A new source of competitive edge
The drive to understand customers
Data mining is the new hot topic among marketeers. It has the potential to help a company find new customers or increase loyalty among its existing buyers. But like all new technologies it suffers from hype — and there are as many horror stories as success stories.
This briefing helps users separate the fact from myth — and provides some best practice pointers to using data mining effectively as a marketing tool. But this briefing is not a general discussion of data warehousing in marketing. Instead, it focuses mainly on the creation of data warehouses as source systems for the data on which data mining techniques are used. It is aimed especially at:
- Marketing managers who are considering whether they should use data mining and how to organise their data.
- IT managers who wish to understand the key marketing and systems issues involved in data mining and warehousing.
Today, there is a new enthusiasm for understanding customers’ needs and how to serve them often driven by competitive market pressures. Some companies now collect client transaction data for analysis. Others are keen to win early benefits from analysis of existing data sources, primarily legacy systems holding product rather than customer information.
A few others — for example, John Hancock in the US — specifically collect new data from customers via surveys, then analyze it as quickly as possible for new insights. This is based on the theory that segmenting customers, understanding them and applying different offerings to each segment delivers happier and more profitable customers.
When you model interdependencies you promote understanding. For example, if we reduce price of product X by 5% for specific customers, what is the effect on volume, cost recovery, profit and the business risk profile?
It is just such questions that companies now highlight in publicity designed to show financial markets how competent their managers are. Companies such as Mellon Bank in the US and Safeway in the UK have showcased data mining projects in annual reports to demonstrate their innovative work is paying off.
Our research suggests industries which use data mining in marketing include those that either:
- Start to develop very large additional datasets because they are now using direct marketing techniques in integrated operations companies — for example, utilities, retail cards.
- Have been gathering datasets for some time and are now in a situation where better analysis is more important, because of recent competitive pressure, or have realised that data holds clues to competitive advantage — for example, frequent flyer programs or retail product data.
Back to hypotheses
So why is data mining so important? Busy marketing managers rarely have time to formulate many hypotheses. They are drowned in operational problems. Yet our research suggests gains achieved from data warehousing and mining projects increase time available to test the hypotheses that sometimes trigger marketing breakthroughs.
This is especially so considering that most warehouses are “data marts” or “mini marts” in which the company makes a positive choice about which variables to warehouse. So note that it is only possible to make sensible decisions about which variables to warehouse if you know the kind of hypotheses you wish to test.
Need for management commitment
Despite the huge potential benefits of data mining every success is matched by a problem. We find that some companies learn what data mining is but do not fully understand how to use it to support their chosen strategies. Unsuccessful data mining assignments are often caused by lack of investment in time, resources or intellect in defining the business issue or opportunity.
Senior management commitment is vital. Without it, a small data mining project can easily be dismissed as an interesting piece of technical analysis. In most cases, outside consulting expertise is needed to assist the company to use the mined information and to get value from it. This is because so often users of the information are not experts. For example, customer contact staff must be trained to manage the “new” information or sell from it.
We found companies who address the data issue by building a data warehouse from which they draw appropriate data. This becomes the platform for future information requirements. Companies with such warehouses find it easier to mine data to manage their customer value strategy and activity.
Results vary tremendously. But if the project is properly defined and has a team of company staff and external marketing and data experts working on it — with knowledge of what is to be done with the information and the systems in place to do it — the outcome can be productive. It may even set the company on the road to managing customers differently, for example, through relationship marketing.
For further information and to purchase contact Colin Coulson-Thomas