Meta Conversion Lift needs 2-3 weeks: freeze holiday hooks by Nov 6

Meta's holiday measurement note says that Conversion Lift results reach statistical significance in 2 to 3 weeks in high-volume holiday periods, against 6 or more weeks in quieter quarters (Meta for Business, read 2026-10-07). Counted back from Black Friday on Nov 27, 2026, a 3-week test must start by Nov 6. That means your hook variants must be finished and frozen by then.

The date math

Nov 27 minus 21 days is Nov 6, and Nov 6 is 30 days after today, Oct 7. Today is Oct 7, so you have 30 days to make the variants. If the test runs 2 weeks, the last start date is Nov 13, but a plan built on the best case leaves no room for a failed render.

Why a freeze matters

A Conversion Lift study compares people who could see your ads with people who could not, so it measures incremental results rather than clicks. A hook test inside that design needs the creative to stay the same for the whole run. If you replace the clip in week two, the result no longer says which hook caused the lift.

Three frozen variants, priced

Plan three variants, one body and one end card. A Gemini Omni Flash 1.1 hook of 3 seconds at 1080p is 3 x $0.1875 = $0.5625. Three hooks are $1.6875. A 9-second body at 1080p is $1.6875 and one end card of 3 seconds is $0.5625. Each final cut is one Timeline render at $0.10, so three cuts are $0.30.

The sum is $1.6875 + $1.6875 + $0.5625 + $0.30 = $4.2375, so about $4.24. Hold a retake budget of the same size, and the whole freeze costs under $9.

Keep a record

Name the output of each render with the variant, for example hookA-v1.mp4, and keep the job ids with it. When the test ends you can say exactly which file ran. Do not re-render a frozen variant. If a clip needs a fix, make it a new variant and start the test again.

A short holdout of untouched creative also helps. Keep the previous best ad running in one arm so the lift has a real comparison.

Also decide in advance which number ends the test. Meta's note ties the 2-to-3-week window to significance, not to a fixed day, so write down the stop rule and the metric before the start date and do not change either after the first read.

Write the variant names into the ad names in Ads Manager as well, so the lift report reads cleanly. Last, the same note reports that one brand with proof of incrementality received a 20% budget increase. You cannot count on that for your account, but it explains why the proof is worth three weeks: a measured lift is the argument for more budget before the Cyber Monday push.