A fellow brought up a great topic, and I’m sharing our conversation here with their permission.
What’s the value of affinity mapping?
Quantitative data (e.g. analytics) and qualitative data (e.g. interviews) go hand in hand. Often, quantitative data only tells part of the story: It tells you what is happening, but it doesn’t tell you why. That’s why we turn to interviews and other qualitative data to uncover insights about user behaviors, motivations, goals, etc.
The challenge with qualitative data is that it’s not cut and dry. Without numbers, how do we make sense of so many voices? That’s where affinity mapping comes in. When conducting interviews with larger sample groups, affinity maps help us identify recurring themes (and outliers) about our users.
When should I use an affinity map?
An example scenario would be if an ecommerce website’s analytics showed people bouncing on the cart page before checkout. You might interview users and organize their feedback in an affinity map to make smarter use of the quantitative data (analytics) you have.
When is it NOT appropriate to use an affinity map?
If you’re on a team where need to “move fast and break things,” you might skip interviewing and go straight to A/B testing two concepts. You wouldn’t know why users chose A over B, but in this case success would be defined by reaching your goal more quickly.
If you only have qualitative data from a few participants (let’s say less than 5 ), you likely don’t have enough data to sort into meaningful categories. Categories uncovered through affinity mapping ideally have at least 5 data points (from 5 separate users)… and the bigger your participant pool, the higher that number should be.
What should I say if asked about affinity mapping during an interview?
Talk about a time where you needed to understand user behaviors, goals, or motivations beyond what you could learn through quantitative data alone. You can discuss how you identified key insights through affinity mapping, and how that helped you move forward in your designs.