Oftentimes when we use metrics we want to find solid arguments to fundament or refute our hypothesis. One way or another we use metrics to take decisions, and these are directly related to the product’s future.
We can measure many things: clicks, performance, navigation, conversion, but there is always a factor considered as subjective: the beauty of a visual interface. How can I say if my interface is beautiful or ugly?
Recently I was working on a sprint where I identified the need to deeply show all the positive values of that study to my stakeholders, I did not just want to show the data I had taken from the benchmark or from the user research, I wanted to go deeper.
How could I prove that “making the screen more beautiful” would reflect in improving the user experience and, consequently, in a better conversion as a whole, bringing real benefits to the business?
How to metrify UI?
This was the question that did not leave my head, how could I deliver value in a data that can be interpreted in different ways depending on the point of view or degree of knowledge about design.
Why do we, designers, tend to think that a “more beautiful” interface is better than an “ugly” one? Whether for personal tastes or “hype”, we carry a quality sense that defines the beauty of something.
With this doubt in my head, I studied about art history and why beauty is beauty, How could I prove this theory that the “beauty” is better than the “ugly”?
After a period studying about ways to metrify and make fair comparisons, I came to a conclusion: numbers do not lie, so if I somehow managed to align expectations and results through a fixed scale I would have a comparative between design versions with different aesthetical aspects.
MY MIND BLEW UP! I had finally found a fair way to make my analysis.
ISSYS — Improvement Score System
Based on the idea of punctuating improvements in a first version of a specific interface I can extract a simple average (maximum score number divided by the number of functions). After I created a new concept of that same interface, working with the same points, I can measure everything again and extract a new average, now you can compare them!
If I could express this crazy count in an equation, would be like this:
To validate the quality of my interface through an interface, I do not ask the stakeholder a number, I just ask if the improvements are identifiable, each response carries a score. I used this values for my sprint:
So this was the interface when I did the first average collection:
This was the result of the first collect:
The redesign result you can see below:
After compiling the data from the first step I did the redesign or creation process. And again I invited the stakeholders to another round of questioning, and to my amazement these were the new collect values.
On a quick observation it is remarkable that the numbers have gone up, and have risen sharply. But, I still wanted to make this clearer for all the team and show the achieved performance by the methodology and how we could use it to compare with our OKRs, or to defend visual improvements hypothesis
In order to reach the maximum score, my scenario was that all the improvements had received the maximum score, that is, all aspects were identifiable and received a score 3. So when extracting the average I would have 12 as a result. And this number ends up becoming my maximum point of perfection of the sprint.
Effectively, comparing numbers is the best way to prove things, numbers are absolute, there is no way to manipulate data if the methodology is applied effectively. At each passing day we have walked to more important business decisions in which designers have been empowered. Design has never been about art, design is about solving problems. We do not solve problems with hypothesis, we solve with fundamented solutions.
I hope with this text to help designers to fundament with more property their hypothesis, show the value of their decisions and even manage business with more certainty of decisions.