Today, CRM is thought of as a combination of business processes, technology solutions and advanced analytics that allows companies to understand customers from a multifaceted perspective. That understanding helps build deeper and more profitable customer relationships.
This is a high-tech approach, but at heart, customer relationship management is an old-fashioned idea. Personalized interactions date back to time immemorial.
When commerce was localized, with only one customer touch point, the corner shopkeeper knew all of his customers by name. He knew their buying preferences. He knew their hobbies, lifestyles, occupations, familial relationships and life-stage situations. He kept these and countless other bits of information about each customer in his memory. The information was readily accessible as needed.
Today, we live in a world where most companies of any significant scale can no longer connect with customers one-to-one. Companies have come to rely on go-betweens to do the connecting for them, like marketing vehicles to mediate communication flows, and contact management channels to mediate service and support flows.
There are three important things to consider in your CRM program.
-
How to identify individual customers.
-
How to gain relevant knowledge about individual customers.
-
How to cross-sell to individual customers in a real-time and context-sensitive manner.
Identify Individual Customers
The corner shopkeeper recognized his customers the moment they set foot in the door. This would be just about impossible for the vast majority of retailers in today’s urban shopping malls.
Loyalty cards are one way retailers try to identify their customer base. At one time an exclusive staple of large grocery store chains, today practically every retailer under the sun offers a loyalty card, using any number of incentives to entice consumers to sign up. The problem is that wallets and key chains can only hold so many pieces of plastic. For consumers today, the invitation to rack up bonus points has lost much of its appeal. This is especially true when the invitation comes from a store they patronize only infrequently or where they make only small purchases.
However, these are the very stores most eager to reach out to these customers and to increase their share of wallet. So what’s a retailer to do?
One answer may be a data-appending solution that uses a matching algorithm to obtain demographic information for individual customers based on data captured at the point of sale. Simply put, the device reads the magnetic strip on the customer’s credit or debit card to capture his or her name. It then marries that name to the ZIP code and/or phone number keyed by the cashier. Instantly, the solution turns around an identification of that customer’s “best” mailing address, with an accuracy of better than 90%.
Now, at long last, retailers can unveil the identity of that “masked man.” They can combine his purchase history and contact information. By enhancing the value of this information using predictive models — and by integrating it with customer data from additional sources — he can be presented with targeted offers likely to reflect his specific needs and situation.
It’s good news for retailers. But what about consumer packaged-goods companies, which operate through a variety of reseller channels? How can they begin to identify their end customers when these individuals don’t interact and transact with them directly? Here the answer lies in direct-to-consumer campaigns. And by now, of course, most major brands have dabbled in any number of offline-to-online name-capture experiments — often in the form of high-profile, heavily promoted sweepstakes games — with varying degrees of success.
In the end, a campaign’s effectiveness hinges on more than just the collection of names and addresses. What really matters is whether this information can be augmented with relevant customer data to create meaningful segmentation schemes and insights that truly can be acted on.
Know Individual Customers
In that largely bygone era of localized commerce, the corner shopkeeper got to know his customers well by talking with them regularly — often daily. How can companies duplicate that trick today? How can they gain insights above and beyond what can be gleaned from transaction history and the geo-demographic information commonly available from data providers?
The answer is to start a conversation. After all, building a relevant customer profile usually requires the use of both explicit data (information collected through direct customer response and entry) and derived data (new insights developed through analysis and modeling). Dynamic survey engines can facilitate the acquisition of explicit data, provided the customer is given a compelling reason to reveal such information.
When designing a response survey, companies have to determine what they still need to know about their customers. What missing data fields need to be filled in? Does the program’s success hinge on knowing their food allergies? Their professional aspirations? Their favorite rock bands? Whether they practice yoga? Whether they buy presents for their dogs? After that it becomes a matter of devising an effective strategy for persuading people to provide the missing information.
In many cases, the initial rounds of a precision marketing program may be devoted solely to capturing the right information, in advance of any actual selling activity. The fact is, you can’t buy a mailing list with the names and addresses of people who earn over $75,000 a year and also like to arrive at airports at least two hours early. Or parents with young children who dine out at least twice a week. Or men who snack on low-fat ice cream while watching late-night TV. Knowing these seemingly trivial pieces of information may be absolutely crucial.
Sophisticated survey designs may deploy discreet choice analysis. Marketers can use this method to simulate all potential offers and determine which ones a customer would be more likely to respond to. For example: How important is it for you to have…the best price? A good return policy? A knowledgeable sales representative? The key is to offer choices at multiple iterations. By asking the same questions in different ways, it becomes possible to isolate the different elements at the forefront of many purchasing decisions. In this way, technology and analytics can begin to simulate the corner shopkeeper’s ability to capture relevant, firsthand knowledge about individual customers.
Cross-sell
The corner grocer could make helpful recommendations to individual customers based on his understanding of who they were and what items best fit their needs and situation. How can companies today make context-sensitive offers? Here again the answer lies in technology and analytics.
Collaborative filtering is one common approach to building a recommendation engine. By sifting through user profiles and usage patterns, it allows a company to serve up products or services to a particular customer based on the preferences of other customers with similar tastes or interests. We see recommendation engines used all the time in online book- and music-buying environments.
Now physical stores that carry all types of merchandise have a similar capability at their disposal. Here the technique uses association rules to discover affinities in transaction data between items often bought together. It allows companies to identify the frequent item combinations that customers purchase, and also the timing between these purchases. While some item combinations are obvious (flashlights and batteries), some are less so (like blenders and salsa music). These can easily be discovered through transaction data analysis.
Retailers that act on these insights can optimize in-store product placement (the right product at the right location) as well as post-sales marketing follow-up (the right offer at the right time). Last week, for example, you bought a digital camera; next week, you receive a coupon for a photo printer, photo paper and ink. Offers also can be made at the point of sale using a rules-based decisioning engine, in which case the cashier receives a pop-up coaxing her to make the specific product recommendation.
Again, it’s something the corner shopkeeper could have done — and without the help of transaction analytics and rules-based decisioning engines!
Maybe there’s some truth in the old adage that the more things change, the more they stay the same. In our complex world, it takes technology solutions and specialized expertise in advanced analytics to approach the level of CRM that was so easily attained in a different era on a far smaller and more localized scale.
Jeff Zabin ([email protected]) is director of marketing for Fair Isaac Corp.’s Global Marketing Solutions division in Oakbrook Terrace, IL.
