Participant Questions From the Recent Data and Analytics Webinar: Round 2

We answer another set of questions from our March 15, 2017, webinar on “Analytics as a Source of Business Innovation.”

On March 15, 2017, we held a free, live webinar to share the findings and insights from our latest MIT Sloan Management Review Data and Analytics Initiative research report, “Analytics as a Source of Business Innovation.” The report summarizes our findings about the increased ability to innovate with analytics and how it is producing a surge of benefits across industries.

If you missed the webinar, the recorded version is available on our website. Thanks to everyone who participated in the webinar — we had a great turnout.

During the webinar, many, many participants asked questions. Unfortunately, we had time to answer only a few during the webinar itself. Last month we answered some remaining questions, and this month we’ll cover a few more. We’ve paraphrased some of the questions to provide context, combine similar questions, and anonymized them.

I read recently that everyone needs to be educated in analytics, starting with K-12 education. Does everyone actually need to know more about analytics, and will this happen in the near future?

We certainly see rising importance of analytics skills in our research — our 2015 report, “The Talent Dividend,” focuses on that topic. As businesses increasingly rely on data to create value, the use of analytics will become more pervasive in organizations. At a minimum, increasing education in analytics will help managers better understand potential biases and discern strong evidence from weak.

But we’ve also found that it isn’t easy. It is hard to keep up with the pace of analytics. In our article “Minding the Analytics Gap,” we describe how the rate of increasing sophistication of analytical results that companies produce exceeds the rate at which organizations are able to consume those results — in other words, it’s easier to develop more complicated analyses than it is for everyone in the organization to understand those complexities. This indicates that increasing education in analytics is important, but it’s also a difficult process.

How do we deal with sparse data in innovation?

Of course, many areas currently have massive volumes of data available. For example, sensors can sample data many, many times per second. Or internet browsing activity can quickly generate detailed logs. Or smartphones — even cameras in mannequins’ eyes — can record shopper behavior. With stories like these, there may be the tendency to think that rich data is available for everything.

However, some areas just fundamentally don’t have many observations to record. All modern analysis techniques that I can think of require many more observations than variables, but as I’ve discussed before, in some contexts, the number of observations is woefully inadequate given the number of variables. In many cases, innovation fits that description — we just don’t get to observe many similar innovations in order to fuel analytical approaches.

But that doesn’t mean analytics is entirely useless in these contexts. Innovation contests, for example, do generate a fair number of observations. While data may not allow for prescriptive estimates at the level of detail and precision that can be found in website optimization, they can still inform. Blended approaches, which we described in “Beyond the Hype: The Hard Work Behind Analytics Success,” can work well when data is insufficient. In fact, the sparseness of data in some contexts can help avoid the myopia that can come from excessive refinement that misses potential for larger gains.

What type of new human tasks do you see rising with the expansion of analytics?

In our analysis of the effects of automation, organizations differ not only in how many tasks are automated, but also how many new human tasks emerge from automating more tasks and processes. This question is crucial for the future of human work. Automation can have vastly differing effects on society depending on whether there is a net positive or net negative effect on human tasks.

Our analysis finds that Analytical Innovators — companies that have made the most progress incorporating analytics into their business processes — are much more likely to report an increase in new human tasks than Analytically Challenged organizations (those making the least progress). The Bridgestone example from our report provides a great example: When Bridgestone increased the automation of inventory allocation, there was a liberating effect. The people who formerly spent time on detailed forecasting and allocation could now spend their time on other activities. Bridgestone reports that these new activities made better use of their employees’ knowledge and creativity to think about new ways of creating business value. For example, jobs previously dominated by prediction tasks can now emphasize more use of judgment.