TDWI Survey Shows Growth in Predictive Analytics
More than half of respondents in a TDWI survey say they're actively using or evaluating predictive analytics (PA) technologies. That's big -- but whether PA ever sees truly pervasive adoption is another matter.
- By Stephen Swoyer
- February 11, 2014
Will predictive analytics (PA) ever take off? According to a recent report from TDWI Research, a sharp uptick in the adoption and use of PA technologies could be close at hand.
Whether PA ever sees truly pervasive adoption is another matter, however.
"[Predictive analytics is a technology whose time has finally come," writes Fern Halper, research director for advanced analytics with TDWI, in her new "Best Practices" report, Predictive Analytics for Business Advantage. The stage is set, the market is primed, and companies are hip to the promise and value of PA, Halper argues. "Companies are looking to use [PA] to predict trends and understand behavior for better business performance," citing the results of a recent TDWI survey. What's more, she says, companies are embedding PA in their business processes. "Predictive analytics is also being operationalized more frequently as part of a business process."
Halper and TDWI see uptake of PA as a function of several factors, including a sustained exponential increase in computing power; a better sense for what PA is and why it's valuable; economic pressure, which has encouraged companies to wring more value out of information and analytics; new delivery schemes, usage metaphors, and interaction paradigms, which combine to make PA easier to build, deploy, embed, and use; and -- of course -- big data.
According to Halper, "slightly more than 34 percent" of survey respondents say they're already using PA, and an additional one-fifth (20 percent) say they're testing or evaluating PA in their environments -- e.g., in the form of a proof of concept (POC) deployment. Collectively, more than half (54 percent) of survey respondents say they're using or actively evaluating PA technologies.
Halper contrasts this figure with that of a 2007 TDWI report, which found that just 21 percent of respondents were using or evaluating PA, while another 19 percent were in the process of developing predictive analytic technologies. "No longer is predictive analytics the realm of statisticians and mathematicians. There is a definite trend toward business analysts and other business users making use of this technology. Marketing and sales are big current users of predictive analytics and market analysts are making use of the technology," she notes.
Predicting Uptake and Success
In that same 2007 TDWI report, Wayne Eckerson -- then director of research with TDWI -- identified several significant barriers to pervasive PA, not least of which was the difficulty of recruiting or hiring business-savvy analysts (the kinds of folks we'd now call data scientists) to oversee the quasi imaginative-artistic, quasi logico-scientific process of building and managing predictive models. Back in 2007, Eckerson thought usability was still a significant problem, too, even though the major vendors (at the time, SAS Institute Inc. and SPSS Inc.) claimed otherwise. Eckerson, as it happens, was right: usability was still a big issue. In other words, and in spite of the most optimistic prognostications of PA's proponents, uptake of PA technologies didn't explode.
Technologies or trends rarely "explode." Gradualism is more the rule, as Halper suggests in TDWI's new report. As of late-2013, slightly more than half (54 percent) of shops in TDWI's survey sample were using or experimenting with predictive analytics. This is a significant increase (about 33 percent) from TDWI's 2007 tally (a combined 40 percent) -- but it works out to about 5 percent compound annual growth rate (CAGR).
Healthy, yes. Explosive? No.
That said, PA has eclipsed the 50 percent mark. A majority of shops are using (or gearing up to use) it. That's significant, Halper argues. "No longer is predictive analytics the realm of statisticians and mathematicians. There is a definite trend toward business analysts and other business users making use of this technology. Marketing and sales are big current users of predictive analytics and market analysts are making use of the technology," she concludes.
You can download a free copy of the TDWI report here. )