Having an integration road map is the best way to navigate through the deployment terrain of CRM analytics. With failure rates of almost 50 percent for implementations of CRM applications and CRM analytic projects, it seems likely that the integration map is the one that many companies leave behind.
Despite that difficulty, the payback of using CRM analytics to identify areas for revenue growth, cost-cutting, customer retention and other business needs will lure more companies to invest in CRM analytics. But the integration effort required to make this possible could cost 10 times the amount of annual sales of CRM analytic software, say some analysts.
CRM analytic software feeds on as much legacy data as it can access, which poses data migration and integrity problems. Data format and middleware standards are vital, as is a rigorous methodology for coding business logic into the software. Failure in any of these areas can doom a CRM deployment, practitioners say.
Success Begins With Girding for ComplexityDavid Gadra planned for the day when customer relationship management (CRM) analytics would arrive at his $5.2 billion office products company. So when the multimillion-dollar IT overhaul known as Project E-Ikon was given the green light last year, he was ready.
Back in 1997, Gadra, CIO at Ikon Office Solutions Inc. in Malvern, Pa., hired developers from Infosys Technologies Ltd. in Bangalore, India, to build an Oracle data warehouse to standardize, migrate and store key data from the "literally hundreds of companies" Ikon was acquiring, he says. Having established standards, policies and processes for migrating external data sources has shaved up to 30 percent off the development time required to integrate analytic tools from Oracle Corp.'s Oracle 11i suite.
Gadra says he learned that if you adopt CRM analytics, you will confront application and data integration problems as well as the challenge of designing business logic into the analytical tools that operate on the data.
He argues that companies that venture into CRM analytics without having their data integration and management strategies in order are "in for trouble." Gadra is particularly emphatic about making certain that enterprise resource planning (ERP) systems are working properly before attempting to use the data with CRM analytic tools. "The CRM layer becomes a giant magnifying glass of an ERP problem," he says.
Data integration topped the list of steps taken by the Texas Education Agency as it ventured intoCRM analytics. The Austin-based agency, which hands out more than $14 billion to primary and secondary schools in the state, began using analytic tools from Pleasanton, Calif.-based PeopleSoft Inc. in the fall to report more than 160 measurements about goals on everything from students' grade-level reading rates to ethnic-group dropout rates.
Dan Arrigona, director of budget, strategy and royalties at the Texas Education Agency, says that before it ran its first report, the agency had to deal with "data coming from different users, legacy student data, financial sources and other multiple sources."
Joyce Mlakar Smith, vice president of customer research at Columbus, Ohio-based Huntington Bancshares Inc., has been using analytics tools from Waukesha, Wis.-based NuEdge Systems LLC since 1996. Huntington uses the software to market services to consumers and small businesses.
For example, the tool has helped the $29 billion-in-assets regional bank identify deposit account holders who have home loans at other institutions, enabling the bank to compete for that business.
After more than five years, the integration work still hasn't ended. Smith says she is always combining external demographic data sources with her in-house data. And she would like to integrate other data from the bank's branches.
Building in Business Logic
Whether IT chooses a best-of-breed product or one that's part of an integrated suite, IT managers still face the thorny problem of business logic.
"On the analytics side, the biggest source of failure has been failing on integrating the business logic into the tools," concludes Gareth Herschel, an analyst at Gartner Inc. in Stamford, Conn. "One out of two large-scale CRM efforts doesn't deliver what users expected." Just charting sales order workflow procedures can be daunting, he says.
And that's scratching the surface. Gadra says, "To e-enable our business processes, we have had to rethink everything we do from transaction processing to working with customers, employees and vendors."
Feeding The Data Omnivore
In a perfect world, data from every conceivable internal and external source would pour into a data warehouse dedicated to CRM operations. From there, data marts would capture subsets of the warehouse data and interact with one or more CRM analytical applications. In our imperfect world, however, the problem of integrating data into a usable and responsive data warehouse is fraught with problems and is often the Achilles' heel of CRM analytic projects.
Analyze Your Staff
Customer data isn't the only thing analytic tools can measure.
Company workers often have comprehensive and insightful data at their fingertips about the customers who call for sales or service. But management doesn't always have information about how those front-line employees are handling those phone sessions.
Spherion Corp., a US$3.7 billion recruitment and outsourcing company in Fort Lauderdale, Fla., had just that problem. "We had lots of tools about customers," says Tom Sultenfuss, a vice president at Spherion. "We wanted a tool that gave employees measurements about how they were doing."
The company bought Emvolve Performance Manager from Performix Technologies Ltd. in Burlington, Mass. The product delivers a real-time score card on each employee's call, measuring hard statistics like the duration of the conversation or "softer" quality issues that managers grade from recordings. Workers can see their performance over time in different skill areas.
Sultenfuss attributes Spherion's 12 percent gain in productivity to Performix's software. "There's no magic," he says. "People get the information and they adjust themselves to perform better."
Integrated Suites Vs. Best of Breed
The inherent data integration problem makes the question of whether to use best-of-breed CRM analytic products or ones integrated within application suites knottier to answer. For David Gadra, CIO at Ikon Office Solutions, who was already running an Oracle Corp. data warehouse, adopting the analytics in Oracle 11i "was very attractive."
Other users argue that the nature of the data needed to feed CRM analytic software, from e-mail to call reports written on personal digital assistants, is too diverse and unstructured to be managed in an integrated way.
"Best-of-breed apps will do best," says Tom Sultenfuss, vice president of operations and customer development at Spherion Corp. "There's just not much of a convergence of data sources, which means a tremendous amount of integration work still has to be done."
Going with vendors of best-of-breed applications has risks in a technology segment that's consolidating, but analysts and users agree that IT can minimize those risks by:
-- Defining standard data formats, such as XML for data presentation or Open Database Connectivity for database-to-database exchanges.
-- Ensuring data integrity through proven and published processes.
-- Establishing data migration processes, such as storing procedures for graphical data.
-- Choosing CRM analytics tools that support Web browser access.
Oracle packages its business intelligence products with Oracle E-Business Suite. It offers impressive integration with Oracle applications.
PeopleSoft Inc. in Pleasanton, Calif., has been in the analytics game since 1997. It targets vertical markets for "cross-chain analysis."
Siebel Systems Inc.'s offering lacks financial, manufacturing, human resources and other applications; however, the San Mateo-Calif.-based company's vast number of integrated vertical market offerings make it the next closest thing to an integrated provider.
Dozens of small companies offer tools for a wide range of niches (for a listing, see www.computerworld.com/q?26892). Guy Crease, an analyst at Aberdeen Group Inc. in Boston, says that with so many business models in so many industries, specialized CRM analytic tools will always be necessary. Many are compatible with the integrated enterprise application suites.