Saturday, April 28, 2007

Customer Segmentation: Think Beyond What You Can See

Article published in DM Direct Special Report
August 17, 2004 Issue


By Shailendra Kumar

The key to business success in today's competitive world is managing customer relationships. Customer segmentation is the first step toward this customer understanding, which allows organizations to build healthy relationship with their customers.

Know thy Customers

Due to the relentless competitive pressure currently being experienced in the financial services, retailing and telecommunications sectors today, a critical issue for business is to establish customer intimacy to maintain better customer relationship.


To build these relationships profitably certainly depends on treating different customers differently. Any organization must understand the customer history and its value to the business, both today, in the past and the future, so that all interactions with the customers are appropriate and lead to the relationship being profitably developed over the period of time. Customer segmentation is the key to this level of understanding.

Customer Segmentation

Customer segmentation is a powerful technique which divides a group of all customers into smaller subgroups that share a number of properties in their relationship with the business. It is obvious that these subgroups should be given very different treatment.

A simple customer segmentation can be produced easily with the analysts exploring the data and applying their understanding of the business. Typically the process begins with a business problem in mind and attempts to determine which measures of a customer are important to that specific business problem and how the chosen measures can be used to segment the data. This is a highly iterative and experimental process, with the analysts constantly increasing their insight into the behavior of the customers. For example, an analyst may want to segment the customer base, geo-demographically and then subsegment any one region by the amount of revenue generated by that region.

Segmentation is now an essential part of the marketing activities of most of the organizations, influencing key strategic decisions. Segments makes analysts think beyond what is there to see.

Practical Issues

There are a number of practical difficulties for analysts to implement the process of segmentation. By its very nature, this type of work needs to be carried out by a business, not a technology expert. A user-friendly, easy-to-use business tool may provide more specific support to help an analyst construct queries and experiment with different approaches to segmentation.

The complexity of the ever-increasing number of queries makes it difficult to track customers in a segment over time. As the customers' data changes, they may no longer fall inside the criteria of the original query for that segment - which makes any form of "then and now" analysis very difficult. For example, the analyst may want to track the response of a segment of customers who have been targeted with a direct marketing campaign or follow the changing composition of a particular segment over time.

In fact, a marketing analyst may have neither the time nor inclination to write his/her own queries. The tasks involved in retrieving customer records from the database and building queries for segmentation are often left to IT specialists, based on the analyst's specification of whom should be in each segment. Not only does this increase the time needed to produce the segmentation model but effectively precludes the analyst from iteratively refining the model. Time constraints will not allow for many cycles and the analyst will lose his/her train of thought while waiting for models to be produced by IT support. Thus, it is suggested that it should be left to the analyst to access their own record set from the customer database and do the segmentation as per their requirement.

Another practical issue is the dirty data which does not allow the IT specialists to give data access to the marketing analysts. A tool which can segment customers and even handle dirty data is the effective need of the hour.

Business Benefits

A very significant part of the value of performing customer segmentation is in providing a common language that the entire business can use to talk about its customers. Indeed, segmentation is an excellent example of the way in which data can be turned into information and then into actionable knowledge. This information is relevant even to the most senior managers where a monthly board report, for example, could include the performance of the product by customer segment over a period of time. Customers can then be tracked as they move from one segment to another, as their relationship with the business matures or, perhaps, begins to go sour.

Customers in one segment with a similar geo-demographic profile to those in another higher value segment can be encouraged to migrate to the higher value segment. Market research can be carried out by segment to determine not just what the customers are doing, but also why they are doing it. New and prospective customers can be matched with known customers based on geo-demographic information in order to include them in an existing segment and understand their likely future behavior.

Today, the need is to provide marketing analysts with a business tool which is quick, can handle large amount of data and can help them do segmentation in simple English language on the customer base along with their transactions. This will make analysts more creative and innovative to follow their "train of thought" and allow them to take those decisions which they wanted to take today and not tomorrow.

source: http://www.dmreview.com/






No comments: