MIT and IBM on Analytics
When I think of advocates of mathematics and statistics for business, I cannot come up with two better exemplars than the MIT Sloan School of Management and IBM.
MIT is perhaps the pre-eminent engineering and technology school in the world, and the Sloan School is recognized as the top business program for quantitative analysis. And IBM, with its recent BI and analytics software acquisitions, its growing quantitative professional services for business, and its ubiquitous Smarter Planet messaging, has built a bet-the-business corporate strategy around analytics and optimization.
The Sloan Management Review and the IBM Institute for Business Value recently collaborated on a survey to understand the challenges and opportunities with the use of business analytics. Over 3000 executives, managers and analysts representing 108 countries and 30 industries completed the 22 item questionnaire. Primary sources were MIT alumni, SMR subscribers, IBM customers and invited academics and industry experts. Survey findings were published in the SMR Fall 2010 research report, "Analytics: The New Path to Value."
Survey questions cover basic information such as industry, respondent role, geography and revenue. In addition, specific questions probe business challenges, uses of analytics, perceived value from analytics, and obstacles to analytics deployment. Finally, there are a number of Likert-like scales that assess the organization's overall performance, as well as its sophistication with information-consumption and analytics. Much of the report focuses on correlating variables from each of the three question groups.
Not surprisingly, organizations discriminate on usage of analytics by their perceived performance. The prevalence of analytics is much higher with top-performing organizations than it is with lower performers, a finding that is consistent across the many functional areas like finance, sales and marketing, and customer service. Fifty-three percent of top performers use analytics to guide day-to-day operations in contrast to just 27 percent of laggards.
At the same time, respondent companies report stark differences in operations surrounding their sophistication as analytics competitors, with categories Aspirational, Experienced and Transformed derived from the question response scale. The primary analytics business challenges for Aspirational companies is cost efficiency, compared to revenue growth for Experienced and Transformed. The motives for each group are to justify actions, to guide actions and to prescribe actions, respectively. Aspirational cultures don't encourage the sharing of information; Experienced cultures share information and insight on a limited basis; Transformed cultures, on the other hand, are effective sharing both information and insights.
One interesting series of items contrasts the top analytic methods in organizations today with those anticipated in 24 months. Forecasting and standard reporting are the most common current techniques, while visualization, simulation, predictive models and mathematical optimization are seen as leaders two years out. Those who've read my last blog will note the similarity of the latter methods with much of computational statistics that is now making its way to the business world.
The study introduces a methodology for operationalizing analytics with mnemonic PADIE, comprised of distinct feedback loop steps:
- Document processes and applications,
- Identify data and insights and
- Embed analytic insights.
An important recommendation of this process? Start with questions, not data. A science of business approach that first generates hypotheses, then looks to data and analytics for test/refine works best.
There's much more to the report for interested readers, such as individual question response histograms and IBM case studies, than I've covered in my brief remarks. Overall, I'd rate this survey very highly, though the bar is set pretty low by the research of many BI/analytics vendors. I would have liked to see a more detailed description of the survey methodology, but the N greater than 3000 is reassuring in the face of a very distinguished-seeming group of respondents. And while both SMR and IBM certainly benefit from a strong showing on analytics in business, there are no direct product sales angles to this research, unlike with many BI/analytics “surveys.” A rising tide of analytics in business benefits an entire industry, not just SMR and IBM.
Those who follow Tom Davenport and Jeanne Harris will find corroborative reassurance with this study. Like research addressed here, the 10-year work of Davenport and Harris has been beneficiary of a large N, with significant senior management participation. Davenport and Harris counter the MIT/IBM analytics sophistication segments of Aspirational, Experienced and Transformed with the more granular Analytically Impaired, Localized Analytics, Analytical Aspirations, Analytical Companies and Analytical Competitors.
And Davenport and Harris take the MIT/IBM self-reported relationship between analytics prowess and company performance a step further, using market returns to show that organizations which invest heavily in analytics outperformed the S&P 500 from 2002-2009. In addition, the authors argue that companies with strong analytical skills recover more quickly from economic downturns. Their Analytical DELTA – Data, Enterprise, Leadership, Targets and Analysts – complements MIT/IBM's PADIE methodology quite nicely.