We live in an era of big data. Whether you work in financial services, consumer goods, travel and transportation, or industrial products, analytics are becoming a competitive necessity for your organization. But even IT people and those who can manipulate data successfully need partners -- business managers who can partner effectively with the “quants” (quantitative analysts) to ensure that their analytics work yields better strategic and tactical decisions.
For executives fluent in analytics -- such as Gary Loveman of Caesars Entertainment (with a Ph.D. from MIT), Jeff Bezos of Amazon (an electrical engineering and computer science major from Princeton), or Sergey Brin and Larry Page of Google (computer science Ph.D. dropouts from Stanford) -- there’s no problem. But many executives -- even those in IT-- find their math and statistics background may be a bit rusty.
How do you help lead your company into the analytical revolution, or at least become a good foot soldier in it? My recent book (co-authored with Jinho Kim), Keeping Up with the Quants, and a related article in Harvard Business Review -- offer a primer for non-quants. However, CIOs and other leaders can benefit from the pointers offered as well -- especially CIOs who oversee centralized analytics groups.
The ability of quants to communicate and collaborate with executives is directly tied to their effectiveness.
At companies such as Procter & Gamble, “embedded” analysts within IT are one of the most important vehicles for connecting the company’s data and analytics with functional and business-unit decision makers.
Adding Business Context
In addition, for decisions affecting the IT function itself, CIOs can begin to think of themselves as consumers of analytics. They might be consuming data and analytics to decide, for instance, whether a new data center is needed, or whether to renew an outsourcing contract. If the quants are well-trained, they are good at gathering the available data and making predictions about the future. But most lack sufficient business knowledge to identify hypotheses and relevant variables and to know when the ground beneath an organization is shifting. As a data consumer, you need to generate hypotheses and determine whether results and recommendations make sense in a changing business environment. In other words, you can help guide the quants toward practical business applications and outcomes by challenging their assumptions -- gently, of course.
If you remember the content of your college-level statistics course, you may be fine. If not, bone up on the basics of regression analysis, statistical inference and experimental design. You will need to understand the process for making analytical decisions, including when you should step in and ask questions. Remember that every analytical model is built on assumptions that producers ought to explain and defend.
You might also learn directly from the quants in your organization by working closely with them on one or more projects. If they report to you, working with them should be that much easier.
What to Ask Your Numbers People
Many managers fear that asking questions will make them appear unintelligent about quantitative matters. However, if you ask the right kinds of questions, you can both appear knowledgeable and advance the likelihood of a good decision outcome.
Many possible questions for various stages of analysis are listed in our book. Three of the most important you can ask about data are:
1. What are the assumptions behind the model you built? Is there reason to believe those assumptions are no longer valid?
You are really looking only for a thoughtful response here. The only way to know for sure about whether assumptions still hold is to do a different analysis on newly gathered data -- which could be very expensive.
2. How is the data you gathered distributed?
If the person can’t describe the distribution, he or she is a shoddy analyst. Good analysts should have already looked at — and be able to show you — a visual display of the distribution of your data on any particular variable. If you are interested in one variable as a likely predictor of another, ask for a scatterplot and look to see whether the data line up in any linear pattern; that would indicate a strong correlation between the two variables. You could go on to discuss whether the data follow a normal distribution, too, and if there were any significant outliers or unexpected values that don’t fit the pattern.
3. What’s the “provenance” of the data in your analysis?
This question is particularly important to CIOs. It’s important to know where the data came from and how reliable the source is. Few are better positioned to judge answers to that question than the head of the IT function. The data may also be external. As I argue in my forthcoming book, Big Data at Work, big data is much more likely to be sourced externally. Discussing external data with analysts may help CIOs understand which types of data are in demand for decisions, and which need to be brought in-house on an ongoing basis.
The main takeaway is that data analysts and quants are excellent at what they do -- culling data and drawing conclusions. But they often need guidance from senior business managers and IT leaders to make sure the information is thoroughly accurate and that it is actionable and useful for making solid, data-driven decisions for the business.
See the related Q&A here.
Tom Davenport is the President's Distinguished Professor of Information Technology and Management at Babson College. He has led research centers at Accenture, McKinsey and Company, Ernst & Young, and CSC Index, and has taught at Harvard Business School, Dartmouth's Tuck School, the University of Texas, and the University of Chicago. He is a widely published author and speaker on the topics of analytics, information and knowledge management, reengineering, enterprise systems, and electronic business. Tom has written or co-authored fifteen best-selling business books, the latest, Keeping Up with the Quants: Your Guide to Understanding and Using Analytics. continues his pioneering work on Data Analytics begun with the bestseller, Competing on Analytics.