Data analytics can provide enormous benefit to a business that uses it properly, but it can’t do everything and should not be seen as the magic bullet for every business challenge. Although it can inform and influence human insight and judgment, it can’t replace those qualities in the operation of a successful company.
It’s easy to get caught up in the technology, dazzled by the staggering amounts of information that can be gathered, but fail to recognize that the successful interpretation and use of all that data often requires more than pure technical knowledge. This powerful tool is still just that—a tool, albeit one with unusually high potential value.
The reason is simple enough. Data analytics is much better at answering the question of what than the question of why. It can tell a company, often in overwhelming detail, what has happened and even what may be likely to happen going forward. But the question of why something happened is less easily defined solely by data. Human insight, perhaps drawn from the business experience of company executives or the in-field realities of employees, becomes a critical factor in sound decision-making.
“When you combine data analytics with a deeper understanding of a customer’s motivation and experience, that’s how you will create better products and services,” says David Schonthal, who teaches innovation and entrepreneurship at Northwestern University’s Kellogg School of Business. Adding human insight to the analytical mix can lead to more profitable decisions by allowing data to “act as a source of inspiration, not just a source of validation,” Schonthal says.
Although the amount of data that can be gathered and examined is huge, more data and more intense focus on that data is not always the best course for a business. Schonthal cites the Netflix Prize, a $1 million offering the company made in hopes of improving the results of its movie recommendation algorithm by 10 percent. After several years of work by various research teams, that objective was reached and the prize awarded—but the algorithm was so complex the company never implemented it.
The prize-driven search for the algorithm had used 100 million anonymous movie ratings provided by Netflix customers. Nuances—customers who liked action films with lots of explosions, but didn’t like a lot of bloodshed, for example—proved hard to identify.
Netflix added user profiles to its customer accounts, which produced far greater improvement in the accuracy of movie recommendations.
After streaming video replaced mailed DVDs for customers, the company gained the ability to gather much more—and much more useful—information, such as what time of day customers watch a movie; at what point they pause, rewind or fast-forward; what device they use to view the movie; and what, if anything, they watch immediately afterward.
Companies make the best use of analytics when it is driven by their business needs, not by the technology itself, as entrancing as that technology may be. In a McKinsey & Company white paper, Shilpa Aggarwal and Nimal Manuel note that analytics should not be “trawl fishing,” but “spear fishing” instead: “To get the greatest value from a stockpile of data, a targeted approach based on clear business cases generates more value than simply throwing out a wide net and hoping something valuable is found among the catch.”
The risk lies in focusing on the technology rather than the strategic justification for it. More may not be better. Simply increasing spending on technology and personnel may provide a poor return on that investment if it is not accompanied by “a clear idea of the business challenges that must be solved or opportunities that can be captured,” they write.
Tools such as dashboards, which provide near-instant data arrayed in easily readable formats, can be valuable for managers, but an over-reliance on them or a failure to place their data in the relevant context can lead to flawed decision-making. Vikas Mittal, a marketing professor at Rice University’s Jones Graduate School of Business, notes that dashboards now “capture an increasingly small slice of time, space and experience.” What not long ago might have been a quarterly examination of information can now be done hourly or even in smaller blocks of minutes.
As impressive as that is from a technological standpoint, there are some potential drawbacks. Information overload can leave managers awash in more detail than they can make intelligent use of in developing decisions. Managers should ask themselves how often they need the data and at what level of granularity.
A potentially greater problem looms because those smaller slices of information are more easily influenced by outliers, by blips that can skew situations and lead to quick reactions prompted by a volatility that may not reflect reality. Daily or hourly data, for example, can be shaped by the actions of a relative handful of customers and create a false picture of a spike or a drop in activity.
Here again, human insights come into play. Managers with a solid understanding of their business, its historical trends, its position in the marketplace and other similar factors are equipped to avoid poorly grounded reactions that might seem at first glance to be supported by data.
The world of politics provides an example of analytics at the expense of insight. In the 2016 presidential election, news coverage often focused on the minutiae of day-to-day polling, expanding small shifts in subgroups of voters and other such slight changes into perceived trends for the eventual outcome of the election. Accordingly, there was a lot more analytics than insight—and the outcome surprised millions of Americans, including the two candidates.
Mittal argues that insights prompted by discussions of analytics are what a company really needs: “Dashboards replete with numbers will not provide insights. To gain insights, embed the dashboard metrics within forums that facilitate discussions, generate ideas and motivate employees to more systematically use the ‘what’ to move into the realm of ‘why’ and ‘how.’”
And don’t think that colorful visualization of data, although sometimes useful, is any substitute for genuine understanding and insight. Mike West of dummies.com warns that people may expect to find the answer to a business question in the simple representations of a bar chart, line graph, or pie chart:
“However, visual patterns can mislead you. Just because a line seems to increase over time doesn’t mean that your conclusions about why it is increasing are the actual causes. Just because two bars on a graph are different sizes doesn’t mean that the difference is significant or meaningful. Only very large differences among very simple comparisons present themselves obviously in visualizations.
“Overreliance on visualization leads to simplistic observations that are not up to the task of producing answers to complex questions. The real world is complex: many factors push and pull in different directions at the same time. These don’t translate readily to visualization.”
Recognizing that push and pull—and the subtle but significant complexities created—is an often overlooked aspect of making the best use of business data. For all its undeniable value, analytics alone can’t tell an executive everything.
