Research plan

This project aimed to reduce the search bounce rate of the branded queries, so we wanted to answer why searching for the exact name of apps does not lead to installation. Based on our question, we needed to see people’s behavior while looking for a particular app and find the problem, so we used the qualitative usability test method for this project.

Who were the Participants?

We conducted this with six people in each user group. We had already defined four groups of our users. This grouping was not very accurate. This was based on the apps they downloaded: Casual gamers, Hardcore gamers, People who download functional applications for their every need, and people who download apps just in essential cases.

Defining realistic task scenarios

1- We reviewed search queries first and separated branded queries that did not lead to install most, to write a test scenario based on those queries.

2- Then, determine which queries seem more realistic for each group of users. For example, Free Fire matched the hardcore gamers segment. So we use this word just for them.

3- And finally, we had a table like this. we use the user’s stories from past interviews to contextualize each scenario.

The applications mentioned in the scenario are different from the reality

Test analyzing

I consider two things in each scenario for each user:

Whether they complete it? If not, what was the reason?

Is there an error or problem in their journey? If yes, what happened and why?

I did not measure the success rate, time on task, etc. There isn’t a need to quantify the result. I just wanted to find the whys behind errors. I made users think out loud and tried to find out why sometimes they faced errors or couldn’t meet the goal.

The findings:

We identified several factors that hurt search quality. One of the main reasons branded searches did not lead to installation was that the user couldn’t recognize the app/game among search results. Even when the correct result is wholly displayed in front of them, this happened for several reasons:

We translated the names of English apps/games into Farsi. But some users know apps by their original names. Sometimes we didn’t translate it exactly, it made users’ jobs more complex.

We put the name, description, and translation of apps/games in one line. This was too long and caused people to skip over that item easily.

Some logos weren’t updated in our store. Users didn’t recognize the old logo because they saw the app/game’s current logo on others’ mobile or other places. Some users thought that the game wasn’t updated or it was a fake version of that.

Sometimes the first result came from search ads. users couldn’t realize that the first item was an advertisement. They get disappointed about finding what they want while the correct result is in the third or fourth stage.

Lots of similar search results cause users to get confused. They saw several apps/games with similar names and logos and couldn’t realize which one was the original.

As you see, the search problems are not only related to search engines. Providing up-to-date textual and visual content and the design of each component significantly impact search quality. As we have experienced, solving these kinds of issues is complex in a big company. It involved several teams and people with different planning, and it could be too time-consuming. According to this research, we planned a long-term plan for a few teams to improve Bazaar’s search channel quality.