One-to-One: Half Your Data Is WastedFor all of the conversation about creating a one-to-one relationship with customers, many managers remain frustrated with how far their companies are from this closer level of relationship with customers. Like "buy low, sell high," one-to-one is an ideal that is easy to say, but hard to do. It begs the question, "How do I actually get there?"
Obviously, serving customers one-to-one starts with understanding customers as individuals instead of poorly differentiated groups that lump together millions of people. How, then, can a company develop and act on the sort of rich customer profiles needed to understand customers at this individual level without losing the efficiencies that make serving your customers profitable?
The answer, surprisingly, is usually not to collect more customer data. A survey by Gartner Group, Stamford, CT, found that 80 percent of companies analyze less than half of the customer data they already have. Most companies collect more information about their customers than they can use but are under equipped to understand and act on the results.
One of the quickest ways to alienate a customer is to collect masses of personal data, and then behave in ways that clearly show you have not used it to gain any real insight into the person. When a telemarketer from a company I do business with calls to sell me something irrelevant, I am tempted to shout, "You know my age, my lifestyle and income, to whom I owe money, and the best you can come up with is interrupting my dinner to sell me something it should be obvious I don't want?"
Is part of consumer concern over privacy not just from the amount of information we collect, but also from irritation that we do such a bad job of using it for the customer's benefit? The difference between invasion and invitation often lies in whether we use data to understand our customers well enough to contact them in ways they appreciate and welcome.
Asking how many pieces of data are needed to develop an effective customer profile is a bit like asking, "How long is a piece of string?" The answer is always, "It depends." But clearly, before collecting more customer data, companies should enrich their customer profiles by improving their ability to analyze and act on the data they already have.
Most companies with any online transaction-processing system already have all the raw data they need to form a rich, useful profile of each customer. However, aggregating this data into a data warehouse is only the beginning of the process of putting it into an exploitable form.
Andersen Consulting divides the customer relationship management life cycle into three stages: integration, analysis and action. Integration is the process of pulling data from online and offline sources into a central repository such as a data warehouse. Analysis of the data provides the customer insight needed to develop strategies for handling each customer. Action is the execution of those strategies. Andersen cites analysis as the phase most crucial to CRM success, and the least developed.
Why is analysis so critical? Because developing a living, actionable profile of each customer has less to do with the number of fields or quantity of transaction data than with the company's ability to conduct analysis down to the individual level and to understand the meaning of changes in behavior data.
A single item of information, such as a customer's purchase frequency, can be incredibly revealing. A customer who had been buying twice a week for two years and has dropped to once a month needs to be treated differently.
Changes in address, shifts in the products being purchased and other information would help round out the profile and understand why the change in purchase frequency has occurred. But it is the analysis of how individual customers have changed their behavior over time that alerts the company as to which customers need immediate, individual attention.
How can you evaluate your company's ability to collect and analyze customer data? By looking at your ability to answer difficult, customer-centric questions such as: Which customers have changed the most today and why? Which customers who first bought during the holiday season last year never made a second purchase? Which customers who were very high-value to the business a month ago are now buying less often?
These questions differ from the much simpler queries you might use for traditional one-to-many marketing in two ways. First, they always refer to specific, individual customers. Therefore, to answer these questions and act on the results, a company must be able to identify and analyze customer behavior down to an individual level, but then dynamically assemble these customers into groups. Second, these questions center on changes in behavior -- changes revealed by analyzing how individuals act over time.
Tracking individuals in time is vital to developing one-to-one relationships, as one leading dot-com recently discovered. The company had assumed that it was successful in retaining its highest-value customers because the total membership of this high-value segment was twice as large as it had been four months earlier. Analysis of the individual customers in this segment, however, revealed that most of the individuals who were in the segment four months earlier had since departed. A large influx of new customers was temporarily masking a serious retention problem.
This sort of time-series analysis is difficult to bring down to the individual level. But companies that want to move toward one-to-one must develop or acquire this capability. Whether you are analyzing a customer profile consisting of five fields of information or 100, there is simply no good substitute for understanding change. Without the ability to analyze changes in individual behavior over time, no number of data fields will reveal the evolutions in individual behavior that present your business with opportunities and threats.
•Charles Nicholls is president of Ithena Inc., San Jose, CA, a developer of e-customer intelligence analytic applications. Reach him at email@example.com.