Your Data Warehouse: A Marketing Information Solution?

Ask And You Shall Receive. Well, Maybe. Maybe Not.

Many companies long ago decided that consolidating their information resources within adata warehouse was the best way to meet their present and future information needs. Thebenefits of integrating information from disparate sources into a single repository werejust too appealing to resist. For the marketing organizations in these companies, having asingle source from which to draw customer data was a dream come true.

For the first time, marketers would have access to the data they needed to targetspecific customers with narrowly focused messages as opposed to the broad, mass marketingapproaches of the past. They would be able to identify a company's best (i.e., mostprofitable) customers and address those customers as individuals. By using customer datamore intelligently, organizations would be able to increase sales productivity, improvecustomer service and increase per-customer revenue and profit.

As more data warehouses come online, marketing organizations are finding that thetypical data warehouse may not meet their needs. For a data warehouse to address questionsabout the profitability of a business, it must usually consider every product, channeland customer, as well as a host of other factors.

Therefore, data warehouses designed to answer profitability questions are among thelargest and most granular. Although summarizing the data can make it more accessible,summarization can inhibit a user's ability to answer ad-hoc questions. In order tosummarize, you must know the questions you want answered as the data warehouse is beingimplemented.

Question Obsolescence

Of course, users of marketing information systems can never predict in advance all thequestions they might ask of the data. Factors such as new products or competitors, changesin government regulations and shifting corporate identities and missions render existingquestions obsolete even as they suggest new ones.

The size and structure of a data warehouse can also limit its usefulness as a marketinginformation system. Data warehouses are designed to take large volumes of disparatetransactional data and do a few specialized things to it repetitively and efficiently,processing it into a consistent framework. Such systems are usually organized in asequential, "flat file" environment. These large, transactional systems cannotafford to have too many users. Therefore, many marketing organizations are turning to"data marts" and "data mining" as a way to provide the customizedinformation they seek.

Data marts and data mining work like Decision Support Systems (DSS) environments inwhich different questions are asked of the data to support decision-making processes.Therefore, DSS environments are more relational. Data is organized in such a waythat it can efficiently accommodate changes in the decision-making process.

Cumbersome And Inefficient

Data warehouses are terribly inefficient and cumbersome when it comes to answeringdecision-support questions. By the time a traditional in-house accounting or operationalsystem can provide a new marketing-oriented report, the opportunity the report was meantto seize may have passed. Relational data marts, on the other hand, are quite inefficientin processing large volumes of transactional data.

The good news, however, is that companies are not necessarily faced with an unpleasant,either-or decision as to whether they should build a data warehouse or more specializeddata mart. In fact, the best approach may be to use their data warehouse as anintermediate storage point to distribute data to various data marts for specializedprocessing and decision support.

One important advantage to this approach is that the data warehouse can apply anenterprise model to the information it receives, reconciling or rationalizing data that isinconsistently recorded, defined or tracked in disparate systems. By cleansing andstandardizing data, the data warehouse enables the data marts that use the information toprovide meaningful and consistent answers.

Buy Or Build

The more difficult issue facing companies in terms of data warehousing and data martsis whether to develop the solutions themselves or outsource the chore to a marketinginformation specialist. The answer depends on two primary factors.

A Corporation's IT Organization. Companies with strong IT organizations may havethe resources and expertise to develop both the data warehouse and related data marts.Keeping development in-house allows greater control of the process, the data and thebudget. However, developing marketing decision support and data mining systems is probablynot the primary mission of most companies' IS organizations. Therefore, they may choose tooutsource such development and keep their IS groups focused on the information needs oftheir core business.

The Relationship Of Marketing And IT. Naturally, IT organizations arecomfortable with the language of databases. They intuitively understand how to structure alarge, data warehouse environment. However, IS organizations are usually less fluent inthe language of marketing and not always aware of a marketing organization's uniqueinformation needs. Unless the marketing organization is intimately involved in the designof a marketing information system, the odds are that the system they receive from IS willbe less than ideal.

If terms such as "customer acquisition," "customer share,""upsell" and "cross-sell" are foreign to the IT group, they shouldstrongly consider outsourcing.

Long Term Storage

Beyond the decision of how the data warehouse and any related data marts will bedeveloped lay decisions on how the systems will be maintained and modified and how thecurrency and integrity of the data will be assured. In all these decisions, theorganization's culture in regards to openness to change, aversion to risk and commitmentto long range goals plays a critical role. When a data warehouse project fails, politics,not technology, is usually the cause.

That's because developing a data warehouse or a marketing information data mart is along-term commitment. Although the payoffs in terms of selling efficiency, customerretention, profitability and market penetration can be tremendous, they won't be realizedovernight. Any information solution implemented requires tinkering as it is put to use andtested in the real world.

Although marketing and sales information systems address areas critical to a company'ssuccess, they tend to provide the quickest return on investment. Despite the challenges indeveloping data warehouses and data marts, when implemented properly, the strongcompetitive advantage they provide will be felt in every area of a company's operation.

--Diane Stuckey is Vice President of Strategic Marketing, IntelliQuestMarketing Information Solutions

(Austin, Texas).

Consultants On A Role?

Consultant Should Be... Percent
Database Designer       51%
Technical Architect 46%
Methodologist 43%
Requirements Analyst 35%
Consultant Should Never Be Percent
Project Manager 32%
Data Warehouse "Evangelist" 27%
Data Administrator 24%
Database Administrator 21%

Source: Data Warehouse 35