Video 1: https://www.instagram.com/reel/DbedKfdqnZS/?igsh=NTc4MTIwNjQ2YQ==

Video 2: https://www.instagram.com/reel/DbjTkhrKONd/?igsh=NTc4MTIwNjQ2YQ==

IMG_1072.png,IMG_1073.png,IMG_1074.png,IMG_1075.png,IMG_1076.png

Video 3: https://www.instagram.com/reel/DblXs68OatF/?igsh=NTc4MTIwNjQ2YQ==

IMG_1088.png,IMG_1089.png,IMG_1090.png,IMG_1091.png

Trial_Reels_Batch_1_Status.pdf

the true answer will be getting the algorithm gods to give us a chance to get some more scale but these insights aren't nothing

Insights from batch 1: Statement hooks beat vibes. The two copy-first hooks (R2, R3) skipped at ~62% vs 78.5% for the scenic "Summer break" opener, and R2 held 40% at 0:03 vs R1's 23%. R3's POV hook won the stop (84% at 0:01) but bled fastest after 0:04; R2's setup-and-turn ("Kids don't decide / Parents do.") converted best, 9m 22s watch time, 70% more than the others. Formula going forward: provocation in line one, turn or payoff by 0:03, event footage as proof, not opener. Two problems to solve: (1) engagement is dead, 5 reactions and zero comments/saves/follows across ~1,100 views, because DM CTAs are invisible to the algorithm, so test a comment-trigger CTA. (2) Trial distribution is going young and majority non-US (43% US, a third under 25), so our real signal with US moms is better than topline numbers suggest; judge on retention shape, not views.

CleanShot 2026-08-05 at 15.41.59@2x.png

That is crazy

Awesome

demographics weren't included for videos 1 and 2 in my first analysis because they were hidden in the screenshots

updating that portion

updated PDF and paragraph:

brain should be doing this shlep for me

Trial_Reels_Batch_1_Status_1.pdf

Insights from batch 1: Statement hooks beat vibes. The two copy-first hooks (R2, R3) skipped at ~62% vs 78.5% for the scenic "Summer break" opener, R2 held 40% at 0:03 vs R1's 23%, and R2's setup-and-turn ("Kids don't decide / Parents do.") converted best with 9m 22s watch time, 70% more than the others. But age data adds a confound: retention rank matches parent-band share exactly (35-44 viewers: R2 16.5%, R3 13.6%, R1 6.3%), so R2 either earned a better audience through early holds or benefited from one, likely both since trial distribution adapts to early signal. Honest read: R2 was the only reel that both stopped people and reached actual parents; R1's failure is mostly audience (half its viewers under 25). Going forward: (1) Keep statement hooks, land the turn by 0:03, event footage as proof not opener. (2) Audience mismatch is the primary bottleneck. Organic trial distribution won't reliably reach US moms 30-45, so either pay for real targeting or write hooks so parent-specific ("your third grader," "the iPad handoff") the algorithm can't route them anywhere else. (3) Engagement is dead (5 reactions, zero comments/saves/follows on ~1,100 views). DM CTAs are invisible to the algorithm; test a comment-trigger CTA like "comment GUIDE." Judge creative on retention shape, not views.