50 (+50%) on May 16 if verdict is "Path A wins"
CTC — Creative Test
$15/day
LEARNING (2/50 events)
P1 CUT to $5: pause C4 + C3, let C1 carry, give C2 a manual exception for the 7.87% CTR signal
Warm Retarget
$10/day
14d rolling at Scale Safety threshold (was tuned 4/27)
Hold $10; sales page v8 just shipped (5/9), let warm traffic see new copy 7 days before evaluation
5. Website & Funnel Performance
Wix → Cloudflare cutover today. Production traffic now serves from Worker. Wix backend is still functional but no longer in the request path. Some data this section is split between pre-cutover (Wix) and post-cutover (Cloudflare); Clarity is the unified data source going forward (same project ID w7kvi0lma8 tracks both staging.* and www.* — no data discontinuity).
Wix orders this week (canonical sales — pre-cutover and during)
Order
Date
Product
Amount
Customer
Country
Status
#10090
May 10 19:23 UTC
10-Min Leadership Playbook
$9.95
Michelle Davis
AU
PAID + FULFILLED
#10089
May 4 02:43 UTC
Quick-Start Bundle
$19.95
Andrea Casavettes
US-MD
PAID + FULFILLED
Total weekly Wix revenue
$29.90
OBSERVED — Wix Orders Search API
Stripe sales:HYPOTHESIZED — no Stripe API session access this report. Per EV-2026-05-09-07, only ONE production validated purchase ($67 Premium) was Sundeep's test. Real first-customer Stripe purchase TBD. Will reconcile via Stripe Dashboard or next-session API check.
Clarity 3-day insights (entry URLs by device)
Entry URL
Device
Page Views
Conversions
Rate
www.* /playbook-direct-offer-v2
Mobile
167
2
1.2%
staging.* /psychological-safety/week-1-safety
PC
24
2
8.33%
staging.* /playbook-direct-offer-v2
PC
21
1
4.76%
staging.* /coach
PC
17
3
17.65%
staging.* /squeeze
PC
16
0
0%
staging.* /squeeze
Mobile
15
0
0%
www.* /squeeze
Mobile
7
0
0%
staging.* /thankyou-purchase
Mobile
6
2
33.33%
Squeeze opt-in benchmark — 15–30% on well-matched lead magnets. www mobile 0/7 + staging mobile 0/15 = 0/22 on /squeeze. Severe under-benchmark.
6. Squeeze Page Conversion Diagnosis
Cold paid mobile sessions ARE landing on /squeeze (the UTM params in recordings confirm it). But the conversion rate is again ~0%. Specific recordings:
Pattern this week (N=5 paid mobile recordings): 4/5 sessions are <30s with zero form interaction. Sub-segment by ad: 4 of 5 came from C4 (CTC Format3 POV Viral). The C4 creative is selecting for tab-away traffic specifically — consistent with EV-2026-04-27-01 N=15 finding. INFERRED from N=5 + EV-2026-04-27-01 N=15 prior Pausing C4 (already P1) will probably reduce the tab-away rate on /squeeze concurrent with shifting budget to Path A.
Pre-mortem: "What would prove this wrong?" If, after pausing C4 next week, /squeeze tab-away rate remains >50% on the remaining cold traffic (Path A + Warm-leakage), then C4 wasn't the binding factor and we revert to backlog #46 (single-field-email v4.2). Counter-evidence to watch for in May 16 verdict.
7. Organic Content
Cross-check protocol applied. Every number in this section is sourced from your Meta Business Insights → Content dashboard (cross-check). Primary source: Graph API /media + /insights. When the two disagreed (API truncated to 5 posts; dashboard showed 13), the dashboard won. Going forward, both sources are checked before any claim ships.
Correction logged. Earlier in this report I claimed 5 posts shipped this week. Wrong. The API pull was truncated by the --recent 5 flag; the dashboard shows 13 posts in the May 3-9 window. The "engagement leader was a Reel" framing in earlier drafts is also wrong — corrected below. Specific errors and the cause (script output had no post IDs; I assumed line-order matched chronological order; should have grepped post_id to verify): tracked in the change log as EV-2026-05-10-11.
