Measuring how people feel about a product or design can be tricky. We’ve been using a method based on Plutchik’s wheel of emotions to calculate our Feelings metric, which gives users eight options like joy, anger, and surprise to capture their reactions. It works, but we’ve heard that the limited range of emotions sometimes makes it hard for our participants to really express how they feel.
Recently, we came across another approach called PrEmo. Instead of words, it uses 14 illustrated emotions to help users express how they feel. It’s a more visual and perhaps more intuitive way to capture reactions, especially for feelings that might be hard to put into words.
Both methods have their strengths, but they also raise a bigger question: What’s the best way to measure feelings in UX research? Is it more useful to give participants a tight, simple set of emotions like Plutchik’s wheel, or a broader, more visual approach like PrEmo?
We’d love to hear your take—how do you think we should be measuring feelings in design?
This is a fun example @Helge! I wonder how the position of the blob around the tree impacts participants’ willingness to identify as one of them. The falling blob is pretty intense
The interesting thing about the sensemaker approach is that it starts with the blobs.
Then asks the respondent to retell a situation relevant to what the researcher is looking for. Note they are not asking for people’s opinions, but to retell experiences.
This is because when you ask people for their opinions they usually start lying to you… e.g. when I ask people what they eat for dinner in a week they would tell me fish two times, pasta, chicken, salad and maybe one day with red meat (welcome to Scandinavia), but when I ask them what they ate yesterday they would say: “I promise, we never usually do this .. but we went to McDonalds”..
The last thing the sensemaker people do is that they say: every analyst has a bias, so the respondent could just as well be their own analyst. They then ask the respondent to rate their experience on a triangle scale (as this creates more friction and thinking than a linear e.g. 1 to 10 scale).
It’s all pretty clever and I’ve used it plenty of times to improve on my own ideas and measurements.
Love this. Feels like a technique that complements needfinding, by trying to go deeper than people’s expressed wants or ideas of how something “should” be.
Captured a comparison table.
Area
Sensemaking
Needfinding
Core goal
Capture real moments and let people interpret their own experience
Uncover unmet needs through stories, patterns, and real behavior
What you ask
A short retelling of a specific situation
A deeper look at tasks, pains, workarounds, and motivations
What you avoid
Opinions, predictions, and surface claims
Leading questions and abstract problem statements
How meaning is created
The respondent tags and places their own story on a triangle scale
The researcher synthesizes patterns across many lived stories
Type of data
Small stories plus structured self-interpretation
Context-rich stories that reveal hidden needs
Strength
Reduces analyst bias and surfaces grounded signals fast
Reveals deeper user needs and systemic gaps
Where it fits
Early sensemaking, pattern spotting, quick cultural insight
Early product discovery, concept framing, and opportunity finding
What you learn
How people interpret their own behavior
Why people behave the way they do and what they are trying to solve
Another thing to keep in mind is that no system for gathering information can remove every form of bias.
In our own system with more than one million people, there are a few factors people should keep in mind.
Up to 5% of people will directly lie for their own perceived benefit. We remove these people, as they ultimately influence a system in a negative way. It’s an ongoing cycle.
Up to 20% of people will white lie, so they are influenced based on how the system directs them. You lean incentives towards
Up to 20-50% will simply not remember correctly. No fault of their own, people are just bad at this. Having multiple points of failure can exclude people from a group.
So in many of our tests, we ask people to react or do something, then comment on that thing the just did to reduce the bias towards recollections, or wants.
We’d love to hear your take—how do you think we should be measuring feelings in design?
