Black Friday 2026 Creative: A Production Calendar Counted Back from November 27

Quick answer:Black Friday 2026 falls on Friday 27 November, with Cyber Monday on 30 November. Meta's July 2026 benchmark report puts the global peak intensity window at "mid-November until early December" and says 54% of festive shoppers start in November. So the sale creative has to be live by the middle of November, and the test that decides what it looks like has to run in October. From today, 4 September, that is twelve weeks, and Sora 2 leaves the catalogue on 24 September, inside the hypothesis-batch weeks. This post counts the weeks back from 27 November and attaches each one to a constraint you can check: a platform's own published data, a documented product change, or a clip cap and credit cost from our own catalogue. The week numbers are our recommendation. The constraints are not.

What Meta's own Q4 data says about timing

Many Q4 calendars start from an agency's rule of thumb. This one starts from the two report pages Meta published in July 2026 in its Festive Season Insights Centre. One is an Ipsos survey of 14,473 festive shoppers across 18 countries, fielded 5 to 19 November 2025. The other is a benchmark analysis of "every purchase-optimised campaign running from October to mid-January", normalised to a 1 October baseline and reported as daily medians at the customer level. Four findings bear directly on when creative has to exist:

• Start dates."Globally, some holiday shoppers begin before October (14%), but the majority (54%) start in November. The peak intensity window globally is mid-November until early December, when 47% of all shoppers are actively purchasing."
• Black Friday participation is regional.81% of shoppers in EMEA, 73% in North America and 72% in Latin America take part in Black Friday sales, against 49% in APAC. Meta's stated response is to "Build a regional launch calendar, not a global one."
• Mid to late December is not a bargain."Western European markets saw CPMs drop 15-30% as advertisers pulled back, and conversion rates fell in step." Cheaper impressions after the sale did not convert in those markets.
• Vertical video with sound held up best in the auction."On Black Friday, when auction competition peaks, campaigns using 9:16 vertical video with sound saw lower CPM inflation, 48% versus 61% for other formats." Meta's advice on that page: "Prioritise building a library of 9:16 sound-on assets before peak season."

That last line is the whole brief. The rest of this post is about what "library" and "before" mean once you are generating those assets with AI video models rather than booking a shoot.

The constraint most calendars miss: fatigue is measured per creative

The reason a Q4 batch has to be a library rather than a single hero ad comes from a paper Meta's analytics team published in May 2023. It defines creative fatigue as "the effect that occurs due to users seeing the same creative (the visual elements from an ad) repeatedly". It counts exposures at the creative level, not the ad level, because "multiple ads may have the same creative elements, but a user would experience both ads as the same."

The measured shape matters for planning. Holding everything else constant, "the reduction in likelihood of a click follows the shape of (N+1)^-.43 where N is the number of times the user has seen the same creative prior to that impression (ie. indexing starting with zero)." Then: "This means that at 4 repeated exposures the associated likelihood of a conversion drops by about 45%." The paper also found no evidence of a wear-in period for direct response objectives: "Clicks and conversions appear to become monotonically more expensive with repeated creative exposures." Meta then ran a split test across roughly 26 thousand cases, and the summary at the top of the paper reads: "Guidance to reduce fatigue can improve conversion rate by on average 8% for high fatigue cases (as shown in the test below)."

Two planning consequences follow, and neither is a number of days:

• Ten renders of the same actor, same script and same background are one creative to the fatigue model. A library that protects you through a two-week peak has to be built from variants that look different: different hooks, different actors, different settings, different formats.
• The remedy the paper tested is "adding a new creative of equal or greater quality being delivered to the same user." That means holding a reserve of finished, distinct variants in November, so you can add them mid-sale instead of generating under pressure on 28 November.

The frequency metric in Ads Manager will not tell you when to do this; the paper explains that frequency is the wrong signal because it averages across a period and sits at the ad or ad set level, not the creative. Plan the reserve up front. The full reading of that paper, and what Ads Manager flags instead of frequency, is in creative fatigue: what Meta's own study says.

What a variant batch costs, and how long a clip can be

A shoot fixes the number of variants on the day; a generation budget lets you keep adding them, but each render has a clip ceiling and a credit cost. The table is read from our own credit-costs.ts and replicate-models.ts on 4 September 2026. Seedance 2.0 runs on BytePlus in production; everything else runs on Replicate.

• The three per-second models are priced on the spoken length of the script, not on a duration you pick. A 25-second Black Friday script on Pruna Avatar is 460 credits (25 x 18.4); the same script on OmniHuman is roughly 820. Cut the script, cut the cost.
• A 30-second sale ad on Veo 3.1 is four 8-second renders stitched together. On Kling 3.0 or Happy Horse 1.1 it is two 15-second renders. Budget the clip count, not the ad count.
• Sora 2 and Sora 2 Pro are absent on purpose. Both leave our catalogue on 24 September 2026, when the Replicate listing we run them through is shut down. That date sits inside the hypothesis-batch weeks of this calendar, so a batch built on them cannot be re-rendered in November.

