Benchmark20
How many ads should you test a month?
About one ad in twenty becomes a winner. Most brands do not make enough to find one.
6 min readRead
OpenAffect Insights
Notes for people who run the ads at an ecommerce brand.
Every piece here answers a question an operator has this week, with the numbers sourced and dated and at least one thing we made ourselves: an angle map, a calculator, a cut list. 4 practical pieces so far, plus a 19-brief research library on how our scoring instrument works, for the buyer who wants to look under the hood.
Start here
All practical pieces
Benchmark21
Your best ad just died. Now what?
Rule out four false alarms, read Meta’s own Delivery column, then iterate before you replace it.
7 min readRead
Opinion22
AI UGC vs. hiring creators, priced honestly
We make AI ads. Here is the honest cost table, and where a human still wins.
7 min readRead
Playbook23
How to read a competitor’s ad account
Sort by what has run longest, tag each ad’s angle, count. An hour in the Meta Ad Library finds your opening.
7 min readRead
The research library
How the scoring instrument works, for the buyer who wants to look under the hood
19 longer briefs on brain encoding models, neuroforecasting, synthetic respondents, creative testing benchmarks and the calibration of our own predictions against real ad performance. Written before the agency existed, kept because the instrument is the same. Start with the calibration study.
- Thought leadershipThe four signals of human response11 min
- Thought leadershipFocus groups at scale are still focus groups13 min
- Thought leadershipWhy we chose to publish our research7 min
- ScienceWhat is neuromarketing in 2026? (And why most of what you have read is wrong)12 min
- ScienceTRIBE v2 explained: what Meta's brain encoding model actually does12 min
- MarketSynthetic respondents vs focus groups: what actually replaces what in 202613 min
- MethodologyHow to test a video ad before you publish it10 min
- MarketThe complete guide to creative intelligence platforms in 202614 min
- MarketHow much does a focus group cost in 2026? (With real numbers)9 min
- MethodologyA framework for pre-publication content scoring13 min
- ScienceEmily Falk's neural focus group, fifteen years later9 min
- ScienceWhat is neuroforecasting?7 min
- MarketThe Meta Ad Library is a gold mine. Here is how to actually use it.11 min
- MarketCreative testing benchmarks: what "good" looks like across paid social14 min
- Thought leadershipPredicting the future is signal integration, not magic13 min
- ResearchCalibrating neural engagement predictions against Meta Ad Library performance15 min
- ResearchA technical review of brain encoding models for content prediction16 min
- ResearchReverse inference, twenty years on: why forward prediction is the right methodology12 min
- ResearchNeural Content Intelligence: scoring video before you publish20 min