Database Marketing: Digging for Business Gold

The way in which professional marketers interact with their customers has been transformed over the past few years. You just can't afford to take a customer's continuing business for granted in the 21st century. Today's customer relationships are a lot more complicated than they used to be. Why? A few factors:

• Compressed marketing cycle times. Customers' attention spans are short and getting shorter, and loyalty is much harder to earn in this age of real time information flow. You must reinforce your value proposition on a continuing basis. And you must do so quickly or the customer will hit the road.
• Increased marketing costs. Everything costs more nowadays - printing, postage, special offers. And the most cost effective direct marketing channel is increasingly the Internet.
• Streams of new product offerings. Customers want products and services that precisely "fit" their needs. One consequence has been a proliferation of the number of products and offers available to the customer. The competition has multiplied like so many rabbits.
• Niche competitors. Your best customers also look good to your competitors. They are focusing on small, profitable market segments and will do all they can to take customers away from you.

Sounds pretty brutal, and it is. However, astute marketers are learning how to cope with the new situation by using powerful tools such as database marketing and data mining.

What are we digging for anyway?

Years ago, as an academic researcher dealing extensively with statistics and data analysis, I crunched numbers like there was no tomorrow. Later, I applied those skills in the field of market research, using techniques ranging from simple crosstabs to logistic regression to discriminant analysis to multidimensional scaling. Whatever it took to get to the bottom of things. But that took a lot of hard work, technical know-how, and infinite attention to detail. It was nose-to-the-grindstone intellectual labor.

What data mining does is essentially automate high-powered statistical analysis for practical marketing purposes. Following are a few typical questions that data mining can address:

• Which customers are most likely to drop their cell phone service?
• What is the probability that a customer will purchase at least P 2,000 worth of merchandise from a catalog?
• Which types of prospects are more likely to respond to a particular offer?

Accurate answers to such questions can help you retain customers, increase campaign response rates, enhance cross-selling, and maximize return on your marketing investment.

Traditionally, direct marketers have targeted their offers based on demographics such as household income, occupation, or family structure. However, that approach requires sweeping assumptions. For example, you assume that because a person or household has certain attributes, they're likely to make certain types of purchases. However, data mining gives you a much more powerful set of targeting methods. If your company sells computer accessories and you know that a consumer has recently bought or intends to buy a home computer within six months - now that's a prospect!

How it Works

The basic principle is that one of the best predictors of future behavior is past behavior. Data mining relies on sophisticated statistical algorithms to profile customers to identify the "hottest" prospects. In practical terms, data mining models generate scores - typically numeric values - and assign them to each record in the database. The score reflects the probability that the customer in question will exhibit a particular behavior, whether that be responding to an offer, switching to a competitor, or upgrading to a more expensive product line. By ranking the data set on the scores, you can select the sample for a highly targeted marketing campaign.

Today's data mining software works in tandem with campaign management software. These programs can deliver timely, pertinent, and coordinated messages and value propositions to customers and prospects. The software monitors customer communications across multiple touch points, including direct mail, telemarketing, customer service, POS, interactive web contacts, in-person sales contacts, and so forth. Campaign management automates and integrates the planning, execution, assessment, and refinement of highly segmented campaigns in a way that would have been unimaginable even five years ago.

Technically, data mining represents an extension of traditional statistical analysis, with some artificial intelligence and machine learning twists thrown in. Today's technology automates the data mining process, integrates it with commercial data warehouses, and presents it in a relevant way for business users. When properly implemented, data mining incorporates the ability to access and analyze data stored directly in a database. To work effectively, the marketing data must be in a format that can be dynamically accessed and usable in a relational database context.

Here's an example. A catalog retailer needs to decide who to send a catalog to. The information consists of a historical database of previous mailings and various attributes of customers - things like age, geographic residence, gender, and previous purchase history. The database software builds a model of customer behavior that generates a mailing list of customers most likely to respond to the new catalog.

Ideally, data mining extracts information or findings from a database that you didn't even know existed. Relationships between variables and customer behaviors that are non-intuitive are the real marketing jewels. However, precisely because such findings are unexpected and out of the ordinary, there is a tremendous art and science involved in turning the output of the data mining process into a viable business/marketing solution. The findings have to be solidly placed into strategic business context.

As database marketing becomes more common, marketing managers should expect to work with an increasing number of suppliers (folks like web site hosts, data providers, statistical modelers, and advertising specialists in various media). Given the level of automation in today's systems, statistical analysts are not as much needed as they were in the past. However, the value that a skilled analyst adds cannot be programmed out of existence. You still need experts to accurately assess model results and validate the plausibility of predictions. Data mining software lacks human experience and intuition, and a computer can't tell the difference between a relevant and irrelevant correlation.

You don't need expertise in all facets of database marketing to use data mining techniques effectively. You don't have to be a computer programmer, a web master, or a high-powered statistician. But you must master direct marketing skills and understand concepts like break-even points and the Lifetime Value of a Customer. And you have to recognize the importance of a good offer, good creative content, and finding the right market segment.

The Future: CRM and Database Marketing

Effective relationship marketing requires a marketing orientation across the entire firm. The customer must be central to the organization's thoughts and actions. This goes far beyond the old fashioned direct marketing philosophy (How do I improve the ROI on this specific campaign?). Instead, a relationship marketing perspective looks at customers as an asset across the board (How do I build these relationships and increase the lifetime value of my customers?).

There are at least two other important prerequisites for success in database marketing:

• A strategic perspective Marketing goals must be aligned with overall strategic objectives, and everybody has to be clear about them. Firms that consider strategic, multi-year variables when evaluating the returns on database marketing will be better able to support sophisticated database marketing systems. This will help them create a sustainable competitive advantage. In contrast, short-sighted firms that require early financial returns for database marketing efforts are placing an unrealistic constraint on the process that will probably ensure failure. This is strategic marketing, not tactical, and you have to have faith in the long-term payoffs.
• Information orientation across the firm Information-oriented companies value information. Such companies have a certain organizational readiness to capture and maintain the correct data - data that are already widely available throughout the organization. Such firms understand that customer-level data is absolutely required for marketing excellence.

Today's powerful database marketing systems empower savvy marketers to collect and analyze customer data to an extent previously unimaginable. By targeting specific benefits to specific customers (micro-segments), you can:

• Niche-market your promotions using the most appropriate channel
• Profile customers and identify prospects who meet certain ideal criteria
• Maximize upselling and cross-selling potential
• Match products and service more precisely to customer needs

In other words, by maximizing personalization you create offers that are more likely than ever before to elicit positive responses from customers. And that's what marketing is supposed to be about.