Research

Value-Added Data Visualization

By Elliot Bendoly, Jan 12, 2016

Let’s step away from the hype for a second.

It can’t be denied that humans are, broadly speaking, intensely visual creatures. Of our senses, vision often dominates our perceptions of the world around us. It is one of the strongest physiological channels that engages us, and it can play a fundamental role in our attentiveness.

It can also be distraction.

There’s a reason the most popular phones today are those that have cutting edge video streaming capabilities; why users are just as likely to browse for visual content with their phone as actually “call” someone; and why according to a recent Pew Research Center survey nearly half of users claim not to be able to live without these devices. There’s a reason graphical massive multiplayer role playing has taken off the way it has, relative to say more historically established text-based multiplayer computer games. And there’s a reason why effective gamers can actually make money (even careers) by doing little more than recording and posting.

Compelling forms that drive visual processing, especially forms that allow for our interaction and higher levels of cognitive processing, draw us in. Or more appropriately, there is a strong match between what we are physiologically capable of visually processing and the nature of many of the visual stimuli being developed today. Our visual and cognitive systems have a thirst – strong visual renderings quench and reinforce that thirst.

For better or worse.

It’s all relative, really. Human visual perception can be capable of permitting remarkable/disconcerting levels of focus under certain conditions (as various selective attention ‘Gorilla in the Room’ tests demonstrate). On the other hand, under alternative conditions, it can pick up on the finest distinctions in compared objects. Not always relevant distinctions. Again, blessings and curses.

The short of it is this: The potential value available to us, and to our organizations, through visual examination is a critical function of the design of what we’re examining.

This is why the most effective approaches to data visualization, those which generate value to organizations, are ones that have been carefully thought out; incorporating critical aspects of the analytical context as well as the attributes of the individuals interacting in visual design and interpretation. This is what separates value-added data visualization from distractions.

Now I could go down the road of discussing why many of today’s highly sophisticated, highly priced, data visualization ‘solutions’ are falling short of their promises… but let’s save that for the next OpEd. Instead, let’s talk now about where the real value-added in data visualization might be.

What value does effective data visualization offer us that we can’t easily get by some other means?

We have to approach this question the right way by considering which kinds of decisions might be better informed by data visualization. We’ll need to come up with a system to estimate the contribution of these decision improvements to our fundamental objectives. That might sound fairly daunting, especially when focusing specifically on data visualization contributions distinct from other related analytical processes. But there are at least some frameworks that can help us.

First it’s useful to delineate a few areas where value from any decision support tool (including but not limited to data vis tools) might exist. Through one cut, we can delineate between benefits to strategic decision making and benefits realized tactically and operationally (in support of higher level strategies). In another, we can delineate between benefits that purposely reinforce competitive strengths from those that critically dampen strategic weaknesses. We can delineate between those that are largely exploratory with regards to searching out opportunities and those that expose threats.

Sound a bit like SWOT analysis? It should. At the core of every consultation I perform on rationalizing data visualization investment is an adapted form of SWOT, with notable distinctions between confirmatory and exploratory use, as well as operational and strategic application (Figure 1). Strong arguments for value added to any of these areas helps justify the nature of decision support expenditures of any type.

Image

Figure 1. Characteristic Elements in Data Visualization Investment Planning

Focusing on specific domains of application such as these, we can more easily evaluate the potentially unique contributions of data visualization. From a confirmatory standpoint for example, data visualization offers a means to fill in the gaps between otherwise black-boxed statistical and computational analysis and the reality of prescription applicability. One of the greatest barriers to the application of analytical findings is the doubt, often well placed, with which decision makers regard assumptions made by analysts. Even in cases where assumptions made appear entirely appropriate, the task of conveying the nature of those assumptions in a convincing manner can be challenging. Layman text descriptions are typically superior to algebraic forms of numerical assumptions alone, but these can also fall short of convincing. Graphical depictions of the tendencies for certain variables to follow trends, correlate with others, vary within system limits… Data visualization can describe the nature of decision making terrains in the kind of detail that can strengthen model application.

From a complementary exploratory standpoint, data visualization can also debunk faulty prior assumptions. Are we sure that the relationship between an operational decision and a performance outcome is best captured as linear? Are there no ceilings to the effect? Is the effect truly continuous rather than punctuated in some instances? Can we not expect to see feedback loops perpetuate the variance we initiate in the decisions we might make? Visual examinations of data in the search for assumed, often simplified, relationships often yield more than what we anticipate.

In other words, data visualization can through attempts to shore up evidence in support of statistical and computational assumptions in fact identify and rectify flawed ones. Again to the benefit of application. Simultaneously such attempts at confirmation and associated exploration often provided the foundations of opportunity and threat identification. Confirmatory data visualization and exploratory analysis are all part of the same system, just as strategic decision making and operational support are.

There are also many audiences for the artifacts that data visualization efforts yield. Some are higher level decision makers. Some are other analysts. Some are workers on the floor. Some are our customers / co-producers. Data visualization can be thoughtfully leveraged in the development of a corporate culture of inquiry and critical introspection. Some of the most effective applications of data visualization have used the most minimally technical solutions in tandem with cultures of practice that reward individuals at all levels for findings that they encounter and clearly, convincingly convey.

It doesn’t always take much investment to make data visualization a value-adding proposition, but it always takes a well thought out plan, some dedication and a willingness to have prior beliefs questioned. Organizations willing to approach data visualization as more than a “just in case” tactic, have the best shot of getting a real strategic return on of their efforts.

About the author

Author photo

Dr. Elliot Bendoly is an IIA Faculty Member and a full professor in the Management Sciences department of the Fisher College of Business, at the Ohio State University. In 2015 he was named the Operations Management Distinguished Scholar by the OM Division of the Academy of Management. Before joining Fisher, Dr. Bendoly was the Caldwell Research Fellow and Associate Professor in Information Systems and Operations Management at the Goizueta Business School of Emory University. His pre-academy industry experience includes work as research engineer for the Intel Corporation. He holds a PhD from Indiana University in the fields of Operations Management and Decision Sciences, with an Information Systems specialization in ERP and Knowledge Management. More recently he has been involved with coursework on modern analytics and visualization, IT-supported service operations and DSS development for managers.

Dr. Bendoly is a prolific author and serves as Senior Editor at the Production and Operations Management journal and as Associate Editor for the Journal of Operations Management. His current research interests include studies into the effectiveness of Operations/IT alignment, and investigations in the Behavioral Operations domain: Collaboration/group dynamics; and Work policies/task complexity/uncertainty.

Dr. Bendoly is the author of Excel Basics to Black Belt (Cambridge Press 2013, 2008), and his LinkedIn discussion forums, “Excel Basics to Blackbelt” and “Operations Management in Practice”, boast 25,000+ and 18,000+ members respectively. He is also the co-editor of Strategic ERP Extension and Use (Stanford Press 2005) and the Handbook of Research in Enterprise Systems (Sage 2010). His most recent textbook project is the Handbook of Behavioral Operations Management (Oxford 2015), a text on learning/training activities highlighting behavioral phenomena in operations management contexts.


Tags

Advanced Analytics & Big Data Adoption

Research Report

The role of analytics and Big Data in enterprise operations, business processes, and decision-making is expanding at an impressive pace.

Read Our Research »