It is important to have good intuition when using data in decisions, and this intuition can be improved. This is a central message of a book that recently emerged from teaching these topics at Columbia University. The message makes a lot of sense and is likely, if anything, to make even more sense as time progresses. We are increasingly able to get data. Furthermore, much analysis may be automated in the future. Despite this, the challenge of knowing how to make a good decision given the data and the analysis presented to us will still remain.
Data, Decimals, And Decisions.
When using data in decisions, consider what is important. Do decimals really matter? Most of the time not, so why show them? Decimals can distract and confuse and give the appearance of false certainty when it does not exist. Decisions are about the future, who can completely accurately predict the future?
The authors argue that we need to do more synthesis, and less summary, of data. Summarizing restates the data. Synthesis adds context, provides a story, and even suggests a solution. Synthesis is much more useful for the person it is shared with than mere summary.
Part of getting better intuition is using the right tools. To help the decision maker get a quick idea of whether something makes any sense at all they discuss Fermi Estimation. You can use this to get a pretty accurate guess at a ballpark number for something you don’t know anything at all about. What is the market size? Not sure, but guessing at the number of people who are in the area, the percentage who might buy, what quantity they might buy, and plausible prices could give you an idea. Getting an idea of what the numbers mean is crucial and need not be especially precise. If your new restaurant needs to attract more customers than exist in your town you can be confident that it won’t be a success without doing much else in the way of analysis.
Using Data In Decisions
This book is not a Gladwellian-inspired tour of research interspersed with cool anecdotes. There are a few stories but they are more about the authors’ experiences and likely aren’t dramatic enough to have made the cut in a Gladwell piece. Instead, the book is more of a how-to guide to using data in decisions. There is a bunch of useful advice. Not least, be judicious in the way data is used. Don’t deluge people with data that isn’t likely to change their decision whatever the data says. Like a car’s dashboard, one should present a workplace dashboard of need-to-know information, not present everything that anyone could possibly know about the topic at hand.
Similarly, don’t bother doing work when you have already decided. Don’t analyze data because it is there. Analyze data because it might change a decision.
Too often we conduct expensive and time-consuming data analysis to evaluate decision that we’ve already made.
Frank, Magnone, and Netzer (2022) page 43
Judgment Not Data
I have to say that my favorite analogy was between asking questions and pulling threads on clothes.
Some loose threads will just come out; others, if pulled, can unravel the whole sweater. Questioning enables you to quickly pull threads to see which are superfluous, which are integral and consequential.
Frank, Magnone, and Netzer (2022) page 10
To be honest I wasn’t really sure if unraveling the sweater was a good or bad thing in the analogy. Still, I think I get the point. Honing your questioning skills allows you to decipher what is critical from the noise.
The book is a helpful addition to managerial decision-making and so I’ll end with a return to one of their key points. Often what is missing is some form of judgment.
[In their examples] …the problem was not in the lack of data or the data itself, but rather in the judgment employed in converting information to sound decision-making.
Frank, Magnone, and Netzer (2022) page XX
For some previous posts on using data in decisions see here, here, here, here, and here.
Read: Christopher J. Frank, Paul F. Magnone, and Oded Netzer (2022) Decisions Over Decimals: Striking the Balance between Intuition and Information, Wiley
