The original use for this Tool was to set the minimum target store stock (called a BTL or 'Build To Level') by balancing the twin risks of lost sales vs. markdowns, through too low or too high a BTL, respectively.

The BTL is a science on which merchandisers and others can then build their art. For example, rounding up to full pack picks, or full shelves, or 'store looks full' and so on. But they can now do so knowing the downsides, usually leftovers, markdowns and margin erosion as customers get more accustomed to fire-sale prices.

For vast swathes of retail (mid priced slow movers) most store BTLs should be 1 or 2. Anything more is 'art' - we're now able to see just how expensive art can be!

A by-product of our knowing (say) that a BTL of 2 meets our service level objective at the defined lead time is that we'll know the service level for BTLs lower or higher than the chosen one. That means we can calculate the downsides of having more or less than BTL yet - in the case of store stock inaccuracy - not knowing it. This page shows one such example.

The Benchmark product (magnified) aims to meet or beat 95% Service Level (SL) on a 1 day Lead Time. The size of a blob shows the amount sold. The first BTL to meet or beat 95% service level (green) needed 2; that achieved 97.6% SL.

Items understocked by one are shown red. These only achieved 85% SL, and thereby lost about 13% of potential sales. How many sales they actually lost varies from business to business; suppose it's 4%. What retailer can afford that size of loss simply through stock inaccuracy?

Items overstocked by one are shown amber. They achieved 99.6% SL, so increasing sales by 2% (and stock by 50%!) The extra item only sells once every 14 months. What retailer can afford that amount of dead stock (possible markdowns etc etc) simply through stock inaccuracy?

Inaccuracy is a prime cause of over- and under-stock, but not the only one. RFID is a tool which, on a solid base, can fix the last and most intractable piece of the BTL jigsaw.

If the base is unsound, fix that first. RFID - like any other techno magic - won't fix poor processes or control.