Research and production in hyper-speed

I did as much user research as I was able to in the short amount of time. This was crucial, because these were expert users. They may not use dashboards like these every day, but they went through this marketing claims research process many times per year. They knew what to expect, they knew the jargon, and I did not want to waste any time on the "low-hanging fruit" if that was not what they struggled with. 

I avoided chasing the wrong rabbit holes and focused on the main barriers I could see users were encountering, which were:

• Labels. The short labels were ambiguous at times. We fixed this by using longer sentences and tooltips for detailed explanations of each graph, as well as descrptive names for the tools themselves (Predict impact rather than TPR)

• Filters. Some filters would slice the same data in different ways (trends and public sentiment, for instance), some would exclude some of the data (sustainability and health, for instance). But they were all together and this difference was not clear, so I grouped filters by functionality and kept them separate.
  
  

• Navigation. The dashboards and filters were split across dozes of pages, and the relationship between them wasn't clear. Users often couldn't tell why COVID data appeared alongside product data. I reorganised everything to make the relationship between dashboards more intuitive.

There was a lot more we wanted to improve, like the chart labels, which users had to hover to see the words in full. Or the chart formats which were not always intuitive. I left behind a list for future improvements.

Results

The improvements we made in that 1 sprint were significant enough to remove the need for training, meaning the data team was spending less time in meetings and more time in improving the tools.