The Data Science team was working on creating a solution for their corporate banking clients that would enable them to use their customer data more effectively. The prototype was a purely technical solution which demonstrated a POC.
Our goal was to create a user-friendly platform that would enable corporate investment banking clients to use the insights that the data provided, in a simple and effective way.
Before our team worked on designing the new UI, I had to analyse the data capabilities, do an audit of what data we had available and match those with the actual needs of the clients. The initial solution was built from a purely technical perspective and it was my duty to bring in the user perspective.
Initial POC Evaluation
While the data team was very proud of their work, and requested only a UI update, I needed to analyse the interface and show them why it was important to bring in the user perspective.
We had a fair bit of discussion around the importance of designing the interaction from a place of empathy.
We had to make an inventory of the pieces of data that we had available to use in the design. Once I had catalogued the data points, I had to understand the user valued that each of these points represented and add meta data to them so it would be easier to work with once we had more information about the user needs.
Getting to corporate banking clients is not easy. You can’t just walk up to them like you would people on the street. Banks take these big ticket clients very seriously and you have to jump through hoops just to get access to someone in the company. Then you have to set up interviews and make and you’ll be lucky if they have the time and the inclination to sit with you and discuss a new product. Sometimes you might have to go with a banker and let them lead the interview, so make sure you brief them about the questions you need asked.
(<- this picture is just for illustrative purposes, you can’t really take pictures with high profile clients…)
The meta data helped me to classify the data according to the mental model of the users. Once the interviews were done, and from talking to our stakeholders, I had a pretty good idea of what users actually wanted to see and what they expected to be shown to them.
Optimising the Data Visualisation
By classifying the data, it was easy to group individual components into clusters and work with them in the visualisation. Different categories of data (according to the user mental model, not the data scientist mental model) helped us understand how to create visualisations of related data. We tested and iterated on these visualisations until we had something that we could go to market with or test with corporate clients.
Working with the data team was an interesting experience. I personally enjoyed the data visualisation aspect of this project. What I learnt was the most users had no clue what to do with the actual data, nor were they interested in looking at it unless it gave them actionable insights very quickly. The more senior the user was in the company, the faster they wanted insights. It is important to use the data intelligently and bring the wisdom out of it and present it as a story to the user. Just random numbers can be a little useless to someone who has no data training.