1999 Jun 01

Six Myths about Database Marketing Systems
David M. Raab
DM Review
June, 1999

Myths help us make sense of reality. But while stories can impose a useful structure on the jumble of experience, they can also block us from recognizing unexpected truths. That’s why it’s worth looking at naked facts once in a while to see whether any new patterns have emerged.

In the database marketing software industry, the annual update of the Raab Associates Guide to Database Marketing Systems provides an opportunity for this type of reevaluation. New statistics derived from the 1999 Guide show that many widely accepted beliefs are not really accurate.

Myth #1: sales of database marketing software are exploding. Given all the attention paid to database marketing, the number of new vendors appearing each season and the size of the trade show booths, it’s easy to conclude the market for these systems is booming. But a look at the figures shows that vendors added about 230 net installations last year–a healthy increase over the 160 per year of the prior three years, but far from spectacular. In fact, since the base of previous sales gets larger each year, the growth rate for total installations is actually declining!

(These figures exclude low cost systems sold primarily to small banks. This is a distinct segment that truly is exploding as banks replace obsolete first-generation products.)

Of course, some vendors are growing more quickly than the average. And, as a group, products using a standard relational database are selling faster than systems based on older, proprietary technologies. But now that relational systems dominate the market–accounting for three-quarters of the new installations in 1998–their sales can only grow with the industry as a whole.

Nothing in these figures proves that future sales will not increase sharply. But they call into question any forecast based on the premise that growth has accelerated steadily in the past.

Myth #2: database marketing software is a mature technology. While software vendors certainly continue to enhance their products, there is a sense in the industry that database marketing systems are rapidly approaching commodity status–that apart from a few bells and whistles, most products provide roughly the same functionality. Certainly any stroll through the aisles of a major trade show leaves the impression of products that are indistinguishable to anyone but an aficionado. List prices of new products are considerably lower than those of older systems, further hinting that newcomers are competing primarily by offering similar functions at a lower cost. And in fact, most systems on the market today do follow the same model: they provide complex segmentations against a standard relational database, with built-in promotion history, response analysis and scheduling mechanisms to handle on-going and event-driven promotions.

But things aren’t quite so simple. Systems introduced over the past year include some significant innovations, particularly in tighter integration of statistical models and in data structures to simplify analysis of large transaction databases. More important, requirements for a succeeding generation of database marketing systems have become increasingly clear. This generation will be designed to handle real-time customer interactions and to guide decisions made within operational systems. This represents a radical departure from today’s standard model, which runs its marketing campaigns as independent batch processes that, at best, feed information to operational systems at frequent intervals. The technologies to manage the new, vastly more intimate connection between marketing and operational systems are just beginning to be defined. But they certainly include real-time model execution, enterprise application integration, and automated policy optimization. So while the current set of database marketing technologies may indeed be considered mature, the industry as a whole still has a great deal of innovation ahead of it.

Myth #3: Customer Relationship Marketing (CRM) is just a fancier term for Database Marketing. True, CRM is used to describe the general concept of treating customers differently to generate greater profits–an idea at the core of database marketing as well. But a closer look shows that the term CRM is mostly applied to systems and processes that manage the customer interaction itself–things like call centers, help desks, sales automation, and Internet interactions. The main vendors in the CRM market are companies like Siebel, Vantive, Clarity and Pivotal, who focus largely on delivering the “front office” or “touchpoint” systems, not the sophisticated segmentation and campaign management offered by traditional database marketing vendors. In fact, the main selling point of CRM is that customers will be treated consistently and efficiently because all touchpoint systems will share the same customer contact database. The ultimate promise, of course, is that the touchpoints will deliver the most effective possible treatment for each individual customer. But today, most of the attention is concentrated on the mechanical processes to build and link the touchpoint systems, not on the marketing processes to define the policies they will implement. There is nothing wrong with this: obviously the foundation must be built first, and installing and linking these systems is a Herculean challenge for large organizations.

The problem is marketing database vendors who think they are at the heart of the CRM industry, and therefore conclude they will inevitably participate in CRM’s growth. In fact, today’s CRM vendors already dwarf the marketing database firms–Siebel alone had nearly $400 million in revenue last year, compared with something under $200 million for all marketing database software (excluding services) combined. Business plans that assume CRM will be $5 billion industry in a few years (probably true) and that a particular database marketing vendor will capture a 10% share of its market (plausible), are wrong when they infer that the vendor’s revenue will be $500 million. The fallacy is that the vendor will have a 10% share of something much smaller than the full CRM marketplace.

If the only result of this myth were overoptimistic business plans, it wouldn’t matter much except to the vendors and their investors. But the real concern is that the CRM vendors themselves will decide to enter the marketing database universe. So far this hasn’t happened, presumably because the CRM vendors are busy completing their suite of touchpoint offerings. (Most started out as experts in a single touchpoint, such as sales automation or help desk.) More ominously, the CRM vendors may not have developed these functions because most customers are satisfied with the limited database marketing capabilities the CRM products already provide. This is ominous because it suggests database marketing may not be much of priority for most CRM buyers–challenging the foundation of the database marketing vendors’ hopes.

