Why I’m always looking at data
Learning how to look at unorganized data can unlock serendipitous discoveries
A couple of weeks ago, a fellow UX researcher was telling me how complicated it is for them to gather and understand quantitative data.
They shared their pain points about uncovering which data points are being collected, where they’re being stored, and how to properly query data. They also said that if it was easier, they would be more inclined to look at data in their day-to-day work.
Pain points aside, it was clear that my colleague saw data as a powerful tool to help us understand our users. Sometimes, researchers think that data is more helpful when it has been organized or analyzed for us — we worry about wasting time querying the wrong data.
Today, friends, I’ll try to tell you otherwise.
The importance of exploration
As researchers, we’ve all had the feeling of joy that comes from witnessing a user share unknown information during interviews. Information that’s not related to the topics we’re researching, but is valuable nonetheless.
This is a perfect example of information we didn’t know we needed, or that even existed, until we heard it. This kind of moment brings us to a new series of questions that are more relevant than what we had originally planned. All of a sudden, you’re making connections that you never thought were possible. Connections that will make your research findings way deeper and more insightful. And that’s all thanks to information that we were not expecting to gather.
We’ve all felt it before, right?
That’s exactly what happens when you’re querying for one piece of data and you happen to stumble across something that you weren’t expecting. Because of this unexpected information, you can now look at your question from a different angle.
Now, tell me this — would you have been able to do this if you were only looking at data that had been nicely organized by someone else? Would you be able to do this with data that was tailored to your initial question?
No, right? Playing around with data and looking at it from different angles is an important part of exploration. And exploration is something that researchers are damn good at.
What you might find
Here’s a simple example: let’s imagine that you’re exploring why a certain flow is not being used as often as your team was expecting.
You put your research goggles on and start to ask questions in order to form a few hypotheses. Maybe an affordance problem on the call to action? Maybe we’re not targeting the correct user need? You start to think about how to get to the bottom of this, perhaps through usability tests or shadowing.
On a hunch, you decide to query some data to get a closer look at whoever was using the flow. You start looking at the data for a couple of minutes without finding much when you notice that the hour when each user entered the flow was being recorded.
You quickly notice that most are accessing the flow after 7pm. You also know that there’s currently a promotional banner running on your homepage that’s been attracting a lot of clicks. Only, this banner isn’t being shown after 7pm. Could that be it?
It’s quite a simplistic example, I know. But it gets the point across. Had you not looked at the data, you could not have easily made the connection.
Ain’t this a job for a data person?
Yes and no. A data person has more expertise when it comes to data, however they don’t always have the same amount of context about the problem as a researcher does. This extra context helps to connect data and insights that others might not easily connect.
Not only that, but some places might not have a dedicated data person or the bandwidth to help you with your data questions.
I like to look at data exploration as an extra tool in the researcher’s belt. In the end, it’s not about being a data expert. It’s about having a curious mindset that will keep you digging for answers and looking for connections. With this mindset, the more data you have, the better.
I’m not telling you to go around building linear regressions or complex algorithms. All I’m asking is that you consider unorganized data as a source of inspiration.
Where do I start?
Some knowledge of descriptive statistics will go a long way to make sense of what you find.
But when it comes to digging around in data I would highly recommend that you learn some SQL or Python. The former will help you with retrieving data and running more simple analysis, while the latter will help you with deal with more complex analysis. It should be very easy to find online courses and tutorials that target SQL and Python for data analysis.
You don’t need to be an expert in SQL or Python since your code won’t be going into production. What you need is just a grasp on how these technologies work and how to adapt it to you daily work.
It’s also worthwhile to share your queries with a data expert — I’m sure they’ll be more than happy to give you feedback on how to improve. For example, a good rule of thumb that I follow is: I make sure to have my queries reviewed whenever one of them somehow changes my understanding of users or their experience.
In the end, data exploration is one extra tool under your belt. Use it as you see fit, just don't shy away from playing around with messy data.