The number of hypotheses that batch needs before a winner means anything is a sample-size question, worked through in how many UGC ads to test before scaling.

The calendar, counted back from 27 November

Dates below are Mondays, and "weeks out" counts whole weeks before the week of 23 November, which contains Black Friday. The assignments are ours; the constraint in each row is the reason the work sits in that week and not later.

Two Q4 traps that come from product changes, not from creative

The disclosure toggle does not survive duplication

The fastest way to build a Black Friday campaign is to duplicate October's. On TikTok that quietly removes a disclosure. The help article is explicit: "Any new campaign creation, including campaign duplication, will reset the AI-generated content disclaimer toggle, even if you reuse an ad that you previously used this toggle for." It adds that "After you submit, you cannot change the toggle or unselect it unless you change the video or image." Easy to lose on the copy, hard to fix after submission. Which TikTok ads need the toggle at all is covered in TikTok's AI disclosure rules for advertisers; on Meta the mechanism is different, as Meta's AI creative policy explains, because Meta labels ads made or materially edited with its own generative AI tools, and its wider AI-label policy is a separate document.

The flexible format is gone, and flexible media is not the same thing

If your 2025 Q4 playbook said "drop all ten variants into one flexible ad and let delivery choose", that option has closed. Meta's help page now reads: "Starting in March 2026, the flexible format will no longer be available in Ad setup." The format allowed "up to 10 images and videos in a single ad campaign". It pitched itself partly as a fatigue tool, since grouping media with different text combinations would "help prevent ad fatigue."

Do not confuse it with flexible media, which still exists (Meta notes it "may not be available to you yet") and does something narrower. Meta's own note draws the line: "Flexible media determines if a media asset could be used across additional placements to help improve performance, while the flexible ad format looks at multiple media assets uploaded for an ad and determines which ad format is most likely to help improve performance." Meta's page keeps the old toggle for one case and moves the rest: "If you're using the awareness objective or any objective with calls as your conversion location, go to the Crop screen and turn flexible media on or off. For all other objectives and conversion locations, flexible media is gradually moving to Advantage+ creative." For a sales campaign without calls as the conversion location, that means the Flex media toggle on the Enhancements screen. The same page warns: "We aren't currently able to detect if objects or text have been cropped out of videos like we're able to for images, but you can use ad previews to better understand how your videos may appear before you publish your ad." For a 9:16 sale video with a price in the lower third, that means checking the preview in every placement before 27 November. Each distinct AI variant is its own ad now, which is more setup than one flexible ad was, and another reason the campaign build sits in the week of 9 November rather than the week of the sale. How to lay the sale variants across A/B tests, multi-ad ad sets and one-creative ad sets is in testing AI video variants on Meta after the flexible format.

Where AI generation actually moves the calendar

Not at the front. The test still needs October audience data, and Meta's shopper data says the peak still starts mid-November. What changes is the back half: a generation budget lets you re-render the October winners as region-specific sale versions in the week of 26 October, and keep a reserve of finished, distinct variants for the fatigue add in the sale week. Those are the two remedies Meta's own documents point at: a regional calendar, and new creative of equal or greater quality added to the same audience. Sizing the reserve is a credit question: decide how many distinct sale variants you want across your regions, multiply by the clips each one needs at the model's ceiling, and read the credits off the table above. If the total does not fit the month's pool, the levers are model choice and script length.

Sources

• Meta for Business, Meta Festive Season Insights Centre (2026 hub page)
• Meta for Business, "Five insights that show how festive advertising on Meta technologies really works" (published July 2026; purchase-optimised campaigns, 1 October 2025 to 15 January 2026)
• Meta for Business, "What 14,473 festive shoppers told us about global shopping habits" (published July 2026; Ipsos survey fielded 5 to 19 November 2025)
• Analytics at Meta, "Creative Fatigue: How advertisers can improve performance by managing repeated exposures" (10 May 2023)
• Meta Business Help Centre, "About the flexible ad format"
• Meta Business Help Centre, "About flexible media in Meta Ads Manager"
• Meta Business Help Centre, "About Advantage+ creative"
• TikTok for Business Help Centre, "How to add disclaimers to ads in TikTok Ads Manager" (last updated September 2025)

All eight were read directly on 4 September 2026. Meta's help pages and TikTok's help articles change without notice; the dates above are how you check whether this calendar has gone stale.

The batch itself is the hard part, and the model choice decides both its cost and its clip length. Kling 3.0 vs Veo 3.1 for UGC ads covers the two ceilings that matter most for a 30-second sale ad, and what makes AI UGC look fake is worth a read before you commit a reserve of variants you will not see perform until the sale week. The reserve of distinct sale variants this calendar asks for is a credit budget, not a shoot day; UGC Vids AI is where we render it, from $49 a month with a free 3-day trial for renders under 250 credits.