Data are just a proxy for reality, they aren’t reality.  As more and more information is being captured, analyzed, and published, this is an important consideration for leaders.  While “management by walking around” might have been the mantra for the 80s, perhaps “data gathering by walking around” should be the theme for today.

In a prior post, I warned leaders about only trusting data that was meticulously measured.  Those people tend to miss out on a lot of what is happening around them.In that post I shared economist Paul Krugman’s story about the development of maps of Africa between the 15th and 19th century [1].  Krugman explained that 15th century maps, while not always accurate on distances and specific locations, were quite accurate on what actually existed across the African continent.  Yet, by the 18th century while Africa’s coastline had become meticulously and accurately represented, most of the interior of the continent had literally disappeared from the map.  The problem was that map makers of the day wouldn’t include things on their maps that hadn’t met their standards for data collection and documentation.  Since the interior hadn’t been explored as extensively as the coastline, the data weren’t as robust.  While what was shown on the map was extremely accurate, the map as a whole no longer adequately represented Africa.

Of course, in the 15th and 18th centuries, people didn’t have access to the same technology and data as we do today. We now have much more data and information at our fingertips. Certainly we should have a better sense of what the world looks like. Maybe not.

In their book Freakonomics [2], Steven Levitt and Stephen Dubner provide an amusing story about how, in 1987, millions of children mysteriously disappeared from the US population.  No, it wasn’t a massive alien abduction.  That was the year that the IRS started requiring people to list their dependents’ social security numbers on their tax forms.  People could no longer cheat when claiming dependents.   As a result, the IRS data reflected about 10% less children in 1988 even though the number of children in the US hadn’t really changed (other than natural changes like births, deaths, and turning eighteen).

In 2014 we have even more data than in 1988.  Surely these problems can’t still persist. Yet, they do.  Yesterday, my brother submitted a correction to Google Maps. He is the owner of a new restaurant in Montauk.  However, when displaying the location of the restaurant, Maps still shows the name, reviews, and other pertinent details about the old restaurant.  Within a few hours of submitting his correction, he received the following response:

“Upon reviewing your suggested change, we have decided not to apply your suggested change to XXXX at this time, as we found the existing details to be more appropriate”

How did Google make this determination?  Certainly within the short “review” timeframe, they didn’t actually go take a look. Had they, they would have discovered that the restaurant was not there.  Yet, it’s still there in the virtual world which is apparently where Google looked to find it. Just like the map makers of the 18th century, they are letting data define reality rather than the other way around.

Ironically enough their decision doesn’t really help anyone better understand reality.  The original restaurant does still exist, it just moved.  So now people can’t find the new restaurant that is currently in that location or old restaurant in its new location.

Big data has many benefits for organizations and individuals.  However, it becomes dangerous when data about reality trumps what is real.

Have you ever been turned down for credit because of a problem in your credit report that didn’t exist?  How long and how much effort did it take for you to correct that?  It seems that it’s much easier to be misrepresented in a credit report than it is to have a misrepresentation corrected.  I once had to prove that I didn’t default on a loan that I never had.  Shouldn’t the burden of proof been on the people who claimed I had the loan in the first place?  My wife was once turned down for insurance due to a rare medical condition that she didn’t have.  We had to go through a three month review process and submit all of her medical records to prove that she didn’t have the condition.

The ubiquity of data has created a new problem.   It used to be that you had to prove that the data matched reality.   Now it seems that the burden has shifted to having to prove that the reality is real when it does not match the data.

Data are not reality. They are an approximation of reality.  Your job is to understand what is actually happening around you.  As more of our world and lives become captured in data, it will be even more important to be diligent, thoughtful, and critical of the data you see.

———

Brad Kolar is an executive consultant, speaker, and author.  He can be reached at brad.kolar@availadvisors.com.