Car Seat Comfort Comparison
Nature of product/system: Automotive car seating
Persona: Car owners and drivers
Stage at which I join the project: At the beginning
My role: UX researcher
Type of research: Formative, Quantitative
Research Technique(s): Non-parametric one-way, within-subject design of five options
What was the problem?
In initial quality surveys, the seat comfort was rated poorly on some of the three models of a new compact car platform. The seat design group had developed four alternative seats and wanted the UX group to determine which design they should use going forward.
What were the stakeholder assumptions?
They understood the underlying causes for the poor comfort ratings
They knew how to change the seat design to alleviate the problem.
One or more of the test seats would prove superior to those in the customers’ vehicles.
How did you determine the forms of research to conduct?
The test was part of an activity called a “car clinic,” in which new customers were invited to give us their input three months after purchasing a vehicle. The sessions were held in a gymnasium so customers could drive their cars within a few yards of the interview stations. This enabled customers to point to the problems they noted in their initial quality surveys.
Which research method(s) did you choose and why?
Given people are much better at comparisons than they are at absolute judgment, the best way to determine the relative comfort of competing designs is to have users rank the test seats along with those in their cars. Due to differences in the designs of each car model’s seats, data were analyzed separately for each model. The numbers of participants for each model were nine, 10, and 12 for a total of 31.
I used a Friedman’s one-way layout and its range test to determine if any of the test seats were significantly more comfortable than those in the production vehicles.
Did it work? If not, what would you do differently next time?
We were able to draw a few conclusions from the experiment.
There are significant differences among buyers in how they assess seat comfort. Those who bought the sportier model preferred bucket seats that “hug” their legs. Owners of the other two models prefer something more like a living room couch.
None of the test seats were significantly more comfortable than those in the production vehicles.
Peoples’ subjective judgments of product options are best assessed in a comparison format.
The Real Story: From other customer feedback, we learned the source of the poor seat ratings had nothing to do with the seats. The top edge of the front windshield in all three models was lower than was standard for a compact car.
Seeing the roof edge in their upper peripheral vision made the interior feel cramped. The only question on the initial quality survey that seemed relevant to this concern was one about seat comfort, so that question received significantly lower ratings than was typical.
How was your research received?
The Seat Group welcomed the results and followed them.
How did your results bring impact?
The Seat Group abandoned the effort to replace the production seats, saving the company the cost of a redesign and refitting project.
PUBLICATION
I presented a poster session on the application of non-parametric statistics to UX research at the 2011 Usability Professionals Association Conference (now UXPA) in Atlanta.