After Black Friday: four questions to ask your sales data

The four days from Black Friday to Cyber Monday are forecast to bring more than $85 billion in US retail sales this year. Four questions show whether the peak paid off for you: check margin after discounts, where shoppers came from, how much comes back, and whether first-time buyers order again.

What the forecasts expect

One forecast of US online holiday sales expects $275.1 billion from 1 November to 31 December, 6.7% more than last year, with Black Friday at $12.9 billion and Cyber Monday at $15.1 billion. A second forecast, expects the four days from Black Friday to Cyber Monday to top $85 billion. It expects them to grow about 8.75%, close to twice its forecast for the season as a whole (about 4.5%).

The first covers online sales only, so the two numbers measure different things. Your own review starts the same way: be clear what your sales total includes and which dates it covers before you compare it with anything.

How much margin went into discounts?

The online forecast expects discounts of up to 30% off listed prices during Cyber Week, for example 30% in electronics and 29% in toys. Revenue alone doesn't show what each order earned after the discount and the cost of the product.

Put the price each customer paid next to the listed price and the product cost. Then group orders by product and discount level. Check which products sold mostly at the deepest discount, and whether the same customers had bought them at a smaller discount before. If a promotion brought more orders, compare total margin as well as margin per order before deciding the extra volume paid off.

Where did the shoppers come from?

The mix of traffic is changing. The same forecast expects visits to US retail sites through links in AI tools to rise 130% on last year.

When one source grows and another shrinks, the total can hide it. Split orders, conversion and margin by where the visit started, and compare them with an ordinary week. A source that brought many visitors but few profitable orders shows up here, and not on the headline chart.

How much of it comes back?

Returns are counted after the sale. Ahead of last year's holidays, a 2025 report from the US retail trade association found that retailers expected 17% of holiday sales to be returned, and estimated that 19.3% of online sales would be returned over the year.

Match each return to its original order, so you can see it by product, discount level and traffic source. Update the numbers as returns arrive, and give the peak and an ordinary week the same amount of time before you compare them. A product that sold well at a deep discount and comes back often may have cost more than it earned.

Will the new customers stay?

Earlier research, on data from 2018 to 2020, found that customers won through paid search and paid social in Black Friday week cost 13% more to acquire than at other times of the year. Their value over the next six months was 9% lower than the average for the rest of the year. Your own numbers may differ, and that's the reason to check them.

Treat customers whose first order came during the peak as their own group. Compare how many of them place a second order, and what they spend over the following months, with customers who joined in an ordinary month. Give both groups the same time since their first order.

What these questions have in common

Each one needs orders next to something else: prices and costs, promotions, traffic sources, returns or customer history. Two of them, returns and repeat orders, also need someone to come back to the data in the weeks after the event, when the team is busy with the next season. Agree now who will look again, and when.

Where DecisionBox fits

DecisionBox asks questions like these on its own. With read-only access to your data warehouse, it explores across your tables and checks each finding again with separate queries, as we showed in how DecisionBox checks its findings. You get findings with the queries behind them and recommendations with an expected impact, so your team can decide what to fix first. Run on a schedule, it keeps asking after the peak, which is when the returns and repeat orders arrive. We wrote more about this for retail teams.

If you want to see what it finds in your own peak-season data, talk to us about a free proof of concept.