What Black Friday taught you

The Monday after Black Friday is the best measurement day of the year, and most brands spend it reading the email platform's dashboard. A week later the attribution has settled into a story, the story becomes the plan for next year, and the plan repeats whatever the story said worked. The weekend's orders are in the store. Read those first, in four splits, before anyone opens a report.

Two brands I planned Black Friday with this year show why.

What a discount bought

A golf and corporate apparel brand, from the full-price case, had run the same Black Friday two years running: 30% off sitewide, early access for its most loyal customers. In 2025, 95% of orders in the window used the discount (count), and the discount came to 29.1% of the list value of everything sold. Orders in the window fell 29.5% anyway (count, 2025 against 2024), and November and December revenue fell 24% year over year. The discount was on nearly every order and did not hold the volume.

Read backwards from the email report, that weekend was a success: every send was credited with orders, because every order used the code the sends carried. Read forwards from the store, it was a smaller weekend than the year before that paid 29% of list price for the privilege. This year the brand kept every product at full price and added a gift card at two basket sizes, sized from a full-price month. The readout below is how we will know whether that was right.

What the list was given

A hand-dyed yarn brand, from the early Black Friday assignment, gave its email list the biggest code of the year in a short window in early November and kept the main weekend to chosen items at different depths, with no sitewide code. Its regulars bring in 79% of revenue (share, six months to July), so the question for that brand is not how many orders the weekend produced. It is whether the regulars bought in the early window and stayed quiet on the main weekend, which is what the plan intended, and whether the main weekend found anyone new.

Two brands, two plans, one readout. The splits are the same.

The four splits

1. Volume, forwards. Orders in the window this year and the same window last year, from the store, counted the same way. Then revenue. If orders fell and revenue rose, the basket did the work; if the reverse, the discount did.
2. New against returning. For every order in the window, was it the customer's first ever? Count both groups and compare each with last year. A Black Friday that was all regulars at a discount moved purchases forward and paid for the privilege. One with more first-time buyers than usual did acquisition.
3. Discounted against full price. The share of orders that used a code, and the discount as a share of list value. If a code sat on 95% of orders, the sale was not a sale, it was the price.
4. Claimed against real. Store revenue in the window next to what email claims and what the ad platforms claim. The claims will add up to more than the store took. The gap is the share of the weekend that was already in motion, which is what "attributed" means; the attribution assignment has the holdout that settles it.

The paragraph that changes next year

The readout ends in writing. One paragraph, dated, with the numbers next to the words: what the weekend did in each split, what the plan expected, and the one thing that changes next year. "Orders down 29.5% with a code on 95% of them" is a lesson. "Black Friday underperformed expectations" is an excuse with a thesaurus.

Then call it even. The point of the record is to stop carrying the weekend into January, where the real work is the customers who bought for the first time and have sixty days to buy again.

Do this Monday

1. Export the window's orders, this year and last, same days, with customer, date, list value, paid value and code used.
2. Count orders and sum revenue for both years. Write both ratios down.
3. Flag each order as first-ever or returning from the customer's first-order date. Count both.
4. Count orders with a code and the discount as a share of list value.
5. Put store revenue next to each channel's claim for the same days, and note how far the claims exceed the store.
6. Write the paragraph, date it, and put it where next year's plan will be written.

In one line: read Black Friday forwards from the store in four splits, volume, new against returning, discounted against full price, claimed against real, and write down the one thing that changes next year.