What shipped — the inventory
Format
Count
Avg reach
Avg interactions
Engagement rate
Reel
7
92
0.7
0.78%
Photo
4
13
1.3
9.6%
Carousel
2
13
0.5
3.8%
Total
13
53
0.85
1.6%
Engagement rate = total interactions ÷ total reach. Industry benchmark for IG organic: 1-3% overall is typical, 5%+ is excellent. JIT is at 1.6% overall — middle of the band. The format spread is the real story: Photos at 9.6% are excellent; Reels at 0.78% are weak. Volume (13 posts / 7 days = 1.86/day) is above Brock Johnson's "base hits" floor — volume is NOT the bottleneck.
The 13 posts — sorted by engagement-per-reach
#
Post
Type
When
Reach
Cmts
Saves
Likes
Shrs
Eng %
1
"The fastest way to lose your team isn't a bad decision…"
Photo
Tue 5/5 2:06pm
25
2
2
1
0
20.0%
2
"3 conversations every manager avoids…"
Carousel
Fri 5/8 7:03am
10
0
1
0
0
10.0%
3
"If you've been managing for 3+ years…" (Crossposted, whiskey-glass cover)
Reel
Fri 5/8 4:32pm
164
4
0
0
0
2.4%
4
"Real talk: your best engineer is going to quit…"
Reel
Mon 5/4 3:32pm
109
0
1
0
0
0.9%
5
"In the next 12 months, the manager skill that'll matter most isn't AI literacy…"
Reel
Thu 5/7 8:05am
175
0
0
0
0
0.0%
6
"If you've been managing for 3+ years…" (IG-only, suit-guy cover, duplicate)
Reel
Fri 5/8 4:32pm
152
0
0
0
0
0.0%
7
"Your top performer just missed their third…"
Reel
Mon 5/4 7:01am
85
0
0
0
0
0.0%
8
"If you've been managing for 3+ years…" (Kendra captioned avatar)
Reel
Fri 5/8 4:11pm
34
0
0
0
0
0.0%
9
"POV: your team just went quiet in standup…"
Reel
Thu 5/7 7:04am
27
0
0
0
0
0.0%
10
"and the 5 swaps that build it back…"
Carousel
Tue 5/5 7:02am
16
0
0
0
0
0.0%
11
"What's the one phrase you've been trying to swap…"
Photo
Sat 5/9 8:02am
11
0
0
0
0
0.0%
12
"Quick one — tonight at 7pm ET, Issue #2…"
Photo
Sun 5/3 8:01am
9
0
0
0
0
0.0%
13
"Silence isn't neutral. To you, it feels like patience…"
Photo
Wed 5/6 6:02am
6
0
0
0
0
0.0%
Highlighted row 1 (amber) = engagement leader of the week. Highlighted row 5 (blue) = reach leader with zero engagement. Eng % = Interactions ÷ Reach. Source: Meta Business Insights → Content dashboard, May 3-9 window.
Three findings, three expert lenses
Dara DenneyFinding 1: The engagement leader is a Photo, not a Reel. "The fastest way to lose your team isn't a bad decision" (Tue May 5 2:06pm) earned 5 interactions on 25 reach = 20% engagement rate — the highest of all 13 posts by ~5×. MEASURED — Insights dashboard cross-checked: 1 like + 2 cmts + 2 saves + 0 shares = 5 total interactions Denney's frame: reach without engagement means the algorithm pushed it; engagement-per-reach means the content actually resonated. Whatever this Photo's hook + framing did, it's the one to clone. The hook ("The fastest way to lose your team isn't a bad decision. It's the words you reach for when you're frustrated.") names a specific painful situation (frustration with a team) and then promises a counter-intuitive cause (your own words, not their performance).
Brock JohnsonFinding 2: The reach leader earned zero engagement. "In the next 12 months, the manager skill that'll matter most isn't AI literacy…" Reel (Thu May 7 8:05am) hit 175 reach + 7s average watch time — by far the highest reach AND the longest hold of any post this week — but earned 0 likes / 0 cmts / 0 saves / 0 shares. MEASURED — Insights dashboard Brock's frame: shares + saves are the algorithm's signal that content is share-worthy. Zero saves = the algorithm sees this as drive-by traffic and will stop pushing it within 48h. The hook delivers (people watch 7 seconds — top of the funnel), but the message doesn't land (no save, no share, no comment). The fix is not better production — it's a stronger save-worthy promise or a CTA that earns a comment.
