Proposed Project for the MSU Industrial Math Students

Steelcase Pricing Value Streams

Title: Uncovering hidden business patterns*

Steelcase sells office furniture around the world, but the majority of
its production and sales are in North America. The sales orders come
to us through many ways ranging from very simple (an order for 10 office
chairs) to very complex (a new corporate headquarters). The attached
slide might help you see how we can intuitively segment the business
into two distinct value-streams. I believe that this split is the most
effective initial split to start understanding the key differences in
our business. This split also allows us to uncover different problems
and their root causes. It also allows us to then identify solutions to
those problems with greater precision. We have transactions that can
be classified as simple with several sub-branches, and other
transactions that are more complex, also with several sub-branches.

I think that the greatest opportunity for a project between Steelcase
and MSU's Industrial Mathematics is with our more simple transactions.
They are numerous in volume and typically lower in dollar value. Since
they are of lower dollar value, they typically don't get as much human
oversight. I have a hypothesis that advanced analytics can provide
some overall insight into these transactions that will lead us to
establish policies and rules that could either help us increase sales or
focus selling efforts into more productive areas.

The students would start with a purposely limited scope of data that
would tend to provide more homogenous situations. The attachment might
provide some more clarity. Including the more complex transactions
seems to add so much noise into the data set that it is difficult to
draw any meaningful results. The initial data is that which is
normally on any commercial invoice....customer number, product lines,
prices, PO numbers, etc. We also have a Customer Information Database
where we collect a lot of data about every customer we add to our
database. This includes information like the company's industry code
(SIC and NAIC's), the number of white collar workers employed by the
company, the primary address, etc.

We believe that there are "patterns" that exist within our own data
(subset) that, if uncovered, could lead us to significantly change
business and sales processes. For example as a possible pattern,
Steelcase is very successful with smaller regional hospitals in
non-metropolitan areas, but we are not as successful with large
hospitals in metropolitan markets. If the students were to receive a
few years worth of sales transactions (order detail) along with the
Customer Information Database, we're hoping they could apply advanced
analytics and data-mining techniques to uncover these otherwise
hard-to-detect patterns.

_________________________________________________________________
* This problem description was prepared by John Shull,
Steelcase Pricing & Contracts, Sales