|
We illustrate the simplest case: The store is
replenished with singles through a BTL
(Build To Level), i.e. the on shelf + in transit
quantity is 'topped up' by the DC to the BTL.
Under steady state this settles down to SOGO,
(or 'sell one, get one') which stays constant
throughout. The DC always has enough stock.
The store sells the same total amount week
after week, and the lead time of resupply (DC
>> store) is constant.
In the real world,
everything is worse than this. Almost 60% of
all sales take place on Fri/Sat/Sun; BTL's
change like the weather; cheap items come in
box quantities, the warehouse shuts at the
weekend so lead times go out, and so on.
No matter, the purpose of this paper is to
illustrate some fundamental truths about
supply chain momentum. However alarming
those results might be, in the real world things
are worse. However large the opportunities
under steady state, in the real world they are
larger.
Configuring and running the tools
The tools are taken from the General Retail
Model Suite, with capabilities including the
trade off between service level and lead time,
between lead time and inventory, between
rate of sale and inventory or lead time or
service level, between box quantity and
service level, 7 day or 'skip day(s)' deliveries,
shelf space and sales and so on.
| |
Most of these are step relationships - at a
fixed rate of sale and lead time the service
level at a BTL of one might be 92% whereas
at a BTL of 2 it might be 99%. There's no
such thing as a BTL of 1.2 items!
In the past that meant it was difficult to
compare like with like … there was no
obvious way to hold the service level constant
and look at the inventory vs. service level
trade off. GRMS now contains proprietary
methods to bridge the gap, although these
runs are on the earlier (integer) model. For
that reason the illustrated rate of sale is higher
than usual … at higher rates of sale the step
increase in service level (between anyBTL
and anyBTL+1) is smaller than at low rates of
sale.
Method
We ran a GRMS tool at a range of rates of
sale, lead time and inventory. The tool mirrors
a real supply chain, so sales fluctuate about an
average. Because sales of each SKU vary
through pure chance and the BTL (on-shelf
plus on-way) remains a constant the resultant
on-shelf stock varies. With a one week lead
time the on-shelf stock today is a consequence
of yesterday's sales (luck, good or bad), the
day before's sales (more luck) and so on. |
