For the rest of the August 2009 issue of CRM magazine, please click here.
Barrie Markowitz, managing director of membership for the United States Tennis Association (USTA), says she had an intuitive sense that her 730,000 members were not all the same—but that’s exactly how they were treated when she first took over the position.
“I knew that we could be more profitable if we were targeted and spoke to people in the right channels,” she says. “We simply weren’t being as efficient as we could be.” The USTA, she says, discovered that the path to that efficiency was lit by predictive analytics (PA).
The USTA isn’t alone. According to the “Predictive Analytics Business Applications Survey” by Predictive Impact, 46 percent of respondents cite “enacting decision automation” as the reason for looking to implement a PA solution. Seventy-five percent confide it was to “obtain strategic insights.”
Those decisions and insights are increasingly sought by line-of-business users. As a result, says Rebecca Wettemann, vice president at Nucleus Research, the business user has a shifting focus when it comes to PA. “More people in the organization [beyond data analysts] are actually able to make decisions based on analytics,” she says.
Wettemann says investments will continue to pour into PA despite the recession, due to the potential competitive advantage provided by PA-powered business decisions. Companies are prioritizing retention of the most-profitable customers, and “PA is good at helping to manage that,” she says. “On the opposite side, you can also predict where there may be fraud or risk.”
Despite the presence of several pure-play PA vendors worth keeping an eye on—see sidebar “Vendors to Consider,” below—Wettemann and James Kobielus, a senior analyst at Forrester Research, both point to SAS Institute and SPSS as the market leaders. Wettemann says, however, that the two PA giants offer very different strengths: Users choose SPSS for flexibility and ease of deployment, while SAS is “more directed in what it provides to customers.”
One area in which Wettemann sees SPSS on the leading edge is in predictive models that integrate unstructured data. “It’s not just about measuring data about transactions, but what customers say about [a company] on the Web,” she says.
Kobielus says that, as more companies turn to PA to improve their decision-making abilities, business intelligence (BI) vendors are looking to add the technology, either natively or through acquisition. Kobielus cites Business Objects as an example: Before being acquired by SAP in October 2007, he says, the BI player could have acquired a data mining vendor but instead chose to license a toolkit from SPSS. “PA really depends on much of the BI tooling and infrastructure,” he says. “There are a lot of synergies between [the two].”
Synergies, perhaps—but some serious differences in price. A high-end toolkit, including data mining, PA, and statistical analysis, can cost between $5,000 and $10,000 per seat, Kobielus says. “It’s considerably more expensive than a standard BI per-seat license, which can be anywhere between $500 and $1,000,” he says. “This is much deeper, specialized tooling used by more-advanced users.”
While it may seem expensive, particularly when most companies are being told to do more with less, Wettemann says that PA is increasingly becoming an integral part of the entire customer experience, whether through having a better touch with consumers or avoiding customer fraud. “It’s going to continue to grow and be a really interesting area,” she says.
The USTA’s experience might be an indication why: Markowitz says that PA has brought clarity—and quantitative analysis—to the USTA, which has helped the organization better target its members and thereby improve retention. Predictive analytics, she says, revealed “things we inherently knew about the business—but now we can prove it.”
SIDEBAR: Vendors to Consider
- Angoss Software Corp. www.angoss.com
- Chordiant Software www.chordiant.com
Product: Cx Decision Hub
- Fair Isaac www.fairisaac.com
Product: Fair Isaac Predictive Analytics
- Insightful www.insightful.com
Product: S-Plus 8
- KXEN www.kxen.com
Product: KXEN Ultimate
- SAS Institute www.sas.com
Product: SAS Analytics
- SPSS www.spss.com
Product: PASW (predictive analytics software)
- ThinkAnalytics www.thinkanalytics.com
Product: Think Enterprise Data Miner
- Unica www.unica.com
Product: Affinium Model
- Visual Numerics www.vni.com
Product: IMSL Libraries
Source: James Kobielus, Forrester Research
SIDEBAR: Questions to Ask
Rebecca Wettemann, vice president at Nucleus Research, argues against the uninformed purchase of predictive analytics tools. The following three essential questions can “lead to a solution that better matches an organization’s needs,” she says.
- What is the business problem I’m looking to solve with predictive analytics?
- Which person or team is tasked with solving that problem?
- What uses of predictive analytics have proven to be successful with similar companies?
SIDEBAR: Green-field Opportunity
According to a recent survey conducted by Predictive Impact, 51.5 percent of respondents have never deployed predictive analytics—but that may soon change.
Time Frame for Predictive Analytics Deployment
In the next 6 months 51.5%
In the next 7–12 months 18.2%
In the next 1–2 years 9.1%
In the next 3–5 years 6.1%
No plans to deploy 15.1%
Source: “Predictive Analytics Business Applications,” Predictive Impact
Contact Assistant Editor Christopher Musico at cmusico@destinationCRM.com.
You may leave a public comment regarding this article by clicking on "Comments" at the top.
To contact the editors, please email editor@destinationCRM.com.
Every month, CRM magazine covers the customer relationship management industry and beyond. To subscribe, please visit http://www.destinationCRM.com/subscribe/.