I have wanted to tell you how I did the analysis of average time per screen in an app and to what extent it helped the product. Knowing more about user behavior is one of the big questions for marketing and product, and we have many ways to find out. Today I come to tell you one, which can serve as a complement to other analyzes (or not).
In my case, we knew that we had a good app design, the usability was good and the user was happy. However, we wanted to know more, somehow quantify the status of each screen, and above all, clear up some issues that had been coming up in team meetings for a long time.
First I will tell you a practical case of a real application, then I will talk about all the advantages that I have found in the analysis, and finally, he explained how it can be carried out.
📊Practical case: Analysis of the average time per screen in an app
It was a time when mobile analytics was not fully dedicated by any member of the team and therefore there were no analytics tools beyond Google Analytics, Firebase …
Does the established flow make sense? Do the registry changes compensate for the increased registration time? Are there any steps / screens left over? What are the differences between new and recurring? Are there differences in use between iOS / Android devices?
The analysis of average time per screen helped us to clear up many of these doubts and above all solve some hypotheses that we had about useless screens. Some examples:
- Hypothesis 1: The registration was too long and tedious: We finally learned that no. The users who increased the registration time were the new ones that we had just captured in a Facebook campaign.
- Hypothesis 2: Mandatory registration with verification SMS was a major obstacle to registration, thus increasing CAC: We learned that noSurprisingly, time and ACC had improved.
- Hypothesis 3: The “view sale” screen was unnecessary and was hardly used by users. It was found that Yes, it was a little used screen. At the business level it made sense but in reality it was not used.
- Hypothesis 4: Chat is the most important part of the product and where we have to focus the improvement focus and incorporate new functionalities. We were able to determine that Yes, is one of the most used screens (even filtering by those users who gave us the most value).
The main advantages of analyzing the average time per screen have to do with understand user behavior when they perform usage flows in an app, for later implement product improvements:
Improve user experience
In our case it was a complex app, with a double role and many different screens. This analysis was a perfect complement to implement improvements at the product level.
Simplify processes and make the app more attractive
Make an app simple, easy to use and understand. If for some reason it can’t be simplified further, use time analysis to look at the differences between new and returning users. Focus your improvement efforts on the most used and important screens for common usage flows.
We accepted several hypotheses as valid, perhaps the most important had to do with the use of a “budgets” screen, which we thought would be very important, and yet the user preferred to use the chat to determine the price of the service, etc. .
Why is this screen used so much? Why this strange flow? In my example, this analysis helped us to detect that on certain Android devices, when you went backwards, it took you two screens back instead of one, thus generating a strange flow of use.
Acquisition cost reduction (CAC)
Very important at the marketing level. The analysis of the average time in an app (along with other metrics) can be focused on the registry, testing different versions until finding the best one in terms of acquisition costs.
✍️How to perform average screen time analysis
1. Access to Firebase or GA
Accede a Firebase o Google Analytics to be able to download the necessary data.
Google Analytics: It can be accessed from “Behavior” or from “Events” -> “All events” -> “screen_view”
NOTE: At the time of writing this article (April 2020) Firebase is migrating the analytics part to Google Analytics, and therefore, it does not yet allow the data to be exported in CSV, so I recommend downloading it from Firebase.
2. Evento screen_view
To carry out the analysis we will use the automatic event screen_view where the average time per screen of the application and the column “average user interaction” are collected. Download the data in a CSV.
NOTE: I recommend having a document with the names of the screens to be able to understand which one is as well as understand if they are reused in various places in the app. If we do not have this documented before the analysis we will need the developer to tell us one by one what each screen is. If there are large screens that are represented in several (those screens with bullets) or that are used in several sites of the APP, we will have to give them a specific name.
3. Data processing
Make the data processing necessary to be able to use the information. In my case I converted the “0.233434” into “23%”, I ordered the screens in homogeneous groups, for example:
Chat = (MainChat + ChatRoomActivity + MegachatActivity)
4. Visually represent the data.
From the use of a table to an interactive graph. In my case it was for a presentation ppt so I did it with Google Slides.
Remember that the most important thing about this analysis is have a clear objective to analyze and perform the necessary segmentation (country, device, time interval, …) and especially in which parts of the app we want to focus: registration, login, purchase flow, etc.
What do you think? If you like it, feel free to share. To share is to live! 💙