Every merchant has a feeling about which day of the week is good, and almost nobody has measured it against their own order book. This study sorts an example store's 137.947 orders by the weekday they were placed on, in the store's own timezone, and totals net revenue per trading day rather than in aggregate, so a weekday that occurs more often in the window does not simply look bigger. Saturday earns 31% more per trading day than Monday, the weekend pair sits clearly above the working week, and the four days from Tuesday to Friday are close enough together to be treated as one.
- Saturday leads at €38.735 of net revenue per trading day, with Sunday just behind at €37.397.
- Monday is the weakest at €29.547, which makes the spread 1,31 times between best and worst.
- Tuesday through Friday sit within about seven percent of each other, so the working week is effectively flat.
- Costs do not follow the pattern: every day carries the same prorated fixed block, which is what makes quiet days look worse than they trade.
The shape, day by day
The weekend is the story and the rest of the week is remarkably level. Monday and Tuesday are the two weakest days, Wednesday through Friday improve slightly, and Saturday and Sunday sit well above all of them.
That is a consumer catalogue's pattern rather than a universal one. A store selling to businesses would very likely show the mirror image, which is the first reason to run this on your own orders rather than adopting these figures.
| Weekday | Net revenue per trading day |
|---|---|
| Monday | €29.547 |
| Tuesday | €29.714 |
| Wednesday | €31.421 |
| Thursday | €31.666 |
| Friday | €31.646 |
| Saturday | €38.735 |
| Sunday | €37.397 |
Costs do not have a weekday
Revenue has a weekly rhythm and the fixed block does not. Rent, salaries and software are consumed every day at the same rate, which is what proration puts on a statement, so a Monday carries the same fixed cost as a Saturday while earning about three quarters as much.
That single fact explains most of the anxiety a first daily statement produces. A quiet Monday shows a loss by arithmetic rather than by failure, which is the first thing to rule out in the anatomy of a losing day.
What it changes about campaigns
Comparing a campaign day against the day before it is meaningless when the underlying pattern moves by a third across the week. The honest comparison is the same weekday over recent weeks, and a campaign launched on a Saturday will beat almost anything it is compared with for reasons that have nothing to do with the campaign.
The same applies to a bad day. Before diagnosing anything from one figure, check what that weekday normally does, which takes seconds on a daily statement and is impossible on a monthly one.
The operational reading
A weekend-heavy pattern has consequences beyond reporting: orders placed on Saturday and Sunday queue for Monday dispatch, so the warehouse's busiest day is the store's quietest one, and support volume follows delivery rather than sales. Staffing built around the sales chart will therefore be wrong by two days in both directions, understaffing the Monday pick and overstaffing the Saturday desk.
None of that is visible in a revenue chart, and all of it is visible the moment revenue is read per weekday rather than per month. That is a small argument for the daily habit in general.
How to run it on your own orders
Export a year of orders, drop test and cancelled rows, strip the VAT, and total net revenue by weekday. Then divide each weekday total by the number of times that weekday occurred in the window, which is the step most people skip and the one that stops a window containing five Saturdays from making Saturday look better than it is.
Two refinements are worth the extra minute. Exclude promotional periods, since a Black Friday inside the window will attach itself to whichever weekday it fell on. And run the same exercise on orders rather than revenue, because a weekday can be strong in volume and weak in basket size, which points at a different audience rather than a busier one.
About this data
An example store's own generated order book, bucketed into the store's timezone so an order at half past eleven at night belongs to the day the merchant experienced. Revenue here is net of the VAT each order carried. The pattern belongs to this catalogue and this audience, so read it as a demonstration of the method rather than as a benchmark, and see how every figure is computed for the rules behind it.