Record

Every test the partner approved, every call made before results came in, and what each new round of results changed.

Results on file

  1. Round 110 ads · results as of 2026-10-03 · loaded 10/3/2026, 4:16:07 PM
  2. Round 211 ads · results as of 2026-10-17 · loaded 10/3/2026, 4:16:07 PM

Drop the Ads Manager export here

Sample data
  • They don’t have to listen · In your wordsMaria, Nurse, Worcester · 41s · 10/17/2026awaiting review
  • No competition · In your wordsJames, Veteran, Pittsfield · 38s · 10/16/2026awaiting review
  • The 51% factAna, Teacher, Lowell · 33s · 10/16/2026scored
  • They don’t have to listenDev, Small-business owner, Quincy · 29s · 10/14/2026scored
  • No competitionRosa, Retired firefighter, Fall River · 44s · 10/12/2026running
Sample data
They don’t have to listen · StephenmadeCut to 20 seconds: keep the opening, the payoff line and the askHow to tell: Interest on the short cut vs. the full take, same budgetDecided 10/8/2026 against a score of 50The 51% fact · RuthapprovedOpen on the number: “51% of races had one name on the ballot”How to tell: Engage rate on the new opening vs. the coin flip, same audienceDecided 10/13/2026 against a score of 56No competition · DaniellerevisedSame script read by a volunteer from an uncontested districtHow to tell: Act rate and comments mentioning the districtDecided 10/14/2026 against a score of 44The 51% fact · DaniellesavedBurned-in captions in the first two secondsHow to tell: 3-second plays per impression, captions on vs. offDecided 10/15/2026 against a score of 39

Same formula (resonance-1), same reference (ref-1). “The score changed” means the number moved. “The conclusion changed” means the take moved across the line: above, below, or not distinguishable from the reference average.

25avg 5075Round 1Round 2
Hover a line to see the take; the scale and the average never move.

New in this round

Stephen · short cut scored 91 (84–94) against its full-length twin's 48 (40–56): clearly higher. It ran a shorter, later window, so treat this as the first measurement of the change, not its verdict.

TakeRound 1Round 2ChangeEngageInterestActReading
No competition, no accountability.… · Danielle43 (37–53)44 (40–49)+0.8+0+6-5Conclusion changed: not distinguishable from average to below the reference average
No competition, no accountability.… · Ruth49 (43–53)50 (47–53)+1.5-1-2+7Score changed, conclusion holds: not distinguishable from average
No competition, no accountability.… · StephenNot scored63 (42–84)Now scorable———First score: not distinguishable from average
51% of Mass. races have been uncon… · Danielle41 (37–49)39 (36–44)-1.7-2-1-2Score changed, conclusion holds: below the reference average
51% of Mass. races have been uncon… · Ruth68 (57–77)58 (51–65)-10.0-13-10-8Score changed, conclusion holds: above the reference average
51% of Mass. races have been uncon… · Ruth · no captions54 (41–67)49 (42–58)-5.3-6+5-17Score changed, conclusion holds: not distinguishable from average
51% of Mass. races have been uncon… · Stephen50 (45–53)48 (47–50)-1.7+1-8+3Conclusion changed: not distinguishable from average to below the reference average
When politicians don't have to com… · Danielle34 (21–44)35 (26–42)+0.5-7-3+10Score changed, conclusion holds: below the reference average
When politicians don't have to com… · Ruth74 (61–82)65 (56–73)-8.6-14-9-5Score changed, conclusion holds: above the reference average
When politicians don't have to com… · Stephen39 (32–49)48 (40–56)+9.2-11+3+30Conclusion changed: below the reference average to not distinguishable from average
Sample data

Partner: 2 of 4 right. The “no change” baseline: 3 of 4.

  • Locked 10/7/2026They don’t have to listen · Ruth: called higher than 66Right · came in higher
  • Locked 10/7/2026The 51% fact · Stephen: called same than 47Right · came in same
  • Locked 10/8/2026No competition · Ruth: called higher than 49Wrong · came in same
  • Locked 10/8/2026They don’t have to listen · Danielle: called lower than 37Wrong · came in same

Stand-in records until the partner's export arrives: the takes are synthetic readings of the library scripts and the paid results are generated in Meta's export format. Upload the real export on Reviews and every screen recalculates from it.