News Stay informed about the latest enterprise technology news and product updates.

Tom Davenport on using analytics to influence business decision-makers

Analytics guru Tom Davenport gives a peek into his latest book, 'Keeping Up with the Quants: Your Guide to Understanding and Using Analytics.'

In his forthcoming book, Keeping Up with the Quants: Your Guide to Understanding and Using Analytics, visiting Harvard Business School professor Tom Davenport offers managers a primer for surviving what he calls "the quantitative information age." For most CIOs, keeping up with the "quants" comes with the job. But as Davenport and co-author Jinho Kim drive home, the numerical data is only a means to the ultimate end: to influence business decisions.

Tom DavenportTom Davenport

In this two-part SearchCIO New Books interview, Davenport talks about the goals of big data analysis, how best to communicate analytical results and why business intelligence (BI) should go back to being called "decision support."

The author of numerous books, including several firsts in the areas of big data and analytics, Davenport also serves as the President's Distinguished Professor of Information Technology and Management at Babson College in Wellesley, Mass. He's also co-founder and research director at the International Institute for Analytics, and a senior adviser to Deloitte Analytics.

Look for a chapter excerpt from his new book on later this spring.

Tell me why your book would be useful for CIOs.

Tom Davenport: There was a McKinsey report on big data that got a lot of press about a year-and-a-half ago, and it argued that we need more than 1.5 million data-savvy managers who are going to be the effective consumers of what CIOs are pumping out. IT by itself is not going to be successful unless there is a group of effective consumers of all the data analysis they are providing.

So, this book is intended to create a more savvy manager. It's about how to think analytically, how to use all the tools at your disposal -- visual and otherwise -- to make effective decisions on the basis of analytics. We try to give some business examples and some real-life examples on how to use analytics: for example, using analytics to improve your marriage and how to measure quite subjective things, like -- well I don't want to go into details about this -- but we provide a measure for male erections to show that almost anything can be measured taking a scientific approach.

I'm sure that alone will sell some books.

Davenport: To college students maybe.

One of the topics you examine is the importance of telling a story with data -- in terms the audience can understand.

Davenport: Over the last several years, as I talked to good analysts, a lot of people would say, 'You know, you need to be able to tell a story with data.' And I'd say, 'Well, what do you mean by that?' I'd never really thought about it systematically. So, in this book we try to lay out what are the different types of stories that you can tell with data.

SearchCIO New Books

Keeping Up with the Quants: Your Guide to Understanding and Using Analytics

Authors: Thomas H. Davenport and Jinho Kim

Publisher: Harvard Business Review Press

Publication date: June 11, 2013

Some of this I got from a guy who is now at American Eagle Outfitters, Joe Megibow. He was at Expedia at the time. Expedia was very much into what he called 'the CSI story.' This was a way to find out what was wrong, or find a problem with their data: One example was discovering that people were dropping off and not completing the transaction because information was being requested they didn't have. For example, in Ireland, outside of Dublin, apparently postal codes don't exist, so people were dropping off the site because they didn't know what to put in that area.

There is the Eureka story, where you know something is out there and you're trying to find it. There are the basic 'what happened' stories, which all reporters know how to tell: the what, when, who, et cetera. There are stories about experiments: Increasingly with analytics we have a control group and a test group and people are randomly assigned to the two and you see if one group is statistically important -- those are becoming important with analytics. There are more predictive stories, survey stories. I tried to lay out what are the different types of stories that you can tell with data. And then you have to think about what's the best medium for it. Is it narrative with words? Is it numbers? Is it visual display? Is it video?

In the book, we talk about a group at the InterContinental Hotels Group who is charged with telling basic financial reporting stories, but to liven it up and get the attention of executives, they make videos that embed the information in them.

It's a completely different world now. We have so many choices of how to convey information, so we have to think about what's the type of story and then the medium -- or mixture of media -- that best conveys that information.

You've talked about another interesting example of an analytics person who made a 3-D model of data results for the business.

Davenport: That was Vince Barraba. I don't really know whether it was more successful, but you have to admire it so much because he's thinking about how these executives understand information and make decisions. That is really the key: to think about who's the audience, what's the context for the decision and how are they going to receive this information. Jeff Chasney, the CIO at CKE Restaurants -- they do Hardees and Carl's Jr. -- has studied cognitive science principles in order to learn about the ways people best absorb information. And one might argue that cognitive science could well become an essential skill or knowledge set for CIOs, because it is not about pumping out the information but whether people are using it effectively.

That is really the key: to think about who's the audience, what's the context for the decision and how are they going to receive this information.

Tom Davenport,
professor, Harvard Business School

You drive home the message that data isn't the point, numbers aren't the point -- the data has to express some idea.

Davenport: Exactly, and the idea might be conveyed in multiple ways. At some point, we'll try to have more experiential ways of understanding data. Right now, you can look at data from a call center and what customers are complaining about, but it is obviously much more effective to require executives to spend some time on the other end of the line in a call center -- and some companies do that. That's much more memorable than seeing statistics. Alfred Sloan used to do this at General Motors. On his vacations, he would go out and visit dealerships and spend time with customers of GM to understand their concerns. So, it is not only about conveying the information, but is the information conveyed in a memorable way so the person will understand, remember and act on it.

You wrote that firms increasingly embed analytics in the processes of the company. Can you talk about this?

Davenport: There are multiple ways of addressing that issue. One is you could create a new product or service, based on that analytic, which is quite common in the big data world of online companies. LinkedIn or Facebook, Google have been doing this for a long time -- they are constantly creating new products and services that are based on big data. Once you create that, it is no longer an add-on; it's part of the basic offering. Another way is to just embed it into the system and the process for running part of your organization.

How can CIOs help their analytics and BI people focus more on using data effectively?

Davenport: I think communicating the thought that it is not their job to present information, it's their job to influence decisions. At Merck, the commercial analytics group has a very good reputation. People will come to them and say, 'I hear you do promotion analysis.' And the analytics people will say, 'Yeah, we do.' 'So do one for me.' The analytics people will say, 'Well, we'll come back with one of three answers: the promotion is incredible, marginally effective or ineffective. Give me a sense of what you're going to do in each of those cases.' You can say it's a little cheeky for an analytics person to ask a decision maker that kind of question, but it basically demonstrates that this is about making better decisions, and if you're not going to make a decision on the basis of this work, we shouldn't waste each other's time.

So, putting everything in a decision context is important. You know, this whole area of BI and analytics was initially known as 'decision support.' And unfortunately, I think we've gotten away from that and often it is not very clear what if any decision is being made on the basis of the data at all. One of the reasons I admire the group at Procter & Gamble who renamed their IT organization 'Information and Decisions Solutions' is that they are clearly working very closely with their senior executives on decision making.

In part two, Davenport digs into the various means to this end, and explains how visualization tools are critical but still quite underdeveloped.

Let us know what you think about the story; email Linda Tucci, executive editor.

Dig Deeper on Enterprise business intelligence software and big data

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.