But let’s assume the demand will eventually be there, once the initial CRM plumbing is in place. When their customers do show an interest, the CRM vendors’ huge resources will make it quite easy for them to add their own database marketing offerings, either through internal development or acquisition. When that happens, marketers who purchased a traditional database marketing system may find integrating it with their CRM system is impractically difficult. This will be a particular risk if the database marketing vendor has not designed their product to avoid the problem.

Myth #4: The Internet is just another touchpoint. By now, most people realize the Internet isn’t just “another” anything, but a world of its own. Still, marketing database systems have so far tended to treat the Internet as another place to deliver messages and execute campaigns–not significantly different from a call center. But the standard database marketing approach to other media is to create a list of messages for each customer and update it periodically–usually daily–based on events. This won’t work for the Internet, which requires split-second reaction to each customer action, with too many possibilities to employ the logic trees used for, say, a telemarketing script. Instead, Internet systems need fully automated decision-making to choose among dozens or even hundreds of alternatives, in the way that an online bookstore might decide which book to recommend.

Making these sorts of decisions requires immediate access to detailed information about an individual customer. This, in turn, implies a data structure build to provide such access–the type of structure used in operational online transaction processing (OLTP) systems (including touchpoint systems), not the structures used in batch-oriented marketing databases.

At a deeper level, the Internet will require rethinking the relationship of the marketing system to the rest of the company. This is really the issue raised in Myth #2–that the marketing system must be one of several that help make decisions made about how to respond to customer actions. Thus, instead of the marketing system deciding on its own what product to offer a customer, it should “consult” with inventory, pricing, credit, and customer service to make the best choice. This will reduce the status of the marketing system to one peer among many. Certain kinds of object technologies are well suited to this approach. But these are not the technologies used in most of today’s marketing systems.

The specific challenge raised by the Internet is that Net-based companies need these solutions now, not in the far-off future when bricks-and-mortar companies will get around to integrating everything with everything else. So marketers seeking an Internet solution must look carefully at what current offerings can really do for them. And marketing database vendors must realize that if they do not address this issue quickly, Internet-oriented vendors will get there first.

Myth #5: the marketing database software industry is a two-horse race. Not surprisingly, this particular myth is most fostered by the leading horses themselves–Exchange Applications and Prime Response. Both are excellent systems; neither is perfect. Nor, in terms of technology, is either necessarily superior to several other contenders (see Myth #2). But these two have done a distinctly superior job of marketing themselves within the industry, to the press and analysts, and to the large hardware and services companies who control access to many projects.

Still, of the 400 or so large-scale relational marketing database installations in place, Exchange and Prime have about 50 each–a combined share of 25%, which is far from a dominant. And they accounted for only one-quarter of the net sales in their segment last year, so their share didn’t grow significantly. More important, they face entrenched competition in several industry niches, including retail, grocery and casinos, and in Europe. They must also contend with upcoming offerings from industry veterans including Harte-Hanks and Customer Insight/Experian, who are finally making the transition from proprietary databases to open solutions. And, of course, there is the specter of future competition from the CRM vendors (see Myth #3) and Internet-oriented systems (see Myth #4).

The danger of this myth (or virtue, depending on your viewpoint) is that it might become a self-fulfilling prophecy. Many buyers prefer to consider only market leaders; a frightening number–particularly those who are new to database marketing and don’t really understand their requirements–will purchase a leading product with little serious analysis of its capabilities. Given that other systems may be clearly superior in a particular situation, and that the current leaders are not engineered for real-time interactions, failure to look past the two leaders would be a major mistake for most buyers.

Myth #6: proprietary systems are dead. Yes, new sales have trailed off for systems using proprietary databases as their primary data store. But proprietary technologies can still execute complicated queries against large files much more efficiently than a relational database. To take advantage of this, several of today’s systems have embedded proprietary technologies in ways that are hidden from users. This may involve proprietary indexes that speed access to the main relational database (Harte-Hanks IRE, Dun & Bradstreet Market Spectrum), a synchronized copy of the database in a proprietary format (Marketing Synergy MSM, TPC Gravity), or a specially formatted extract of the promotion universe (Paragren One-by-One).

In addition, the past year has seen increasing attention to specialized analytical data stores like Broadbase, Digital Archaeology and Sand Technologies Nucleus. These are not generally used for marketing selections. But they illustrate a growing recognition that analysis and execution are distinct functions that can be done against separate databases. This reinforces the expectation that future marketing systems will move away from today’s model of using the same data store for analysis and selections. Instead, they will use the interaction-friendly approach of developing policies on an analytical system and then implementing them through touchpoints that access a separate OLTP-style database.

* * *


Year-to-Year Change in Installations by Active Products (large-scale marketing database systems only)

1998 1997 1996 1995 (18 month period) prior
proprietary 54 38 45 111 86
relational 117 139 102 50 30
combined 231 177 147 161 116
% change 31% 20% -8% 26% (annualized)


Active Installations (figures do not match prior table because of discontinued products) (large-scale marketing database systems only)

1998 1997 1996 1995/4 (18 month period) prior
proprietary 331 280 242 197 86
relational 395 306 171 69 30
combined 726 586 413 266 116
% change 38% 42% 53% 92%

Source: Raab Associates Guide to Database Marketing Systems

* * *

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.

Leave a Reply

You must be logged in to post a comment.