#
Optimizing Inventory with XpertMart^{TM}
##
Using Minimums with No Sales History
One of the myths that keeps retailers from aggressively using Minimums
to control their inventory is that you need months of historical sales
data in order to properly set the parameters. Minimums are fine and well
for styles that have been around for a long time, the myth goes, but are
irrelevant when it comes to new styles being stocked for the first time,
for example at the beginning of a season. However, we believe that you
should always use Minimums, even as you just begin receiving merchandise.
Suppose you receive a new shipment of men’s casual western boots you’ve
never carried in stock before. To start off, you are obviously going to
have to use an educated guess to set your initial stock level. The best
thing to do is to look for a style that is similar to the new style, in
this case another casual western boot, and use its historic sales data
to set your initial Minimums. If there is no similar style to rely on,
try picking another style that is in the same Type in your Departments
catalogue and move up to Sub-Class and Class until you find a good fit.
For example, if you cannot find a men’s casual western boot, perhaps you
can find a men’s dress boot. So far so good. Now once the boot is on the
shelves, we begin collecting sales data. As soon as we get our first sale
for a particular item, we can begin to calibrate our Minimums. Suppose
we are looking at a size 11 brown boot. For the first three days there
is no sale, and then on the fourth day there is a sale. This means that
so far we can expect one sale every four days.
Now suppose your goal is to keep 45 days of sales in stock. If we were
to calculate this item’s Minimum, we would need to use a multiplication
factor of 11.25 (since there are 11.25 4-day periods in 45 days). However,
we do not yet have enough reliable data on which to calculate a Minimum
(your multiplication factor should never be higher than 3.5 or 4). Therefore,
the best option at this point is to set our Minimum so that we are only
ordering merchandise for the next two weeks and then recalculate our Minimum
when we have more data. If we want to keep 14 days of sales in stock, we
need a multiplication factor of 3.5, barely within our acceptable range
for a multiplication parameter. So far, we can conclude that we need to
have 4 pairs of the size 11 brown boot in stock to cover the sales we are
anticipating over the next two weeks.
Now obviously this Minimum is not terribly accurate since we are only
relying on four days of data. For all we know, that sale was a fluke and
there won’t be another sale for days to come. Or maybe the first few days
were flukes, and the size 11 brown boot will really start to sell. As we
mentioned earlier, the larger the Multiplication Factor, the less accurate
the Minimum. However, it is better than our earlier educated guess since
it is using actual sales data, and it is far better than no guess at all.
Obviously, as time goes by and we gather more data, we can continue to
calibrate our Minimums and increase their accuracy.
To continue with our example, let’s suppose that after two weeks of
sales we have sold 5 pairs of the size 11 brown boot. Now we can recalculate
our Minimums, and this time we have enough data to aim for 45 days of sales
in stock. Our multiplication factor of 3.2 (45 divided by 14) is within
the acceptable range. We would use the following parameters:
**Sales Period:
**April 1 – April 14
**Multiplication Factor:**
3.2
**Sales:**
5
**Minimum:**
16
Our new Minimum is 16. As we’ve gathered more data, it turns out that
the first four days were somewhat atypical and that our first Minimum was
not very accurate. Now we have a new Minimum and it is based on 14 days
of actual sales, not 4—so our educated guess is getting increasingly educated.
The closer our multiplication factor gets to 1, the more accurate our Minimum
will be. The ideal, of course, is to have 45 days of stock as our Minimum
based on 45 days of actual sales.
Now suppose that after three weeks of having our new boot on the shelf,
we have sold 10 pairs of the size 11 brown boot. Let’s calculate our Minimum
once more, with the following parameters:
**Sales Period:**
April 1 – April 21
**Multiplication Factor:**
2.14
**Sales:**
10
**Minimum:**
21.4
Our new Minimum is 21, or 22 if we choose to round up. Notice that the
difference between this Minimum and our last Minimum is not as great as
it was at first. This is because our accuracy is increasing (the multiplication
factor is down to 2.14). We are starting to settle on a Minimum that is
most likely in the 16 – 22 pairs range—not bad after only 3 weeks of sales!
Let’s finish our example by supposing that during the first 45 days, the
size 11 brown boot has sold 20 pairs. Let’s look at our parameters one
last time:
**Sales Period:**
April 1 – May 15
**Multiplication Factor:**
1
**Sales:**
20
**Minimum:**
20
We now know that our Minimum should be 20 pairs. We can have great confidence
in this number as our multiplication factor is 1, i.e. we are relying on
actual sales for the full 45 days we want to keep in stock. We can now
use this Minimum to determine future purchases and stocking levels. More
importantly, along the way we were able to use Minimums for the same purposes
that became increasingly more accurate and useful (we began with 11 as
our Minimum but after less than two weeks were already at 16, not too far
from the real Minimum). So even when we are receiving totally new merchandise,
we can begin to optimize our inventory within a short period of time. There
is some question about how often we should calibrate our Minimums. In the
beginning, our Minimums become increasingly more accurate, as even the
difference between 4 days of sales data and 10 days can be significant.
At this point, we can keep a higher percentage of new styles at the warehouse:
if Minimums go up, then we ship more to the stores; if the Minimum is lowered,
then since it is still early in the season, the excess will be sold down.
A good rule of thumb is to recalculate your Minimums once a week until
you reach a Multiplication Factor of 1. After that, it’s still a good idea
to recalculate your Minimums every week. As we’ve mentioned, you will need
to calibrate your Minimums with greater frequency if sales seem to be rising
or falling more than usual.
Copyright © 2002 XpertMart |