Why You Don’t Know More About Your Customers

By Alexander J. Black

Our 2006 Customer Intelligence Survey confirmed a key finding from our previous survey: many companies that invested in customer relationship management programs are still not getting the benefits that they expected. The problem with most of these under performing efforts has been an over emphasis on customer touchpoint automation and an insufficient emphasis on using customer information more effectively to enrich the customer experience. Speedy and knowledgeable service requires the organization to effectively obtain, store and provide access to the wealth of customer information that has been growing in leaps and bounds over the last decade. And in today’s environment of constantly increasing customer expectations, speed and knowledge are requisites for retaining customers.

The barriers to using customer information effectively range from the mundane: incorrect or redundant data, to the complex: determining members of a household and the decision maker in the household. Too often, companies have defined data attributes inconsistently, which ultimately makes it untrustworthy. “A single version of the truth” has become the rallying cry for customer data advocates. But even when companies have trustworthy data, they do not do the kind of analysis that would identify their most valuable customers or those customers that show the highest propensity to buy a given product or even to defect to a competitor.

Not all companies need to do the most sophisticated kind of data analysis, but most companies would benefit by doing more than they are doing now. The astronomical increase in the amount of customer information that is available today, plus advances in data management and data analysis, allow companies to know their customers better and to give them better service.

The problem of untrustworthy data

Banks are an example of how companies can have lots of data about their customers and still not know much about them. Suppose a bank has records for Jane Jones and for J.E. Jones. They are both the same person, but the bank can’t tell that from the names alone. It also can’t tell from the addresses. Jane’s record shows that she lives on Main Street in St. Louis, while J.E. lives on Main St. in Saint Louis. To add to the confusion, there is also a Bridget Jones living at one of those two addresses. Bridget is Jane’s teenage daughter, but the bank has not figured out how to tell that from all of the information that they have on these customers.

The bank has untrustworthy data. Using that data, the bank may send two copies of a marketing brochure to Jane, when it should send only one. It may also send a copy to Bridget, who should not get the brochure at all. These are the kinds of problems a company creates for itself if it doesn’t scrub its data and standardize records on names and addresses. Address standardization problems are very common. Unless a company is rigorous in administering cleansing and aggregating techniques, customer data become redundant and untrustworthy.

Even taking the trouble to scrub data manually may not always work. Suppose I am designing a marketing program and I need to collect customer data. Because my company has not aggregated data into a central repository, I have to pull it from different sources and enter it in a spreadsheet. I also take the trouble to eliminate duplicate names and addresses. The customer data in my spreadsheet probably is the cleanest, most trustworthy data the company has. But when I share that spreadsheet with people in another department, they won’t accept it because the data didn’t come directly from a company system. Then they will start putting their own spin on it.

If you don’t have a single reliable source of data, you don’t have a single version of the truth. Any company that reaches a certain size and carries on a relatively complex business will eventually face the issue of untrustworthy data.

Operationalizing trustworthy data

Untrustworthy data wasn’t a problem most companies foresaw when they first embraced customer relationship management. Broadly speaking, CRM includes sales, service, and marketing, and is about developing relations that will result in profitable acquisition and service of customers. The first CRM software applications, however, were about automating customer interactions, mainly sales force and call center automation. Automating these interactions reduced costs, but did not increase revenues. Expectations for CRM fell short because the early applications offered only touchpoint automation, which produced efficiency gains but not effectiveness gains.

Which is why CRM evolved into the capabilities needed to make customer interactions more productive and profitable. Those are the capabilities to bring more information to the point of customer contact to make that contact mutually beneficial. Bringing all that information to bear is what we call customer intelligence.

Bringing all that information to bear is not an easy task, however. Take the example of a cable or satellite television company that has millions of subscribers. Most of the company’s interaction with its subscribers takes place at call centers, which get as many as a million calls a day. If trustworthy customer data is going to be used to increase revenue, call center employees will have to sell more services. That means operationalizing the data to give employees the tools to do that.

Giving them those tools means thoroughly scrubbing customer data and adding data from third parties. Then the data will have to be run through an application that will assign subscribers to a segment based on an attribute such as their lifetime value to the company. In this example someone who has been a subscriber for five years, has bought pay-per-view and other services, has a monthly bill of $150, has recently upgraded her equipment, lives in a high-wealth zip code, and has a good credit rating will have the highest lifetime value. The final step is to give call center employees a quick way to identify a caller’s ranking —an easily recognized symbol for their value segment — then provide a list of offers to make to callers in each segment.

Centralize to move up in the Customer Intelligence Maturity Model

Why don’t more companies do this? Because they haven’t looked at the interdependencies of customer information. And they have not been able to get past departmental organization issues. Data spans departments and people and therefore a highly integrated interdisciplinary structure is required to take advantage of customer information across the enterprise. An example of the need for such a structure is customer analytics. Our survey revealed that a key to developing and using advanced customer analytics is centralization of the analytics function. 

Typically, four different groups in corporations do customer analysis: corporate strategy, marketing, finance, and IT. Each of these groups has a different reason for analyzing customer data and some are more familiar with the analytical tools. The CEO’s direct reports will get a higher priority than people trying to launch a marketing or customer retention campaign, but it may be the marketing people whose need for data is more important to the company’s competitive position. When these groups operate independently, as silos, they come up with different versions of the truth, which does not benefit the company as a whole.

These are the characteristics of a company that is in the Basic rank of our Customer Intelligence Maturity Model (see Figure 1). We developed the model to classify companies’ capabilities and assess their performance. Companies that want to move up will have to centralize their data management and analysis to avoid the problems just described. That is important in all three dimensions of customer intelligence capabilities. It is important in the Customer Information Integration dimension because it is the only way to ensure a single source of reliable data. It is important in the Customer Insights dimension because it prevents the use of different platforms and eliminates redundant questions. And without coordination across the company, there is no way to do something as challenging as operationalizing customer data.

Getting beyond the lowest level of customer intelligence requires an investment of time, money, and training, but two factors are making that easier: New, more powerful, applications are providing a bigger return on investment and the growing trend toward offshoring will greatly reduce staffing costs. More companies will be making these investments as they realize that it is not their products that distinguish them from their competitors. As products become increasingly commoditized, companies will begin to distinguish themselves by how well they take care of their customers. And how well they do that depends on how much they know about their customers.