The basic architecture of a customer relationship system is familiar to most people who care about such things. Picture a dumbbell: at one end are touchpoint systems such as call centers and Web sites that interact directly with customers. At the other end is an analytical database where data is gathered and results are assessed. Connecting the two is an intermediate system that transmits data from the touchpoints and sends customer strategies back to them.

Touchpoint systems and analytical systems have traditionally been produced by different vendors and run by different corporate departments. As the drive for unified customer management brings these two sets of vendors into contact with each other, the intermediate system becomes a key battleground in their struggle to control the entire customer management process.

The specific issue is how customer strategies are deployed to the touchpoints. There is no real question that the touchpoint and analytical systems both need to exist as separate entities; the touchpoint systems require normalized data structures for real-time transaction processing while the analytical systems require consolidated, denormalized data structures for segmentation and reporting. The analysis of customer behavior necessary to design and later evaluate customer strategies clearly belongs in the analytical system. So does the assignment of individual customers to primary segments based on deep historical data. (Primary segments reflect fundamental customer attributes, such as expected long-term value, that change slowly if at all.) But where will the segment codes reside so they can be used to drive day-to-day interactions?

The analytical vendors would argue for a separate interaction management system that holds its own customer database and whatever campaigns or business rules are needed to apply customer strategies to touchpoint interactions. This system would be independent of any individual touchpoint but attached to all of them. Thus, it would provide a single place to define and deploy campaigns, ensure consistency across touchpoints, ease access to advanced tools such as real-time scoring and allow best-of-breed customer management regardless of the capabilities built into individual touchpoint systems.

The touchpoint vendors would argue that the segment codes should be posted directly into their operational databases and that customer management rules should also be built into their regular business processes. Working directly against the live operational data would ensure decisions are based on the most current information without the burden of copying each transaction into a separate interaction management database. Similarly, avoiding a separate interaction management system places all the rules governing each transaction in a single location, rather than splitting them between the standard operational processes executed by the touchpoint and the marketing embellishments added by the interaction manager. This makes it easier to manage the experience that is ultimately presented to the customer and saves the effort of inserting links, such as Web page "slots," that call the interaction manager from within touchpoint processes. It also avoids the nagging headache of ensuring the marketing messages called for by the interaction manager actually have been created and deployed in the form of content – such as Web pages, e-mail text or call center scripts – available to the individual touchpoints. Of course, these arguments are most convincing when a single operational system delivers messages through all the different touchpoints. But this sort of cross-media "customer interaction center" is exactly what leading touchpoint vendors are now trying to sell.

Which approach makes the most sense? The answer will be perfectly clear to people who work for one type of vendor or the other, but not to marketers or IT managers who lack a strong vested interest. In theory, the advantage lies with the independent interaction management products: these can be expected to provide the most powerful campaign definition tools and tightest links with the back- office analytics needed for true relationship optimization (but that's another column). However, the reality is that today's interaction managers are far from mature. The most sophisticated campaign management systems, from vendors such as Xchange Inc., Prime Response and Recognition Systems, were built primary for batch-selected outbound campaigns rather than real-time interaction (although they are all adding real-time capabilities as quickly as they can). Products built explicitly for real-time interactions, such as E.piphany Real-Time Personal-ization, Yellow Brick Experience Manager and Harte-Hanks Allink Agent, are limited in the sophistication of the campaigns they support. These systems can also be expected to improve over time, but today are not necessarily more advanced than the (also limited) campaign managers built into touchpoint systems.

This doesn't necessarily make touchpoint systems a superior choice. Only a few touchpoint products can truly support multiple channels or provide respectably powerful campaigns. Just a handful of vendors – most notably Siebel Systems – have attempted any meaningful integration with a proper analytical database. Apart from some Web-specific products, it's hard to think of any touchpoint that incorporates automated statistical modeling.

It's tempting to defer any decision until a superior solution emerges, but most companies can't afford to wait. For companies that can find a single touchpoint system to handle all, or at least most, of their interactions, it pays to assess whether this system has the campaign management functions needed to execute meaningful customer strategies. If so, buy it and do analysis and planning offline. If multiple touchpoint products are unavoidable and coordinated interactions can't wait, there is really no alternative to deploying a separate interaction manager with its own database. You will have to wire that system to the different touchpoints and continuously synchronize data, campaigns and content. At least you will have a platform that lets you deploy true customer management and can be expected to grow over time.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access