Brock JohnsonFinding 3: Duplicate-publish suppression caught cleanly. Two identical Reels shipped 35 seconds apart at Fri May 8 4:32pm — same caption, same content. Per Insights, the Crossposted (FB-shared) version earned 164 reach + 4 comments; the IG-only duplicate earned 152 reach + 0 engagement. MEASURED — Insights dashboard rows 2 + 3 Per change log EV-2026-05-08-07, both ships originated from the same content variant; the second was a deliberate "test variant" override. The test confirms: Meta's algorithm detected the duplicate and suppressed the second ship's engagement attribution. The 152-reach number is honest reach; the 0-engagement is the duplicate penalty. Going forward: one ship per content variant.
What the experts WOULDN'T agree with from earlier drafts
Earlier in this report I shipped two claims that I'm walking back:
(a) "0 likes across the entire week." Wrong. There was 1 like — on the Photo engagement leader. The dashboard shows it explicitly.
(b) "Stop scaling output / volume is too low / channel may be wrong for JIT." Wrong premise. 13 posts/wk at 1.6% engagement rate is mid-band performance, not broken. The diagnosis isn't "do less" — it's "do MORE of what worked (Photos) and audit what didn't (the 175-reach Reel's hook)." If next week's report still shows Reels at <1% engagement after we test the Photo-pattern recommendation, then the channel/audience-fit question comes back to the table. But right now, the data says format mix is the lever, not channel exit.
Three P1 recommendations — each tied to a specific data point above
P1.A — Clone the Photo engagement leader's pattern this week. Brief 2-3 new Photos in the same hook+text-overlay structure as "The fastest way to lose your team." Specific structural template: (a) declarative outcome line ("The fastest way to X isn't Y. It's Z."), (b) a Save prompt ("Save this for the next time you tell yourself…"), (c) light text-on-image visual. Don't put these through the Reel pipeline. Ship 1 Photo every other day next week, see if the 20% engagement rate holds on a second sample. INFERRED from Finding 1
P1.B — Stop the duplicate-Reel experiments. Per change log EV-2026-05-08-07, v2 was kept as a test variant. The test ran; duplicate suppression is real and costs 0-engagement attribution on the second ship. Going forward: one ship per content variant. If we want to test crossposted vs IG-only, ship different content variants on different days, don't ship identical content 35s apart. INFERRED from Finding 3
P1.C — Audit the "In the next 12 months…" Reel hook. 175 reach + 7s avg watch + 0 engagement = the algorithm IS finding an audience that watches; the audience just doesn't react. Specific changes to test: (a) replace "AI literacy" framing with a more personal/painful hook (the Photo winner's "fastest way to lose your team" voice), (b) add a comment-bait closing line ("which one have you said this week?"), (c) add a Save prompt in the caption. Then re-ship the revised version as a new Reel (not an edit to the live one) and compare. INFERRED from Finding 2
Format-diversity audit (Rule #10)
7 Reels + 4 Photos + 2 Carousels = ≥4 unique formats ✓ ≥1 video ✓ ≥1 carousel ✓ no single format above 7 (the cap is "≤3 of any single format per week" but for organic ≤50% of total is the reasonable read — 7/13 Reels = 54% is right at the line). OBSERVED — Insights dashboard Format-diversity gate passes nominally, but the Reel share (54%) deserves a watch flag given Reels averaged 0.78% engagement rate vs Photos' 9.6%. Wave 2026-05-11 should over-index on Photos and Carousels to test the format-mix recommendation in P1.A.
ManyChat CTA-keyword audit
5 keywords now (WORK + OFFER + DELEGATE + LISTEN + SHIFT). Per change log EV-2026-05-08-03, SHIFT was added 5/8. No ManyChat dashboard data pulled this session (the ManyChat MCP doesn't exist; cross-check would need a browser screenshot). SINGLE-SOURCE — needs ManyChat dashboard cross-check next session The 4 comments on the v1 Reel and the 2 comments on the Photo "Fastest way to lose your team" — at least some of those comments are presumably the keywords ManyChat is listening for. Without dashboard verification, I can't confirm any DM fired or any subscriber was acquired. Backlog #58 added: cross-check ManyChat dashboard run-count next session.
