Data mining develops the MARKET
The most
powerful force in today’s supermarket wars is
information. Loyalty cards and EPOS systems provide the
ammunition. Data warehouses and data marts serve as the
arsenal. And perhaps the most powerful weapon is data
mining - the process of discovering unexpected patterns
and trends in data.
Yet this
process is often misunderstood and undervalued. Data
mining claims a novel kind of data exploitation - it is
not simply the hypothesis confirmation of statistics; nor
is it simply the data visualisation of graphs and plots.
Data mining is becoming a force to be reckoned with, say
its supporters, because of the way it can generate new
ideas.
In short, a
good data mining tool should be able to gain the kind of
insight that a skilled human might gain from looking at
the detail of the data. That is, if humans were capable
of truly comprehending data with hundreds of fields and
thousands (perhaps millions) of records.
So what can
data mining do for retail? That question is at the heart
of the MARKET project. Part funded by the European
Commission, MARKET uses the Clementine Data Mining System
with records from Somerfield Stores to demonstrate how
the numerous techniques that embody data mining can
provide real business benefit to the supermarkets of
Europe.
MARKET is
being run by a trio of specialists. Somerfield Stores are
a top five UK supermarket chain, serving more than seven
million customers each week. The Parallel Applications
Centre (PAC) is a wholly independent organisation that
specialises in bringing early adopters of new technology
together with pioneering suppliers. Integral Solutions
Ltd (ISL) were the first company to produce a
comprehensive data mining tool for non-technical users,
and their Clementine Data Mining System continues to be a
market leader.
The aims of
the project are twofold. Firstly, Somerfield will benefit
directly. Studies will be undertaken with them on their
data to produce business knowledge; but they will also
develop a wide scale method for deploying this knowledge
throughout the business, customised to their needs.
The second
result of MARKET will be the development of an on-screen
demonstration of how data mining can be used in retail
The benefits achievable through the MARKET approach will
be demonstrated to a Europe-wide audience through an EC
funded promotional programme. The idea is to take
advantage of the user-friendliness of Clementine to
demonstrate business advantage, and push forward the
burgeoning data mining market.
"This
kind of project is about ensuring that European business
can fully take advantage of new technology. Besides the
natural tendency to fear change, there is also a wealth
of technological development happening at the moment.
This is complex and potentially confusing, but my opinion
is that data mining will be crucial to successful retail
in the next century," says Paul Allen, of PAC.
These are
sentiments shared by other members of the partnership.
"Somerfield Stores have been enjoying the benefits
of a three year investment in turning information into
profit," says Jerry Warren of Somerfield. "We
have uncovered a great many opportunities for data mining
within our organisation - for instance, one of the
projects will be to optimise customer choice in
Somerfield's smaller stores."
This is the
kind of attention to detail that characterises data
mining. "Clementine allows users with little or no
technical expertise to get right into their data,"
says Tom Khabaza of ISL. "I’ve been using these
techniques for many years, and I can tell you that even
as a specialist in data analysis your work rate is
boosted tenfold by Clementine."
The project
is still in a relatively early stage, but current studies
include looking at bread buying patterns. When should
bread deliveries be scheduled or baked in store? What are
the effects of promoting a product at a keen price on
other products? These micro studies are fundamental to
the bigger picture of how people shop. And it is by truly
understanding how people shop, claim the MARKET team,
that their needs can best be catered for.
The bread
exercise is just one of a number of issues that the
project will address. They are also looking at promotions
and basket analysis. But its fundamental importance, says
Warren, is in defining the parameters for data mining in
the future. "Data mining is powerful, and Clementine
makes it accessible and practical to do. It is not a
panacea, but it is one of the most efficient and
effective ways to add value to data - and the kind of
investments many companies are making in data
accumulation and storage right now may be hard to justify
without it."
|
