More isn't always better. These days direct marketers have oodles of data to choose from, but finding usable information can be a challenge.

Experts suggest looking in the databases themselves — but only if you're willing to embrace new analytic approaches, freshen database design, and reconsider the ways you use customer information.

For starters, you might try studying customer behavior rather than static transaction data.

“The classic case is in the subscription business,” says Anthony Power, senior vice president/chief scientist at Alterian, a database marketing firm in Chicago. “When I subscribe to a newspaper it shows up in a database as a transaction. When I cancel, it shows up as another transaction.”

Behavior, as Power defines it, is observing those two activities together and realizing the value of the resulting metrics. In this case the key questions are: How many times has this individual been a subscriber? How many times has the reader signed up for the premium rate and let the subscription expire at the end of the promotion period?

There's been a move recently toward combining transaction data and uncovering embedded information, says Power. The goal "is enriching what data you have to focus on the behavior the marketing department is ultimately trying to change.”

Then there's the problem of storing and using data from ever-increasing sources. For example, online and offline marketing information initially was stored in different warehouses, but that is no longer an option.

“It's a challenge to integrate online and offline data,” says David Ehrenthal, senior director of integrated services at direct marketing agency PreVision, Lincoln, MA. “For e-mail campaigns you want to pull data from a central database and drop it into an e-mail deploying system. And some systems do not integrate well.”

Ehrenthal continues, “The keys to developing a data warehouse and predictive analytics is the integration of business rules, householding, and data analysis technology.”

In some instances, the barriers to integration can be system-based. Data systems that touch customers — such as call center, point-of-sale and customer service mechanisms — often are either proprietary or heavily modified from a standard design, says Andy Cutler, chief strategy officer at database marketing services provider BeNow, Wakefield, MA. And getting modifications to ease data integration can be tricky, especially when a company has invested in getting its call center reps up to speed on the system currently in place.

New Info, New Problems The online channel provides a new stream of information about customer behavior, but it also creates new headaches. Electronic marketing has heightened both customers' and managers' expectations; e-marketing campaigns operate at a faster pace than traditional efforts, and managers usually want to begin seeing results and running analyses a few hours after launch. For those using a centralized database designed to be updated weekly or monthly, this can be frustrating.

So why not keep data in separate systems? “At some point marketers want to aggregate all the information on the household and develop a single truth about a customer,” Ehrenthal says. “Even though the responses [to an e-campaign] may be in the database, you haven't gone through the process of matching them to households. There's a lot of quality control, a lot of processing, and it's expensive.”

Another disadvantage is that the anonymous nature of Web surfing limits its usefulness as a DM tool. If a marketer is able to require site visitors to register — in return for delivering unique content, for instance — the ability to follow individuals around is hugely valuable in developing well-rounded pictures of customers, according to Cutler.

There's a tangible benefit to linking online prospects with their offline identities. Cutler notes that e-tailers generally pay for referral sites that bring in new customers. If a company is able to prove that someone coming through Google or Yahoo! is a pre-existing customer, the cost savings can be considerable.

Having personally identifiable information leads to the chore of maintaining a database that will allow marketers to swiftly integrate a lot of data. But if this data isn't analyzed properly, results from it can be meaningless.

Therefore, with the exception of a few bold early adapters, the DM industry is taking baby steps. “I don't think [Web data analysis] will go too far in its first iteration,” Cutler says. “Marketers don't need that level of granular data yet.”

Cutler anticipates it will take between one and two years to figure out how to integrate Web data into DMers' offline systems. The current buzz at electronic retailing conferences, he says, is that people are just now realizing they have to do this.

Keeping Everyone Satisfied If processing power is one obstacle to creating a high-powered enterprisewide marketing database, human egos are another. Sometimes resistance to a new data approach comes not from technology but from tradition, according to Carmen McKenna, solution leader for Acxiom Corp. in Little Rock, AR.

“Take retail banking,” she says. “Mortgage information vs. consumer deposit vs. all the other different lines — all that information is siloed. The net-net [of integration] is that you get to understand who your customers are [and appreciate] how you want to interact with them. But marketing budgets are for a specific line of business, each of which is trying to optimize its piece of pie in the bank.”

One trend helping to satisfy all parties concerned is what PreVision's Ehrenthal refers to as the democratization of data. “More and more end users — people who are decision-makers and action-takers — are using the warehouse,” he says. “You need to have a central database, and [information integrated from all sources] has to be consistent across all views of the data.”

Customers vs. Campaigns Databases also have to fuel reports that look at customers rather than campaigns, says Power. The move toward customer centricity did not end with product design. Reporting programs have to be able to segment by a host of demographic, attitudinal or behavioral attributes.

Cutler agrees that reporting tools now have to answer a set of queries they never did before. “CEOs say they're in the business of making products and selling them,” he says. But marketing, Cutler feels, follows Peter Drucker's mandate that the purpose of business is creating and keeping a customer.

This means marketing executives need reports with information that's valuable to big-picture individuals — those who have to demonstrate to Wall Street how increased customer satisfaction or loyalty can translate into top- or bottom-line contributions.

The move toward this type of reporting correlates with the advancement of the chief marketing officer, says Power. “Look at the rise of the chief information officer from vice president of IT. The vice president of IT had to learn a different language and set of issues to earn the right to sit at the executive level.

“We are seeing the same thing in the marketing department,” Power continues. “In terms of database design and structures, the analysis has to be aligned with the metrics the CEO is interested in. How do acquisition programs improve corporate performance [as reflected by] cash flow, equity and balance statements? If you don't speak in that language you'll always be a separate and distinct part of the organization. The data's there, [and now] the technology for analyzing data is there too.”

To this end, many analytic packages now include “validation pieces,” as Acxiom's McKenna calls them — reports that specifically address the concerns of non-marketing managers who often sign off on campaigns. Because for every other theory about database design, there is one long-standing truth: The folks who control the purse strings have to be softened up first.

More isn’t always better. These days direct marketers have oodles of data to choose from, but finding usable information can be a challenge.

Experts suggest looking in the databases themselves