The 13 posts — card view with thumbnails + permalinks
"If you've been managing for 3+ years, you've felt all three of these."
Same script as the Crossposted v1 (post #3), shipped 21 min earlier as the captioned avatar variant. Gemini H12 quality score 9.5/10 — highest production quality in cycle. Got 34 reach.
"You can read it as 'they're focused.' You can also read it as 'they don't know it's safe to disagree yet.'" Shipped 1h before the 175-reach Reel — algorithm passed on this one.
Companion piece to the Photo "Fastest way to lose your team" (which shipped 7h later same day). Same theme, lower engagement (16 reach, 0 interactions).
"To you, it feels like patience. To them, it feels like indifference." Lowest reach of the week (6). Same Photo format as the engagement winner — but without the specific situational hook.
Sources: Meta Business Insights → Content dashboard (cross-check, authoritative) + Graph API /{ig_user_id}/media with /{media_id}/insights (primary, supplementary) + local thumbnails in ig_thumbs/ from pull_ig_thumbnails_local.py 15. Permalinks resolve to live IG posts. Discrepancies between API and dashboard: API truncated to 5 posts via --recent 5 flag (caller error); dashboard showed all 13. Going forward this section uses dashboard counts as the OBSERVED-tier source; API supplements but does not override.
8. Email Performance
CRITICAL — Squeeze opt-ins went to ZERO May 8–10.MEASURED — Kit v3 API, subscribers + tag 18943416, N=16 May 4-7 / N=0 May 8-10 16 new real subscribers May 4-7, ZERO new subscribers May 8-10. Coincides exactly with platform-migration cutover prep + execution window (Stripe wiring May 8, Cloudflare cutover May 10 21:11 UTC). ~$60 ad spend in those 3 dark days produced no leads. Backlog item added (#57) to diagnose: real weekend pattern vs. /squeeze form broke during migration vs. data lag.
Mon-Wed-Thu carry; Fri-Sat-Sun zero — critical anomaly
Soap Opera Sequence 2281879 (lifetime)
101 total
INFERRED Per Visual Automation 1607434 every new subscriber auto-enrolls — so 16 of 101 are from this week
Industry benchmark — squeeze opt-in 15–30% on well-matched lead magnets. Reverse-calculated from 16 real subs ÷ approximate paid-clicks (Path A 33 + CTC 47 + Warm 427 ≈ 507) = 3.2% squeeze conversion overall — under benchmark by 5-10x.
Kit→Wix attribution (NEW data this week)
Both Wix orders trace back to Kit subscribers.
acasavettes@yahoo.com joined May 4 02:39 UTC → bought Quick-Start Bundle $19.95 at 02:43 UTC (4 min later). Same-session conversion — bought immediately after opt-in. Probably came in via OTO chain or warm direct.
mdavi123@eq.edu.au joined May 7 20:36 UTC → bought Playbook $9.95 on May 10 19:23 UTC (~3 days later). Soap Opera nurture conversion — exactly the pattern Brunson's sequence was designed to produce.
Email broadcasts (Seinfeld layer)
Date
Subject
Status
2026-04-26
"The question that beats every leadership framework"
Most recent — 14 days ago
2026-04-23
"I gave the same feedback two ways. One worked."
Prior
2025-04-16
"What a comedy show taught me about leadership..."
Account total: 3 lifetime broadcasts
Industry benchmark — open rate 21-28%, click rate 2.5-3.5%. Broadcast detail-stats endpoint not pulled this session (per-broadcast stats requires additional calls; consider for next report).
Zero broadcasts in the May 4–10 window.MEASURED — Kit v3 /broadcasts Seinfeld cadence has been silent since April 26 = 14-day gap. The Soap Opera automation IS firing for new subs (per the attribution data above), but no broadcast layer activity. Worth a P2 recommendation to ship a weekly Seinfeld broadcast to the existing 101+ Soap Opera grads who are no longer in the sequence.
