Data Analysis
I use data to answer a product question: where people struggle, whether a problem is widespread and whether a change helped. The work starts with a question and a baseline. Behavioural data becomes much more useful when it is connected to the task someone is trying to complete and the feedback they give along the way.
01 / The practice
How I work.
Choose the question before the chart
I define the decision we need to make and which signals could inform it. That helps distinguish a useful success measure from a number that is simply easy to collect. It also makes the limits of the evidence easier to discuss.
Read behaviour in context
I combine product analytics with user feedback and journey observations. A drop-off can point to friction, but understanding the surrounding task helps explain it. I look for patterns that can be investigated rather than treating a dashboard as the entire answer.
Return to the original hypothesis
After a change, I compare what happened with what we expected. The useful question is whether the experience improved for the people it was meant to help. Unexpected results are a reason to learn and adjust the direction.
02 / Something useful
What comes out of it.
- A question, baseline and success measures
- A view of friction and opportunities
- A clear readout that informs the next decision
03 / In practice
The work behind the words.
Measuring the content-search problem
Pendo analytics helped quantify the difficulty of finding educational content. The project also used stakeholder feedback from teams working on published courses to understand the effect of the new search experience.
Understanding the purchase journey
At Emerald, Hotjar, Google Analytics and journey mapping supported hypothesis validation and decisions about the e-commerce experience.