Black Friday, measured
Did Black Friday sell more, or just sell earlier?
The spike is easy to see. The weeks shoppers waited for it, and the December it borrowed from, are not. Measure what the event actually added, on your own sales file, with the season held constant.
Invitation only while in beta. Tell us who you are and we get in touch.
Weekly sales, with and without Black Friday
October to January, sample retailerAn illustration on one of our sample markets. On your file the line without the event is estimated from your own years, and every number comes with its confidence interval.
Four ways a Black Friday flatters itself
The report says sales doubled in the event week. All four of these sit inside that number, and none of them is new demand.
The weeks before
Once shoppers know the discount is coming, they stop buying at full price. Early November sinks, and the event week collects what it would have sold anyway.
The weeks after
The present bought on 27 November is not bought again on 14 December. The lift in the event week is partly a hole in the weeks that follow.
The season underneath
Late November rises every year as Christmas comes closer. Compared with October, Black Friday gets the credit for the calendar.
Last year is not a baseline
The date moves, from 29 November 2024 to 28 November 2025 to 27 November 2026, the weekdays around it change, and the year grew or shrank for its own reasons.
From the booked lift to the real one
Start from what the event weeks sold over the line, then take back what the weeks around them lost. On the sample retailer, of the booked lift was new demand. The rest was sales moved inside the quarter.
That is before the discount, which was paid on every sale of the event weeks, new or not. Next November’s plan starts from the net figure.
Where the booked lift went
sample retailer, 2026The path in tea
Three analyses, on the same file, each answering one part of the question.
-
Seasonality & Event Impact
Declare Black Friday, the weeks before it and the weeks after it as events of their own. Each gets its effect with its 95% interval, measured with the calendar held constant. Windows that overlap are flagged. How it works
-
Promo Elasticity
With a column for the discount, the weeks after every promotion enter the model as terms of their own, and the lift comes out net of the dip that follows. It needs at least three promotions in the file, and says so. How it works
-
Contribution Decomposition
If media ran harder that week, a Black Friday column enters the model as a promotion next to media, price and the season, so the week is split between them instead of all going to the discount. How it works
What your file needs
Two Black Fridays at least
Daily or weekly rows over two years, three is better. One event is one observation.
date week
The number to explain
Sales, orders, units. One column, the same on every row.
sales orders
What else moved
Discount depth, media spend, price. Optional, and each one is a part of the answer.
discount search_spend
National holidays load in one click for Italy, Germany, the UK and the US. Black Friday is yours to date, and the Data Scout tells you before the run if the file is too thin for the question.
Black Friday incrementality, in a paragraph
The incremental effect of Black Friday is the difference between what you sold and what you would have sold in the same weeks without it. The second number is never observed, so it has to be estimated: from the years you have, with the trend, the season and the weekday read at the same time as the event. The weeks around the event belong to the measure, because a discount that everyone sees coming changes when people buy before it changes how much they buy.
What comes out is three numbers with their intervals: the lift in the event weeks, the loss in the weeks before, the loss in the weeks after. Their sum is what Black Friday added. When the interval around it crosses zero, the honest answer is that the file cannot tell, and that is an answer too.
How the event model reads a calendar, and where it misleads: Seasonality and event impact, explained.
How do I measure the real impact of Black Friday?
Against the sales you would have had in the same weeks without it, not against October. Late November rises anyway as Christmas approaches, so the comparison has to hold the season constant, and it has to count the weeks before, when shoppers wait for the discount, and the weeks after, when the gifts they bought early are not bought again. What is left is the incremental part.
Is one Black Friday enough to measure it?
It gives a number, with a wide band and no way to tell the event from everything else that happened in that one year. Two years of data, so two Black Fridays, is the minimum; three is where the estimate starts to settle. The interval around the effect says how far to trust it.
What is pull-forward?
Sales that would have happened anyway, only later: the Christmas present bought on 27 November instead of 14 December. They show up as lift in the event week and as a dip in the weeks after. Counting the first and not the second is the most common way a promotion is overstated.
Can I compare with last year’s Black Friday?
Only with care. The date moves every year, the weekdays around it change, and the year itself grew or shrank for reasons that have nothing to do with the event. A model that reads the trend, the season and the event together separates them; a year-on-year percentage does not.
What does it cost?
During the beta, nothing: every run is measured and no balance moves. After the beta, every plan includes a monthly token allowance: 150 on Free, 3,000 on Pro, 6,000 on Plus, 20,000 on Business. What is left of it expires when the month renews; tokens you buy never expire. The plan never limits which questions you can ask.
Plan next November on the net, not the spike.
Drop the file, measure the event, keep the number that holds. Free while in beta, by invitation.
Join the betaLearning
A report says what happened. An analysis says why.
Most of what lands on a marketing desk is reporting. Useful, and a different job. The blog tells one case a day where the difference decides the answer.
Sales were up 11% in April, TV ran six weeks. What happened, by channel, side by side.
4.2 of those points came from TV, 5.1 from Easter moving. What caused it, and what the next euro does.
When a budget optimiser says move 60% into one channel
Rolling cross-elasticity: checking whether a rivalry is stable