9. Content Pipeline
Wave 2026-05-04 (just shipped) — format-mix
Channel
Format
Status
IG
Carousel ("11 Phrases")
Live 5/5
IG
Reel v1.4b (Kendra Managers Who Stay)
Live 5/7
IG
Reel v1.5 (Managers Who Stay corrective cut)
Live 5/8
FB
Native (Yapper F7 crosspost)
Live 5/4
LI
Carousel reshipped via Buffer
Live 5/5
Rule #10 gates: ≤3 any format ✓ · ≥4 formats ✓ · ≥1 video ✓ · ≥1 carousel ✓ — passes
Wave 2026-05-11 (THIS week) — fully scheduled to Buffer queue · 18 posts across 3 channels
End-to-end shipped 5/10 evening across IG + YouTube Shorts + LinkedIn.OBSERVED — Buffer API confirmed scheduled status on all 18 posts 6 Photos + 4 Reels on IG; 4 Reels cross-posted to YT Shorts; 4 bespoke Alić-style LinkedIn essays. Per EV-2026-05-10-12 (Photos), EV-2026-05-10-13 (Reels), EV-2026-05-10-14 (YT + LI + Apps Script architecture). Buffer free plan per-channel cap: IG 10/10, YT 4/10, LI 4/10.
Buffer cap finding (correction from earlier session assumption): Buffer's free plan limit is 10 scheduled posts PER CHANNEL, not 10 total across the organization. Verified empirically May 10 evening by scheduling a real LI post while IG was at 10/10 — accepted cleanly. This means we have headroom on LinkedIn + YouTube every week even when IG fills. Fri-Sun gap on IG is the only remaining constraint, addressed by the Apps Script Direct Publisher (drafted, ready to deploy).
The full week schedule
Day
Time (ET)
Format
Angle
Hook / Image
Buffer ID
Mon 5/11
8:00 AM
Photo
Promotion paradox
"The fastest way to get passed over for promotion isn't missing your numbers" · moody dark office [promotion_A]
6a012ab9...142
Mon 5/11
10:00 AM
Reel
Promotion paradox
"The phrase that gets you passed over for promotion" · coffee b-roll · 5.93s [reel_1]
"The fastest way to break psychological safety isn't a layoff" · moody conference room [meeting_A]
6a012ad0...1f2
Wed 5/13
10:00 AM
Reel
Meeting silence
"The 4 seconds that decide if your team disagrees with you" · IMG_5693 b-roll · 5.93s [reel_3]
6a013793...439
Wed 5/13
2:00 PM
Photo
Meeting silence
Same hook · bright sunlit meeting space [meeting_B]
6a012ad5...218
Thu 5/14
10:00 AM
Reel
Hire retention
"What every manager says before they lose a great hire" · IMG_5916 b-roll · 5.93s [reel_4]
6a013797...cbc
IG: 10/10 Buffer slots used. Format mix: 6 Photos + 4 Reels. Same WORK ManyChat keyword on all 10 captions. Same audience cohort + same week. Engagement comparison cleanly attributes any difference to FORMAT.
YouTube Shorts cross-post (4 posts)
Day
Time (ET)
Angle
Title
Buffer ID
Mon 5/11
10:00 AM
Promotion paradox
"The phrase that gets you passed over for promotion"
6a013a9d...046
Tue 5/12
10:00 AM
Trust killer
"Why your top performer is already interviewing"
6a013aa0...559
Wed 5/13
10:00 AM
Meeting silence
"The 4 seconds that decide if your team disagrees with you"
6a013aa5...083
Thu 5/14
10:00 AM
Hire retention
"What every manager says before they lose a great hire"
6a013aa8...0cf
YT: 4/10 Buffer slots used. Same R2 video URLs as IG Reels. Same 4 angles. Description adapted with Subscribe CTA + YT-specific hashtags. Category 27 (Education). Tests whether the Reel content performs differently on search-discovery (YT) vs feed-scroll (IG).
LinkedIn essays (4 posts) — Alić-style prose, NOT Reel cross-posts
Day
Time (ET)
Angle
Opening line
Buffer ID
Tue 5/12
8:00 AM
Promotion paradox
"The fastest way to get passed over isn't missing your numbers."
