Supplier Familying: Behind the Hype
Grouping together multiple instances of the same supplier —
"familying" — provides better visibility into the total
spend with that supplier, and, for very large companies,
|
When you've grouped spelling variants together, you've solved most of the supplier familying problem.
|
|
may make it possible to take advantage of higher discount
levels from the supplier. Suppliers are entered multiple
times in the payables system for a number of different reasons,
which may include different billing addresses, different people paying bills,
different divisions
utilizing the same supplier, acquisitions where the acquired
company continues to run its own accounting system, and
clerical error. A clear view of the total spend with a given supplier
is therefore difficult to achieve without a "familying" exercise —
i.e., a grouping exercise that enables the total activity with that
supplier to become visible, as in the example below:
The familying problem is a classic 95-5 type problem, and in this case 95%
of the problem is simple to solve. In practice, it turns out
that almost all of these redundant suppliers are misspellings
or re-keyings of the same supplier name — "I.B.M.", "Int'l Bus Mach",
"IBM", "International Business Machines", "Internat. Bus. Mach.",
and so on. If you can group these variants together, you're pretty
much done with supplier familying. BIQ provides tools that make this process easy.
Some spend analysis vendors will tell you a more complicated story.
According to them, it is critical to know the "who owns whom" relationship,
so that I.B.M. in Saudi Arabia, which operates under a different name,
can be familied together with the other IBM entries. The vendor typically
sports "automated tools" to accomplish the match between their database and your supplier list.
There are three problems with this reasoning: (1) often the "who owns whom"
relationship is of no value — Hilton Hotels, for example, are independently
owned, and therefore grouping them by owner is a mistake — in fact, they should
be grouped together, even though their owners vary;
(2) even if
A owns B, often neither A nor B is interested in the
fact that you are doing business with the other (for example, UTC owns both
Carrier and Otis Elevator, but you won't get any discounts from UTC); and
(3) automated matching tools aren't very good at
matching, so the "automated tools" usually contain a rather large
(and unmentioned) "manual" component. We suspect it's unlikely
that 10 strangers in Bangalore or Boston can do a better
job at familying your vendors than your own commodity managers.
We've listened carefully to our
customers and we've applied engineering principles to what they told us.
So, BIQ provides a Dimension Editor that enables efficient familying
of supplier and other dimensions. Even an inexperienced BIQ user can
family thousands of entries per hour. This makes familying a once-only
chore that seldom takes more than a day, and solves pretty much all of the problem.
One BIQ customer completed a familying exercise on 80,000 vendors in less
than two days. Maintaining vendor families as new vendors come in every month
is a 10 minute exercise.
All of the above notwithstanding, if your company is large enough
that discount level differences can add up to significant savings —
and you can't trust your commodity managers to be aware of major vendor
relationships — then you can always employ Austin Tetra or
D&B to take a pass through your suppliers to catch stragglers.
We haven't encountered a situation where this effort has proved
to be worthwhile, but we can't discount the possibility that there
might be one.
