What is this 'Saturday Effect'?

Under various guises, the belief that Saturday sales require extra stock.

In a well tuned supply chain, almost all do not.
Badly tuned chains send much wrong stuff and don't send some right stuff. This is most visible on Saturdays, so reinforcing the myth.

Summary - The Colloquial Version

Most non-food SKUs (stock lines, incl colour, size and subsize) in a single shop don't sell every day. Just suppose an SKU sells once every 7 days - this is faster than average for great swathes of goods, from clothing thru consumer electronics

If Saturday were a normal day it would sell on a Saturday once every 7 weeks. Other weeks it would still sell, but on a different day.

If total sales each Saturday were twice as high as the average day, and we had one in stock, that one would sell on the Saturday every 3rd or 4th week. (Twice in 7 weeks, i.e. 2x more often than the 'level days' example we started with.)

With predictable sales and immaculate replenishment, a stock of 1 is enough to fill sales. In this simplified example, stocking an additional item (the crie-de-couer from those who would solve the Saturday Effect by adding stock) makes no sense.

Sales aren't predictable; Replen isn't immaculate.

Both right and wrong.
Sales First: Something selling once a week doesn't sell one every week. Some weeks it will sell none. Other weeks (perhaps 1 in every 6) it will sell 2 (if it has or gets 2). And there's a chance (albeit ever decreasing) the demand will be 3, 4, 5 and so on. We can only sell what's in stock, so sales are capped by a combination of what we stock and how fast we replenish any sale.
Sales, be they weekly or daily, aren't even. They vary - they are subject to chance, the whim of the buying public. They vary, but that variation follows a pattern which repeats time and again. Sales are 'predictably unpredictable', if you will. These underlying patterns are the demand curves. They alter with rate of sale, with sector (fashion is 'spikier'), the weather and competitive action. But when they do alter, it's not by much … much less than myth and legend which surrounds the subject.

Replenishment: Replen will always go a bit wrong some of the time. The DC doesn't have enough stock, returns won't have been processed in time … if you have less than 10 of your own, try harder!
My plea is to give it a chance, then take the rough (by now, rather less rough) with the smooth. Rather than curse what we can't control, let's look at what we can:-

1. Top of any list is store stock accuracy. If in-store stock accuracy, SKU by SKU, is wrong then so will replenishment be. On the one hand
   'You said you didn't need another, so I didn't send one' is the stupid computer's response when your book stock is higher than physical.
   On the other hand, 'You said you had none and needed 1, here it is' illustrates why the DC keeps sending stuff you don't need and may have to markdown.
   Both persist until you fix the underlying cause. If you do sell that extra item and still haven't corrected the book stock, the DC will send another. If you lose sales thru insufficient stock and the computer still thinks you have enough you'll lose yet more sales.
2. The DC may not have stock, a common problem with multiple DCs, where we've found 40% stock mis-disribution. (A small amount is inevitable, but 7 in every 8 were avoidable)
3. Another store may be overstocked with exactly the SKU you desperately need. Sometimes just luck, but if low stock accuracy is endemic their (avoidable) overstock - the computer thinks they need stock which they already have - will always happen. It's worse; you'll be overstocked with something they could sell, and you won't
4. We aimed to run out. Pardon? If we set a target service level of 95% ex-stock sales[1] that means 1 time in every 20 we aimed to run out. That we did is a victory, not a defeat.
   This is doubly true in fashion goods - approaching end-of-season we want to run out of stock.

Does it matter?

Although an aside, that's a really good question. In one study the same shop stock, regularly replenished, met the target service level for all sales between 7 and 27 a season. If the lowest rate of sale were to triple we'd still keep the same stock.

1. Anything much higher is a fast route to bankruptcy. That's because the additional stock sells so infrequently that most times it gets discounted. Having already clogged the shelves all season, taking space from items which could and would have sold.