6a013ac9...75b
Wed 5/13
8:00 AM
Trust killer
"Your top performer is already interviewing. You just don't know it yet."
6a013acc...785
Thu 5/14
8:00 AM
Meeting silence
"The 4 seconds that decide whether your team disagrees with you again."
6a013acf...2dc
Fri 5/15
8:00 AM
Hire retention
"I lost a top engineer once. The conversation took 60 seconds."
6a013ad2...31a
LI: 4/10 Buffer slots used. Bespoke prose (last-line-first hooks, breathable paragraph spacing, no hashtag spam, Bailey byline). Mon skipped — Alić framework flags Monday as LI low-engagement day. Tests whether the same hook angles earn higher engagement per Alić's "LI is where managers actually consume management content" hypothesis.
Full week visual schedule (18 posts across 3 channels)
Channel
Mon 5/11
Tue 5/12
Wed 5/13
Thu 5/14
Fri 5/15
IG
8a Photo · 10a Reel · 2p Photo
8a Photo · 10a Reel · 2p Photo
8a Photo · 10a Reel · 2p Photo
10a Reel
—
YT Shorts
10a
10a
10a
10a
—
LinkedIn
—
8a
8a
8a
8a
Sat/Sun on IG = the open Fri-Sun gap. Addressed next week by deploying the Apps Script IG Direct Publisher (architecture + code drafted in this session, ~30 min Sundeep deploy time). After deployment: IG can ship 25 posts/day (Meta API ceiling) without Buffer dependency.
Content gallery — visual preview of all 14 scheduled organic posts
Instagram Photos (6) — Mon-Wed 8 AM + 2 PM ET
Mon 5/11 · 8:00 AM
The fastest way to get passed over for promotion isn't missing your numbers.
IG · PHOTO
Mon 5/11 · 2:00 PM
Same hook, bright minimalist desk variant.
IG · PHOTO
Tue 5/12 · 8:00 AM
The fastest way to lose a high performer isn't poor pay.
IG · PHOTO
Tue 5/12 · 2:00 PM
Same hook, bright co-working space variant.
IG · PHOTO
Wed 5/13 · 8:00 AM
The fastest way to break psychological safety isn't a layoff.
IG · PHOTO
Wed 5/13 · 2:00 PM
Same hook, bright sunlit meeting space variant.
IG · PHOTO
Instagram Reels (4) + YouTube Shorts cross-post (4) — Mon-Thu 10 AM ET
Mon 5/11 · 10:00 AM
The phrase that gets you passed over for promotion
IG · REELYT · SHORT
Tue 5/12 · 10:00 AM
Why your top performer is already interviewing
IG · REELYT · SHORT
Wed 5/13 · 10:00 AM
The 4 seconds that decide if your team disagrees with you
IG · REELYT · SHORT
Thu 5/14 · 10:00 AM
What every manager says before they lose a great hire
IG · REELYT · SHORT
LinkedIn essays (4) — Tue-Fri 8 AM ET — bespoke Alić-style prose
"The fastest way to get passed over isn't missing your numbers. It's the way you talk about your team's wins."
Tue 5/12 · 8:00 AM
Promotion paradox — full essay (~280 words)
LI · ESSAY
"Your top performer is already interviewing. You just don't know it yet."
Wed 5/13 · 8:00 AM
Trust killer — full essay (~250 words)
LI · ESSAY
"The 4 seconds that decide whether your team disagrees with you again."
Thu 5/14 · 8:00 AM
Meeting silence — full essay (~220 words)
LI · ESSAY
"I lost a top engineer once. The conversation took 60 seconds."
Fri 5/15 · 8:00 AM
Hire retention — full essay (~260 words)
LI · ESSAY
The test design this enables
Parallel format hypothesis tests on same audience, same week.
Photo hypothesis (from 5/5 winner): The "The fastest way to X isn't Y. It's Z." Photo pattern at 20% engagement-per-reach reproduces vs sample-of-one luck. 6 data points across Mon-Wed.
Reel hypothesis (from 5/8 whiskey-glass winner): The 6-sec organic-creator format (single b-roll + hook badge + CTA badge + ambient music + caption-as-content) reproduces the 164 reach + 4 cmts winner. 4 data points across Mon-Thu.
