Using Data Mining on the Road to Successful BI, Part 1
Information Management Special Reports, September 2004
Given the ever-increasing amount of wrongly targeted direct mail and telesales calls we receive, one might think that many organizations still believe that customer acquisition is better than retention. Financial services companies, retailers and telecommunication (telcos) companies know little more about you than their mailing lists or partially completed Internet registrations reveal. They guess at demographics from an address and attempt to formulate psychographic profiles from individuals' magazine subscriptions details. This same information is available to almost any company. There is nothing unique about this data; little which can help one company differentiate their "unique" offer from another. They don't know what unique characteristics differentiate one customer from a potential consumer. They don't know a person's preferred method of paying their bills or seeking advice - is it credit card, debit card or cash for payment? Do they contact a shop, branch, call/contact center or the Internet for advice? Organizations that can answer these questions end up doing the bulk of the customer's business.
They will know what you buy; and when, where and how you do it. If they use a data mart/warehouse and have invested in the right tools and people, they have an even greater chance of success. They have the opportunity to understand your lifetime value, your potential profitability, potential needs, preferences; and if they are really clever, they can predict your purchase triggers. They may also be on their way to maximizing their efforts with CURARE. 1
The Organizations Unique Differentiators
All businesses have a range of unique differentiators. Arguably their most valuable differentiator and asset is the data they hold on their customers. It then rests on their ability to use that data to create individualized services that are unique to their own customers. They can extrapolate that understanding to customers that look like their good customers and move them to the customer acquisition list. However, the major problem for businesses is to understand what the data they have really means and turning that meaning into actionable information.
Advertisement
The supermarkets introduced the so-called loyalty card for more than just providing a small discount to the consumer. They hoped to get some "loyalty" and some useful data to analyze purchase profiles and customer characteristics.
Richard Neale of Business Objects, a software company based in San Jose, California, tells the story of a British supermarket that was about to discontinue a line of expensive French cheese which was not selling well. But data mining showed that the few people who were buying the cheeses were among the supermarket's most profitable customers - so it was worth keeping the cheeses to retain their custom.2
Telcos are attempting to understand their customers through analysis of the data contained in call detail records (CDRs). These records are created automatically whenever someone or something uses a phone link and include every aspect of a call. Generated by the millions each day, these records are used for billing and other purposes. Financial services companies continually collect data for regulatory purposes and to use in their customer relationship management (CRM) applications.
Magical Software
Organizations are often depicted to be awash with data but without the right information available to run their business. Some collect all sorts of data for years and years, across many disparate systems and then hope that some magical software product can hook them all together (maybe magical middleware that we have been promised for years). Some of these magical tools are often nothing more than glorified report writers or report cubes that enable only the simplest of queries to be answered by the creation of two-dimensional reports.
The different systems an organization uses define and describes transactions, entities and relationships differently. Some have found their systems use multiple naming conventions and hidden encoding structures, and sometimes crucial historical data was only retained for a short period. One bank, when putting together its customer information file for its first data warehouse, found it had millions of customers more than it possibly could have. Each disparate system recorded customers as unique entities with specific attributes which created duplicate, triplicate and in some cases even more entries related to the same customer. The customer file had to be cleaned (data cleansing) and merged to reduce the repetition so that one version of the truth could be created.
Why Bother Using Data Mining - To Make CRM a Reality
Why bother using data mining? Because you need to get closer to your actual and potential customer and given the large number of relationships an organization must manage and the readily available technology you can! It has always been good form to have a customer focus. We have been badgered for years by the marketing gurus that understanding the customer and only making goods or providing services they will buy makes good sense. For many, it is making the customer the heart of the business and in that way they will be the focus of the organizations strategy and business tactics. The manifestation of that has been the growth and hype surrounding customer relationship management (CRM). CRM solutions are not about call/contact centers and sales force automation (SFA) but about building organizations that focus on the customer to enable a "potential" one-to-one relationship.
For many the concept of mass customization or the segment of one may be enough, but the technology today enables organization to move toward one to one. In addition, the tools and approaches available enable an organization to work out whether it is worth the effort and cost to actually service customers one at a time. To satisfy the right the customers and suspects desire for personalization many tools, techniques and physical activities must reengineered, retuned and rebuilt to meet the needs of the arbiter of all profit the consumer.
To get closer to the consumer and business customer, to understand what they want, who they are and predict what they might want at the right time, in the right place at the right price requires a little analytical effort - hence, data mining.
Clickstream and Web Traffic Analysis
Page 1 of 3.
