When you start optimizing your app stores listings, you always see that the app icon is your best shot at getting an amazing increase of your conversion rate.
This article describes how we approached and executed this optimization. We hope this feedback will encourage you to explore these experiments yourself.
A New App Icon? What For?
Most smartphone users treat the search feature on Google Play or Apple App Store just as they treat a normal Google search. However, the only information available to them to make a decision after they have searched for an app is the icon and name of the app.
We wanted to make sure we maximized our exposure to potential customers, but also conveyed the following:
- Communicate our brand early on and make it consistent with the actual app experience;
- Get away from letters inside the icon (different market, different alphabet);
- Be able to keep the current Wi-Fi symbol and dominant blue color;
Those bullets were listed while keeping a need for consistency in the user journey in mind, as well as a need for internationalization, since we were targeting different countries with different languages and sometimes different alphabets from day one.
We simply searched the internet to get information about an achievable target. Overwhelmed by all the exciting figures, we decided to challenge them and create a new app icon that will increase our store conversion rate by at least 30% in 1 month.
I. First Steps with Freelancers
We initiated a contest on a freelancer website.
This was a quite straight-forward process. We posted (i) the brief of what we were looking for, (ii) the number of app icons we would select (3, in order to A/B/C test the winners), (iii) the amount each one of the selected freelancers would be paid (for their work and for the propriety transfer).
After one week, we got 156 new app icons created by 33 designers. Out of those 156 app icons, we thought 22 were ‘good’ to ‘really good’. We selected (and bought) 3:
The ‘Rocket’ was selected to internally explore the opportunity to picture the speed idea as well as the surf idea; the two Pin/Wi-Fi icons were selected to let us mix what we thought was the strongest design, albeit not perfect.
So after only 10 days, we were already able to start testing the new app icons in a Google Play Store Listing Experiments (as 80% of our users are on Android devices).
The results were… really disappointing. We got a harsh -5% performance with the new app icons. Not only did we have to pay for something that performed worse, but we suddenly realized the process would be longer and harder than expected.
Lesson learned: 30 days aren’t enough…
II. Back to Creativity
We had to be agile and fast. We hired a freelancer, but onsite this time. We asked the person to explore at least 5 different paths and to be as creative as possible, showing only 4 app icons for inspiration and feelings:
After another 10 days (not fulltime), we got a serious amount of app icons:
In order to avoid any new disappointment in stores, we went out in the streets of Stockholm to ask people about which app icons they would want to click on if they were looking for a Wi-Fi app. Nothing scientific here, just a desire to get honest, real and human feedback. The only selection we made was to target a young audience (aged 18–34 to match our customer base).
We created a simple canvas printed on a white A4 paper, with some of our new icons inserted in between existing Wi-Fi apps, and just went for it, stopping pedestrians in Hamngatan.
3 designs clearly won the street survey, and got approximately 80% of the votes: (i) a blue Wi-Fi symbol, (ii) a green Wi-Fi symbol, (iii) a Wi-Fi with a city background.
Results were still painful (as our icons were not the favorites) but we could useful data from it and better prepare our next listing experiment.
Lesson learned: Simple is best / Contrast seems somehow to stand out / People in the street don’t generally want to answer surveys…
III. Google Play Experiment
Thanks to the last survey and a pool we created with our Power Users Google+ group, we could set up a new listing experiment with 3 new app icons we believed in.
The experiment confirmed the survey done in the street, with a dark background that stands out and where simplicity attracts attention.
However, the second app icon did not fit our requirements: we had a problem with the colors and the user experience consistency. We produced new icons, with even more focus on contrast and simplicity.
The results for the app icons #2 & #3 made us believe that: (i) the blue color brings a lot of confidence, (ii) the city background gets the attention by ‘provocation’ (as it got a higher degree of variability in the results).
Therefore, we just mixed the two effects together (icon #3 Wi-Fi has slightly more round angles compare with icon #4).
City background (contrast) did best, as always — the Wi-Fi symbol with sharp angle also performed best.
The results were now getting really exciting since we had a potential app icon matching most of our requirement thanks to the blue color and the use of our current Wi-Fi symbol. Although the performance was still not adequate, we still thought we could get more out of this.
We created a new app icon with an extra-large blue round shape and another one with a darker background. This process made us look at the city background and understand that finally, what matters here is more the contrast effect than the city feeling. The contrast effect could be achieved with something else than a city, and since our app is community-based, we simply decided that we should ‘replace the city with humans’!
The darker city background was good; the larger round shape was also good; but the faces background was our best app icon ever!
We applied it to all our users and stores (as an app icon and app launcher for Android and iOS). It took 10 weeks to achieve what we first thought would be a 4-week process.
Lesson learned: Be bold in your experiments // Get insights from every experiment.
IV. When Attention to Details Nails It
Even though we, as well as our users, were really satisfied with the new app icon, we wanted to finish A/B testing of all possible details: size of the blue round; shape of the Wi-Fi symbol; background contrast; shape of the app icon itself.
We ended up with 6 slightly modified versions of the same app icon:
Some versions had no effect at all on the performance, but as we already learned previously, the darker contrast was again a success. We applied it straight away.
Lesson learned: Get into details // Keep on trying
The whole process took us 16 weeks, which is a lot longer than we expected, however we were rewarded after our efforts with an overall +30% on our Play Store performance.
One important takeaway is that you first need to be bold in your experiment to be sure you are exploring all the required paths.
Although creativity is important, consistency and relevance are certainly even more important, as they help users when looking for an app that will quickly tell them what it does. Unless you are already famous, most of your new users are not looking for a brand but for a solution. Your app icon has to be clear in terms of what it’s about and also stand out. For us, it’s all about Wi-Fi, and all our attempts to get away from it (rockets, animals, etc.) were unsuccessful.
Lastly, keep in mind that this process is relatively cheap, and will increase the effect of every single other action you will take to grow your app.
Just go for it. We will also re-visit the process when the new app icon is one year old.