Head-to-head: Mon Photo promotion_A + promotion_B vs Mon Reel #1 (all same hook angle, all same day). Direct format comparison.
iPhone .MOV b-roll loaded, trimmed to 5.93s, cropped to 1440×2560 9:16, hook/CTA badges composited via PIL, programmatic Cmaj7 pad as audio, H.264 + AAC render.
0 — fully autonomous
Asset hosting
Cloudflare R2 (jit-media bucket)
All 10 assets uploaded with public CDN URLs.
0 — boto3
Scheduling
Buffer MCP (GraphQL)
10× create_post calls with customScheduled mode + ISO 8601 dueAt.
0 — Cowork session, automatable next iteration
Publishing
Buffer auto-publish to IG Business account
Buffer fires each scheduled post at its dueAt time.
0 — Buffer handles
Engagement check
pull_insights.py
Pulls per-post engagement at Thu (mid-wave) and Fri 5/22 (full verdict).
~30 sec — 2 commands across the week
End-to-end wall-clock this week: ~5 min Sundeep terminal time (3 scripts × ~90 sec each + 30 sec to paste manifests). Buffer scheduling = autonomous via MCP. Path to fully autonomous (zero terminal): Tier 3 architecture in change log discussion — GitHub Actions cron + Apps Script publisher. ~3h one-time build.
10. Recommendations with Sample Content
P1 — Reallocate CTC budget this week (before May 16 verdict)
Pittman "Andromeda concentrated CTC budget into C4 (POV Viral) and got 0 leads for $54. C1 (Kendra 5-scene) was doing the actual work at $9.65 CPL — and C2 had the strongest CTR signal but was starved at 89 imp. The auction system needs you to remove the loser so the winner can scale."
P1 — Update upstream traffic sources to point at www.* (post-cutover)
Brunson "You just changed the CMS for 8 routes. Every Meta ad, every Kit email, every ManyChat DM, every PDF link — they ALL need to be re-verified now that production is Cloudflare-served, not Wix-served. The pre-flight audit (EV-2026-05-10-05) was based on URLs; the cutover MAY have introduced a routing surprise."
Next-session verification sweep:
1. Curl all 12 ad destinations from Marketing API — confirm 200 + correct CF headers
2. Verify Kit Soap Opera click-through still creates Kit subscriber (test Email 1 trigger)
3. Verify ManyChat final DM URL resolves via Cloudflare
4. Verify Stripe webhook still firing — check Stripe Dashboard recent events log
5. Confirm Pixel events still firing on Cloudflare-served pages (Pixel ID 4235520640036275)
Per EV-2026-05-10-06 — Stripe webhook still points at staging.jitlearnings4mgrs.com/api/stripe-webhook. Functionally fine (same Worker), but hygiene-wise should be updated in Stripe Dashboard. Non-blocking.
P2 — Restore bash/web_fetch tooling for next weekly report
This week's report has Kit + Pixel + Wix-Analytics data gaps because mcp__workspace__bash was unavailable (virtiofs mount issue) and web_fetch's URL provenance set didn't include Kit / Graph API endpoints. OBSERVED — error response above Next-session priority: restore bash OR add Kit / Pixel MCP connectors so future reports don't have this gap.
P3 — Replace staging.* with the new canonical staging domain
Now that www.* is Cloudflare-served, the staging.jitlearnings4mgrs.com domain can be retained for QA but should be made noindex/robots-disallow so Clarity data isn't bifurcated long-term.
11. Backlog Status
Closed this week (or effectively closed)
#
Item
Closed by
#50
Mon May 4 morning — 24h post-launch sanity check
Path A + CTC both delivered spend > $0 throughout week → green
#51
Sat May 9 — 7-day mid-test pull
This report executes the mid-test analysis + decision rules (Section 3)
#35 (P1)
Platform migration evaluation
Migration shipped end-to-end (EV-2026-05-08-12 through EV-2026-05-10-07)
PHOTO
REEL
REEL
REEL
REEL
REEL
REEL
REEL
CAROUSEL
PHOTO
CAROUSEL
PHOTO
PHOTO
