The two vendors lead a market that is moving away from complex data mining workbenches, which can do a wide variety of customer data analysis, to packaged applications that use some of the same analytical techniques but are designed for specific business processes, according to Gareth Herschel, research director with the Stamford, Conn.-based research firm.
A shortage of skills demands the shift in approach. There are simply not enough analysts with the skills to effectively operate the sophisticated data mining workbenches.
"It [doesn't take] a PhD in statistics, as people often joke about, but it does require a fairly skilled user to use that type of application and to use it correctly," Herschel said.
As a result, more companies are turning to specialized customer data mining applications that focus on just one type of analysis – like recommending which products to up-sell or cross-sell to which customers -- requiring less skill on the part of users.
Customer data mining technologies help companies analyze and predict customer behavior to better allocate resources. If a company knows -- based on experience -- that a customer is likely to buy a certain product anyway, for example, it can spend less time and money marketing the product to that customer.
Packaged customer data mining applications take "those generic workbench capabilities and put a more user-friendly wrapper around them, more workflow, more templates, more guidance about how to achieve the specific business objective," Herschel said.
The market for packaged customer data mining applications remains immature but shows significant potential for growth, he said, as companies want the insight provided by the technology but lack qualified personnel to use it effectively.
The market for complex data mining workbenches, by contrast, is still growing, he said, but at a slower pace than in years past.
SPSS, SAS on top
SPSS has done the best job of transferring its legacy data mining workbench technology to packaged applications, Herschel said, landing the Chicago-based vendor at the top of the leaders' quadrant. Gartner's Magic Quadrant methodology places vendors that meet its inclusion criteria into one of four quadrants based on "completeness of vision" and "ability to execute."
Leaders are those vendors that excel in both ability to execute and completeness of vision; challengers have the ability to execute but lack strong vision; visionaries are market-thought leaders, but they struggle with functionality issues; and niche players concentrate on just one or two specific segments of the customer data mining market, but do it well.
SPSS also enjoys a high customer satisfaction rate, the report says, "with the sales process, technical implementation and deployment, user on-boarding and training, and post-sales support all highly rated."
Cary, N.C.-based SAS, the only other vendor in the leaders' quadrant, was cited by Gartner for its experienced and talented staff, its long and established track record, and its efforts to solicit and learn from customer feedback.
"These are very reliable tools," Herschel said of the two leaders. "They'll give you the right answers to the right questions."
Among the challengers, Angoss Software Corp., based in Toronto, has positioned itself as the go-to Software as a Service (SaaS) customer data mining vendor, Herschel said. U.K.-based Portrait Software plc, the other vendor in the challengers' quadrant, is particularly adept at data mining for direct marketing purposes, according to the report.
Rounding out the list, Infor CRM Epiphany, Unica Corp. and Viscovery GmbH were placed in the niche players' quadrant, while ThinkAnalytics Ltd. was the sole visionary.
Pick a target, match tools to skill set
Companies making their first foray into customer data mining should pick an area of analysis that fits their business strategy, Herschel said. "Think about the critical uncertainties that data mining could help us answer," he said, rather than just implementing data mining for its own sake.
Herschel also cautioned that although data mining gives companies better insight into their customers' behavior, it doesn't necessarily tell them what to do with that insight.
"They really need to think about: 'If we use a data mining tool to identify our 1,000 riskiest customers -- those we're most likely to lose in the next month -- what business processes do we have in place to retain those customers?' " he said.
Finally, companies must match customer data mining tools to the skill sets of their users, Herschel said. There's no sense buying a complex data mining tool if nobody has the expertise to use it.