A Look into the Future of Analytics
Industry luminaries Claudia Imhoff and Harriet Fryman peer into the future of analytics.
- By Stephen Swoyer
- September 16, 2014
At the recent Pacific Northwest BI Summit, Claudia Imhoff of Intelligent Solutions Inc. and Harriet Fryman of IBM Corp. spoke about the future of analytics.
More specifically, they wanted to talk about the future of analytics today. The future isn't quite here now, but it could be, they suggested.
For at least two decades, the emphasis among business intelligence (BI) practitioners, implementers and integrators, consultants, vendors, and others has been on making business more "BI-driven." As a differentiating strategy, however, "BI-driven" is passé; to be BI-driven is to miss the mark, to come up short. It's long past time, Imhoff argued, to make business more analytics-driven.
"We need to be able to expand who is using analytics throughout the organization," she said, calling attention to the disconnect between the BI industry's obsession with tech, features, and functions and how people actually use and interact with software and technology.
"The industry focus has been predominantly on the usability of technologies that produce these analytics rather than on how people understand them, how people consume them," Imhoff continued, stressing that analytic technologies need to be easy to find, easy to search for, easy to navigate, and easy to use.
Imhoff, a protégé of Bill Inmon and one of the most respected thinkers in information management, was paired with Harriet Fryman, a BI veteran who's regarded by many who know her as one of the sharpest and most imaginative observers in the business. Neither Imhoff nor Fryman could be described as reflexively or unduly optimistic; for a number of reasons – not least among them social and economic factors – both women believe that the shift to analytics-driven business operations is unlikely to result in a reprise of the dysfunction of the past.
Fryman, for her part, pointed to an ongoing cultural shift that's transforming how people expect to consume and act on information: namely, at almost all times and in almost all conceivable contexts.
The effect of this (mostly) consumer-driven transformation has been to address the very problem -- the traditional myopia of BI and enterprise software vendors -- that Imhoff bemoaned. In a mobile context, especially, app developers are trying to tailor the in-app user experience (UX) to the preferences and desires of consumers. They have every incentive to do so.
The advent of pervasive connectivity -- from mobile phones to automobiles to microwave ovens to climate control systems, to say nothing of industrial-strength connectivity in the form of sensors and other connected devices -- is another key piece of the puzzle. Analytics aren't just more usable (from a consumer's perspective), but more valuable, too: richer, more detailed, more complete.
Fryman highlighted the ubiquity of potential analytic sources -- e.g., sensors, devices, applications, and other connected signalers -- and cited the "connected car" as a practical demonstration of how analytics at large scale can actually work.
A third piece of the puzzle, one that has to do with what it actually means for a business to become "analytics-driven," involves leveraging automation to orchestrate or accelerate business processes -- even if this means automating the kinds of non-critical (and, in some cases, critical) decisions that would otherwise have to be made by human beings.
"If you look at what people are trying to do with decision-making, they're trying to make an outcome happen based on data and the conclusions from the data," Fryman said. "Automation is trying to use [an amount] of information to make a better decision; it's just that another object is making a decision, not a human being."
The rub, as Imhoff noted, is that the great mass of potential adopters is still very much stuck in the past. To put it another way, even if the data is available, in all of its richness and depth, will IT organizations actually make use of it? She cited her own experience as a frequent presenter and educator at BI industry conferences and events.
"I always ask [attendees], 'How many of you are using your data warehouses for something other than reporting or multi-dimensional analysis?" she related, noting that out of several hundred attendees, she'd typically get just a few positive responses. "We're still woefully behind the curve in terms of ... getting organizations to figure out that they can do more with all of this beautiful data than just create some spreadsheet or this report."
As a positive development, Imhoff cited an uptick in successful adoption of predictive analytics. "We're starting to see some good stuff come out of the predictive side of the house," she observed, citing cases in which predictive technologies are used as prescriptive aids in decision-making or leverage to help optimize business processes. "It can lead to guided decision-making, but I think that's kind of the ultimate goal of BI and analytics is to be able to say, 'Something is off. What are we going to do about it? Here's the best solution, the best outcome, [here are] our what-if scenarios."
Imhoff stressed that the importance of traditional BI -- OLAP and reporting -- is undiminished. There's the potential, however, to do so much more, she urged.
"OLAP and reporting are still a mandatory part. It is, after all, how people get funded to begin with, [and] it's usually the most understandable project," Imhoff said. "[However,] it's all descriptive: here's what has happened or is happening right now. [This isn't] as valuable in my humble opinion."
Fryman picked up on this theme. Traditional assets -- such as dashboards -- need not be limited to backward-looking or state-of-the-business-type views, she said. "A lot of times, people hard-code the word 'reporting' to [mean] 'historical data,' whereas dashboards can contain [advanced analytics]."
The final piece, Imhoff suggested, is nothing less than a crash course in what it means to think and act analytically. "I don't think the business has been educated on analytically thinking: what does it mean to think analytically, what does it mean to be a diagnostician, what ... value does it have in your everyday life, how do you use analytics, how do you create them? That's education," she said.
"We can train people to use a tool. Unfortunately, that's what most business intelligence and analytic environments do. We're still in this mode of just presenting the analytics and keeping it separate as opposed to embedding it in something." That, Imhoff stressed, must change.