Skip to content

The Black Friday Post-Mortem You Cannot Run in January

By January the data behind Cyber Week is a model of itself. Meta's windows have closed, GA4's default retention has expired the events, and nobody remembers what was live on the Sunday. The post-mortem window is the week of 7 December.

Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free

By , Founder & CEOUpdated 4 min read

Run the numbers for your store: the free attribution window calculator.

The Black Friday post-mortem has its own deadline, and it is earlier than the January review it usually waits for. Meta's default 7-day click window closes on every Cyber Monday click by 7 December. GA4's default retention deletes Cyber Week event-level data by late January. The exports, the windows and the memory of what was live all decay on a schedule. Run the post-mortem in the week of 7 December, or run it on a model of the data instead of the data.

Octalysis would call the January review an Evanescent Opportunity in reverse: the reward was available in December and quietly stopped being available while everyone was on holiday.

What closes, and when

WhatClosesWhy it matters
Meta's default 7-day click window on Cyber Monday clicksMon 7 Dec 2026After this the platform's Cyber Week number stops moving and can be captured as final
GA4 event-level data for Cyber Week, default retentionAround late Jan 2027Explorations on last peak stop being possible; the aggregated reports remain
Your own record of what was live each dayFades from the first weekDiscount depth, email sends, stock-outs and creative swaps are the covariates a read needs
Platform definitionsAny date, undisclosed in chartsThe book documents Meta's March 2026 click-definition change; the next one will not appear in a year-on-year line either

The export list for the week of 7 December

  1. GA4: export the event-level data for 20 November to 7 December, by CSV or by the BigQuery link if it is on.
  2. Each ad platform: export Cyber Week conversions under the default window and under a 1-day click window, side by side. The gap is the "touched earlier" population.
  3. The store: orders and revenue by day, the fact about money.
  4. The email and SMS platform: claimed conversions and the attribution window setting that produced them.
  5. A one-page diary: what discount, what stock, what creative, what send, by day. It takes twenty minutes on 8 December and is unrecoverable by February.

The three numbers to compute while the data exists

Coverage. Attributed conversions divided by orders. The Price of Being Found reports the publisher's own census at 48.6% of sessions with no resolvable source over 17 October 2025 to 2 September 2026, one company's data, not a population estimate. Yours is computable in an hour. The number nobody checks covers it.

Claim ratio. Every platform's claimed conversions, summed, divided by orders. Peak week is when it is highest.

Incremental contribution. If a holdout started by 2 October, its read is due now, against the pre-period fixed in September.

Why January is too late even when the data survives

Even with retention extended and the exports taken, the post-mortem that happens in January answers a different question. In December the question is what to change about the plan that is about to repeat in a smaller form for January sales and in a larger one next November. In January it is a retrospective with no decision attached, and reviews without decisions are the ones that get skipped.

What to do this week

  • If you have to defend the number: book the post-mortem meeting for the week of 7 December now, with the export list above as its agenda. The meeting will not book itself in the last week of November.
  • If you own the budget: decide today that the December read, not the January deck, is what next year's Black Friday budget will be based on, and say so to whoever prepares the deck.

The Black Friday 2026 calendar has the dates before the peak. This one is the date after it.

Dates and default settings as of 9 September 2026. The coverage census and the March 2026 definition change are from The Price of Being Found (Edition 2.10), Chapters 8 and 9, with the book's caveats.

Frequently asked questions

  • When should I analyse Black Friday results?
    In the week of 7 December 2026. By then Meta's default 7-day click window has closed on Cyber Monday clicks so platform numbers are final, and GA4's default 2-month retention has not yet expired the event-level data. Export everything that week, because the data decays on a schedule.
  • What data should I export after Black Friday?
    GA4 event-level data for the peak (CSV or BigQuery), each ad platform's conversions under the default and a 1-day click window, orders and revenue by day from the store, the email and SMS platform's claimed conversions with its window setting, and a daily diary of discounts, stock and creative.
  • Why can't I just run the Black Friday review in January?
    GA4's default retention deletes Cyber Week event-level rows by late January, the record of what was live has faded, and platform definitions can change on undisclosed dates. A January review also has no decision attached, whereas a December read still shapes the plan that repeats in January and next November.

Go deeper: Incrementality testing, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.

Keep reading

Terms in this article

Browse the full glossary

Your platforms guess.
We run the math.

Upload a GA4 export and see what each channel caused, next to last-click, in 1–2 minutes. The read is yours to keep.

The free finding runs in your browser. Your file never leaves it. The full read is €99, refundable within 30 days, no questions asked. Prices exclude VAT.