Everyone wants to make data driven decisions, but really, anyone can do that. Data is ubiquitous. It’s in spreadsheets and databases. Even if it’s bad, incomplete or hard to extract, it’s still there, waiting.
Since data is everywhere, what decision makers, strategists and researchers should really be aiming for is using good data interpretation practices to drive decision making.
Using data is like using information about your surroundings to drive a car. To extend the metaphor: when you want to be data driven, you can either make it to your destination safely in a timely manner or you can speed during a snow storm and end up with your car stuck in a ditch.
Here are 3 Don’ts of Data.
#1 Causation is not Correlation
If data is ubiquitous then it can be used to say anything. This is a poor data interpretation practice though. My favorite example of this is the correlation of pirates to global warming.
A data driven solution can be made from this chart to increase the amount of pirates in the world in order to bring down rising global temperatures. To do that though would be to fall into the causation equals correlation trap. As you review data, always remember to make sure you are reviewing whether the data points you are reviewing are correlating or not. For most instances, you can apply additional scrutiny when trying to form a hypothesis, but if you are making a big decision it might be time to statistically prove that the prediction is strong through a model. There are simple ways to do that in Excel or by using some of the models in the book Score! [NOTE: I’ll be exploring this in another post.]
#2 What are you solving for?
When preparing data it is crucial to know what you are trying to solve for. The most common problem I run into is when someone using data for something it is not intended for, like exporting data on constituents and then trying to solve for the amount of gifts coming from a certain region. The unintended consequence of using constituent data to total gifts would be that duplicates will severely inflate the amount of money raised.
When pulling data from your database, make sure you are pulling data that will allow you to solve for what you intend.
#3 Unintended Consequences
With so much talk about data available to analyze, sometimes it can lead to decision making that is divorced from the real world. Making data driven decisions in a vacuum can lead to unintended consequences. Recently, Steve Forbes, commenting in his own magazine, opined on the overwhelming value of e-cigarettes:
The real world shows the opposite: Since the invention of e-cigarettes a few years ago, teenage smoking in the U.S. has fallen by half! Teenage use of vaporizers has kept teens from getting hooked on a truly lethal habit.
Seems good, right? A decrease in teenage smoking! Hooray! But, reading between the lines, it does not take long to realize that what is really happening here is an increase in teenage vaporizer use. Teens are still hooked on something: nicotine. It’s just changed vehicles. The data is there: teenage smoking cut in half. But, if we let the data drive us to the conclusion Forbes desires, we are in a world with tens of thousands of teenagers puffing on vaporizers. Is that a world we want to live in? Maybe, but we should at least think about it before letting data pronounce such a dramatic reduction in smoking without taking the time to explore the unintended consequences.
Before diving into a spreadsheet and coming up with your next great solution, make sure you are not accidentally performing one of the 3 Don’ts of Data.