How non-technicians can master the basics of data analytics to make better decisions
Most of the volumes in the “HBR Guide to” series are anthologies of articles previously published in Harvard Business Review in which various contributors share their insights concerning a major business subject such as Better Business Writing, Getting the Right Work Done, and Project Management.
As is also true of volumes in other such series, notably HBR Essentials, HBR Must Reads, and HBR Management Tips, HBR Guides offer substantial value in cutting-edge thinking from 25-30 sources in a single volume at a price (each at about $15-20 from Amazon in the bound version) for a fraction of what article reprints would cost.
What we have in this volume — now on sale by Amazon US for only $16.86 – is a wealth of information, insights, and counsel that can help leaders in almost any organization to establish or strengthen a workplace culture within which managers understand the numbers and make better decisions based on those numbers.
For example, when it comes to requesting new data or analytics from data scientists, a lot of managers don’t know the right questions to ask to get the information they need. But if you don’t frame the request correctly, you might leave your analyst uncertain about how to proceed. Here are three questions to think about:
o What will you do with the data? Be as specific as possible about what decisions you and the company will make based on the information. The data scientist, by hearing what you hope to achieve, can then collaborate with you to define the right set of questions to answer and better understand exactly what information to seek.
o Is the data readily available? Ask if someone has already collected the relevant data and performed analysis — either in your company or using public data. The ever-growing breadth of public data often provides easily accessible answers to common questions.
o How do we get the data? Data scientists must decide between using data compiled by the company through the normal course of business, such as in observational studies, and collecting new data through experiments, which can be expensive. As part of your conversation with analysts, ask about the costs and benefits of these options.
Data analysis can help us sort through complexity and make decisions, but even with the best analytics tools, we’re still vulnerable to human mistakes. For instance, we’re likely to pay more attention to findings that align with our beliefs and to ignore other facts and patterns in the data. This is called the confirmation trap. You can avoid it by trying to embrace information that counters your (or your boss’s) beliefs by doing the following:
o Specify in advance the data and analytical approaches on which you’ll base your decision. This will reduce the temptation to cherry-pick findings that agree with your prejudices.
o Actively look for findings that disprove your beliefs. Ask yourself, “If my expectations are wrong, what pattern would I likely see in the data?” Enlist a skeptic to help you.
o Treat your findings like predictions, and test them. If you uncover a correlation from which you think your organization can profit, use an experiment (or perhaps more than one) to validate that correlation.
This volume will help managers improve how they identify the most important information needs, complete knowledge transfers, work more effectively with data scientists, conduct business experiments, assess the value of data, make early (albeit tentative) predictions, distinguish between correlation and causation, avoid cognitive biases (e.g. confirmation), and maximize the potential value of machine learning.
I urge those who share my high regard for this book to check out a companion volume in the series, HBR Guide to Finance Basics for Managers.