Since the 1960s, data analysts have built reports to collect and share information from computer systems to allow people to better make decisions. In my experience, though these reports are full of information, it’s often not the right information our data customer wants or needs to make the decisions that are part of their job.
This is due to several factors:
- Data needs are changing rapidly as the variety of data available increases. People now recognize the power of the data—if only they could get it in a format they could use
- The volume of data is exploding. Our world has changed, and now there is so much data available, it is becoming a challenge to organize and share the data in ways users can consume it. A dashboard report that is ten pages long is not really a dashboard, and the volume of information lessens its value
- The quality of the data is a challenge. With the rapid increase in data sources, the ability to ensure the data is all valid and in the expected format has become another issue for us to solve
Today’s data analysts are solving these problems, but I think they’re missing the opportunity to do more.
On a recent project, I was asked to look at inbound customer call volume to see if there were ways we could improve customer satisfaction. Nobody likes being on hold, especially when dealing with a billing issue, so we were eager to find better ways to serve our customers.
First, I gathered a year of call details: date and time of call, time to answer, etc. I organized the data by splitting the date from the date-time and then filtered out the invalid records.
It became clear that we had high call volumes during certain periods of the month. This trend was already known, but I also saw a large spike in the abandon call rate that had previously not been known. These are customers who gave up after being on hold too long.
We implemented changes to our call system to address this gap, including a call-back feature so people don’t lose their place in line, and added more staff on the days when we are confident call rates will be higher than usual.
As a data analyst, all I was asked to do was look at call rates. But I saw the abandon rate and used that to find an opportunity to improve customer satisfaction. I felt like a detective noticing a clue.
The testing world needs more data detectives. Let’s look at some other ways we can put our data sleuthing skills to good use.
5 Steps of Data Detection
Columbo was a classic detective on a TV show of the same name. (I guess this reference shows I am not in my twenties!). He used his observational skills and ingenuity to solve even the most complicated crimes. What did Columbo do?
- He entered a scene with lots of evidence to consider (data)
- He organized the data in a method that allowed him to filter and group the data into clues that helped him see what to investigate next (information)
- He used these clues to develop theories about how the crime happened and who was responsible; often, he made false starts and accused people who were innocent, then had to revise his conclusions based on getting a better understanding of the actual events (knowledge)
- Finally, he would figure out the true story and explain how he arrived at this point (wisdom)
- The bad guy would confess and get arrested (action)
OK, he was a TV detective and the dog did all the real detection. So what about a real-world example? We’ve seen these same steps practiced by a brilliant data detective: Albert Einstein .
Einstein is famous for his discoveries and theories, but he was also a data detective. Computers did not exist yet, so data had to be collected by hand and organized on paper. Much of his work required observing behaviors and, from those observations, developing ideas and proposing solutions.
I am confident that he did not just wake up one morning, scribble down “E = mc2,” and then go to breakfast. We can imagine him considering the datahe had available, processing that information into a story and creating rules that we would call knowledge, gaining a true understanding of the issue and using this wisdom to summarize the concept in a way people could understand, and finally taking the action of publishing his theory.
I am not claiming to be the next Einstein—or even be capable of mowing his lawn—but I am saying that the work of data analysis has real value. We have the potential to do more than just report the facts.
Data Detectives in Action
We can do more than just analyze the data we have available; we can use the knowledge revealed by our work to help change happen.
Here are some real, useful actions a data detective can take:
- Listen to your customer to understand not just what they are asking for, but what problem they are trying to solve
- Look at the data available and what elements are the most valuable to serve your customer or their reporting need
- Filter and organize the data into information your customer can understand and that is relevant to their needs
- Demonstrate your understanding of the problem to be addressed and what the information you have collected does to provide knowledge on the subject
- Finally, provide conclusions, recommendations, or observations on the opportunities for change that your work has revealed
- For bonus points: Develop a dashboard or another simple way to quickly measure the impact of any changes made. Your customer will truly see you as an asset to the team if they can report the impact of the changes made on a timely basis
If you are a data analyst, you are likely doing the first several bullet points in your job already. This is a call to action to do more. You may have insights based on your efforts that nobody else has. Take a risk and talk about the changes you think could improve the situation.
You have the evidence in front of you. This is your chance to show you can do more than just generate reports. Become a data detective and be a change activist.
