Keep the same attribution window you use the rest of the year, and change how you read it instead. Switching windows for BFCM makes your Black Friday numbers impossible to compare with anything else, and comparison is the whole point of measuring the weekend. What does change is that many BFCM purchases were set up by ads people saw weeks earlier, so a short window hands more credit to retargeting, email, and branded search than they actually earned.
Why does BFCM distort attribution windows?
Because much of the demand is pulled forward, not created. Plenty of people who buy on Black Friday decided earlier and waited for the discount. A 7-day click window credits whatever they clicked that week, which is often a retargeting ad or a promo email, not the prospecting campaign from October that introduced them to the brand. Platform-reported ROAS looks spectacular during BFCM for the same reason, as I explained in how should you read ROAS during a sale or promotion.
Should you lengthen your attribution window for Black Friday?
Not as your main view. Add a longer window as a second lens instead. Compare 7-day and 28-day click credit for the prospecting campaigns you ran in October and November. If prospecting picks up a lot of credit in the longer window, it was doing the work that BFCM promotions got credit for. Keep the standard window as the number you report, so this year’s results line up with last year’s and with next month’s. The longer window is a diagnostic. It tells you how to plan next October’s prospecting budget, not how to grade this November.
What attribution window should you use the rest of the year?
For most e-commerce brands under $20M, a 7-day click window for daily decisions and a longer one, 28 days or more, for judging prospecting is a sensible default. Higher-priced products with long consideration periods need longer windows across the board. I covered the trade-offs in choosing attribution windows without overthinking it and how do you attribute sales with a long purchase cycle.
How do you judge whether pre-BFCM spend worked?
Look at the whole season, not the weekend. New-customer revenue and MER from October through December tell you more than any single week. Compare cohorts as well: customers acquired at a deep discount often repeat less than full-price customers, and gift buyers inflate new-customer counts in ways that never turn into repeat orders, which I wrote about in why gift shoppers distort your Q4 new-customer numbers. For a tool view, tools to measure whether BFCM actually worked covers the options.
How does ThoughtMetric handle attribution windows?
ThoughtMetric lets you switch between five attribution models and configurable lookback windows on the same data, so you can compare a short and a long window for BFCM campaigns without touching your standard reporting. (Disclosure: ThoughtMetric sponsors this site.) Post-purchase survey answers are folded into its model, which helps with the early deciders and gift buyers that click tracking misreads during BFCM. It starts at $99 a month for 50,000 pageviews, with every feature on every plan.
What should you set up before Black Friday?
Lock your attribution window and model now and leave them alone until January. If you also report numbers from Meta or Google, keep their attribution settings fixed too, since changing them mid-season breaks the comparison with last year. Write down the dates of every promotion and major email send so you can line them up against the data later. Turn on a post-purchase survey if you do not already run one, since “how did you hear about us” answers are most valuable exactly when click data is noisiest. Then decide what success looks like before the weekend starts, in new customers and contribution margin rather than ROAS. The full checklist is in what to fix in your attribution before Black Friday.
Quick answers
- Should you change your attribution window for BFCM? No. Keep it consistent and add a longer window as a second view.
- Why does BFCM ROAS look so high? Short windows credit the last touch before purchases that were decided earlier.
- What should you judge BFCM on? New-customer revenue and MER across October through December.
- Want to compare short and long windows on the same data? ThoughtMetric (our sponsor) supports configurable lookback windows.
- Unsure whether prospecting drove BFCM sales? Compare its credit at 7 and 28 days.
Leave a Reply