Neevo avatars
Study and creation of avatars for the Neevo platform
2019 | Defined.ai
Time frame: 1 month
My roles: Researcher, Visual Designer
Disclaimer: The visual language could not be changed completely as we had to stay as much within the company's design system as possible and maintain visual coherence with other platform jobs.
Contextualizing
Addition of non-binary option for users
The Neevo platform previously only featured two gender selection options for users, male and female. After some complaints, it was decided to add the non-binary option for those who don’t identify with the two previous genres.
Note: This is important for later data segmentation.
Also, not all users identified with the old avatars, and there was only one option for each when female or male was selected. Even if the user has identified with the gender, it does not mean that the image generated represented them in matters of color, religion, age, culture, etc.
This scenario could generate a bad relationship between the user and the product and/or harm people with non-dominant identities or who did not feel represented.
Goal
Offer the user a more inclusive option
Target Audience
All Neevo users, which currently include people between 18 and 70 years old of different nationalities.
How
By giving users the option to choose their own avatars (in addition to photo selection) and offer options that are inclusive.
Research
Types of avatars
The first stage of the research was to analyze the types of avatars that exist and try to identify the strengths and weaknesses of each one. The goal was to overview, see examples, and shorten our possibilities to whatever we thought most suited our situation.
Here on the side, you can see some of the types found.
Initials, for example, are impersonal and quick to identify, but you can easily have users with the same letters.
Animals, as well as aliens and monsters, can be fun and increase engagement, but care must be taken so that the image is not degrading, such as automatically choosing a donkey or a turtle for a user (can be associated with stupidity or slowness).
Human characters are the ones that can best uniquely represent each person, but for this to be possible, it is necessary to offer as many possibilities as possible.
Possibilities Test
Three categories
Of all the categories seen, three were selected for us to test and analyze. The chosen ones were "Character,” "Abstract," and "People.” These were chosen because, of all, they were the closest to achieving the desired goals of engaging, causing identification, and being fun.
Analysis
Quick versions of avatars were made following the brand guidelines for the three options to make viewing and analysis easier. The options were evaluated by the team and 5 real users (through an online questionnaire) in terms of how fun, customizable and inclusive they were.
In addition, we thought about how the type of selection should be so that each category delivers the best possible result. In this case, if they would be ready-made options, random or customizable part by part, such as skin color, hair, accessories, clothing, etc.
We also considered the number of variations needed for each option to work and understood the effort required for future development. The more variations needed, the greater the design effort.
And finally, we associate a value from 0 to 5 to keep in mind the possible risk of negatively affecting the user. In this case, the more similar to a person, the greater the risk, because the greater the chances of the user not feeling represented. Therefore, the more abstract, the lower the risk.
Then the analyzed points were superimposed on a radar chart, and a positive (fun, customizable, and inclusive) and negative (risk and effort) average were generated. The characters had the best average between the least risks and the most qualities of the three options. People had a higher positive mean, but the risk and working time were much higher. Therefore, considering the expected time of the project and the final objectives, the characters were chosen.
Final result
Neevo avatars
Neevo, the platform's mascot robot, was used to develop the final avatars. The accessories and the background colors were other factors added to make them stand out even more and generate more identification with the users.
In addition to the first twelve made for the analysis stage, eight more were designed to achieve the project manager's initial minimum number of options.
Final screens
Different scenarios were thought of, such as skeleton loading and error situations. Some things that didn't work so well had to be kept. For example, the error only appears after selecting a huge file. It was suggested to switch to advance notice, but it was not approved as it required more time and worked on the development side.
User feedback after implementation was highly positive. What most caught our attention was this:
"What a joy to see all these colors when choosing my avatar! I had already noticed that the previous illustration was very simple and I, as a black woman, didn't like it at all. I was delighted with the new options! I chose the kitten because I love cats :)"
Learnings
In this project, I learned a lot about representation in product design and the importance of listening to the user. It was a very rich research process.
Unfortunately, there were limitations of time and resources, which prevented us from generating the best possible result in the selection stage. The error scenarios didn't have the best outcome because of this.
The response from users was excellent and reaffirmed what we had seen in the survey, showing that the simple redesign of avatars can bring people closer to the product and engage them even more.