Better decision making with data

Each and every one of us has to make decisions. This is part of life. The problem with making decisions is the fact it is not always easy. Sure, deciding whether I want to eat cereal or bread for breakfast is not that a hard decision. However in life we meet bigger decisions to make such as who to date or marry, where to live and what careers we would like to have. But most decisions in life are somewhere between those two ends of the scale.

Decision scale: harder decisions on the right

Decision scale: harder decisions on the right

As employees we make decisions all day long, and as your job in a company hierarchy is higher the more decisions you make without approval, and they effect more people.

No Ladder

Let’s clear a myth here. I will use engineering here as an example, but it is not limited to engineering. People think an engineer has a “ladder” to climb in order to make more decisions, and get into management. This is a false belief in my opinion and Nathan Potter wrote it much better than I can at this post. Engineers have much more merit when it comes to taking engineering decisions. Only fool managers will guide engineers what technology to use without establishing the facts, pros and cons for a technology. Those who will build and use the product should decide this, they should inform their manager about the decision, and the reasons led to it, but they should be engineering ones, and not top-down instructions, or gut feelings.

On the other hand, managers tend to have more merit when it comes to staffing, company goals and direction and similar areas of expertise. This is no ladder to climb, it is two distinct jobs and people might, or might not be good at both. Taking an excellent engineer and pushing or forcing him or her into a managerial position can be the most foolish thing to do. You might loose a great engineer and get a poor manager. Of course there are other outcomes, but at the core, it is the choice of the individual. Now that the ladder myth is gone, we can talk about good decision making.

Decisions have a toll

Good decision making vary greatly throughout the day. Did you know there is something called decision fatigue? Or that judges grant much less paroles when they are hungry? Since we want to maximize our ability to make quality decisions we need a method to help us pick the right decision. Of course making decisions when we are happy, satisfied and not hungry can help, but this is not possible at every decision we need to make.

We have many other biases when we try to make quality decisions, so we need to make sure we address them. I am not going to dive into each bias and how to cure it, but do read and acknowledge there is a problem that awareness and good defaults help reduce.

Photo by rawpixel on unsplash

Photo by rawpixel on unsplash

The importance of data

In order to have quality decisions we need quality data to support our decision making. So here comes a short story. In one of my jobs i was managing a group of people doing on call. Now we all know there are many ways to do on call, and they don’t all have to suck, but employee there felt it did.

I wanted that fixed. So I dived head first into data. Who does it, when, when are they paged, how often and so on. The data was outstanding, and it greatly helped us see where are the pain points and fix them.

The same can work for example when you want to analyze and fix other problem such as uptime issues or even in other industries or fields.

This approach is called DDDM, or data driven decision making and it is basically a way to say you make your decisions after you collected data and not just work with instincts or educated guesses. It is a derivative of an approach called DIDM or Data informed decision making, which is mostly in use and grew in the fields of education. The difference between the two is of course debated, as well as which one should be used.

Analysis scale

Analysis scale

So the sweat spot is around 75%–80% of possible data to collect. I am highly against getting to the Analysis paralysis zone, as you never make a decision on time, and probably not a quality one, if it is ever made. The 75%-80% mark is in my opinion another example of Pareto’s principle.

The best decisions in my opinion are timely, accurate as much as possible with the given data, and done after a good period of research, but not too much research.


To Summarize, here are the main points you must have in your back of your mind before making a decision

  • Don’t make decisions when you have decision fatigue
  • Be aware of your biases
  • Collect 75–80 percent of possible data to collect
  • Do not get into analysis paralysis
  • Make a timely decision