App Marketing

Just released! Our Mobile Benchmarks for Q1 2016

Simon Kendall
Head of Communications

We’ve updated the Mobile Benchmarks - with brand new numbers from Q1 available now. The dataset, taken from our app statistics platform, averages the figures out, and finally compiled them into a single file. What we’ve created is a fascinating study that makes it easy to compare the current benchmarks with your own data, and give all marketers a better understanding of current user behaviour on all kinds of apps.

The best thing? You’re only a few clicks away from finding out where you stand.

To give you some ideas on how to interpret the data, we’ve put together a few insights comparing our current report with our last set of data, that of Q3 2015, to understand how the app world has changed in the intervening months.

Here are three quick trends we’ve found from the data...

‘Utilities’ are a steady performer, particularly on Android devices

Apps categorized as Utilities have performed consistently, and well, when comparing between 2015 and 2016. In Q3 2015 Android utilities apps finished with a lower retention rate by day 30 (.069 as opposed to .87), but performed much better across all three other categories on all periods versus six months ago. Perhaps this could indicate that while users may know what they want, once they find an app they like they’re likely to keep it. As such, they’re likely to drop off quicker, but use the app more if they find a use for it. This is further evidenced by the number of sessions per user on Android, which average at 2.12 per day - a figure only challenged by business apps and games.

Are business apps getting better?

Speaking of business, retention rates are slightly down from last year in the category (moreso on Android), but time spent by a user on an app has seen an increase overall. The trend is interesting to spot, and could suggest that apps are improving in quality on the year before. Either that, or people who use business apps regularly received an upgraded mobile device for Christmas, and are now using their phones more. Out of all apps business has the third highest retention rate on Android (behind publications and utilities) on day one, but by day seven it’s one of the lower categories for returns.

When comparing time spent by year, in 2015 users were spending 30 seconds on an Android business app by day 30. This was over double in 2016, at 70.5 seconds. On iOS performance was still up from last year, but apps on Apple devices did not perform quite as well versus phones running Android.

Reduced engagement with social apps?

Let’s look at social, one of the strongest performers of all. Interestingly, social apps have seen a fall in sessions over the last six months, dropping an average of a session per day for each respective OS. Retention rate has seen a fast drop, although on Android the fall has been slower.

It’s not all doom and gloom! Time-per-session has seen a big increase on both platforms, following but improving on the general trend of 2016. Perhaps in general users are a little more demanding of a quality experience, but also more likely to keep on returning to the apps they love.

That’s all on analysis for now. Later in the month we’ll be releasing a longer article about retention rates - keep an eye on the adjust blog, our post will go into much more detail about how retention can be leveraged for reporting and optimisation too.

Using the CSV

Each KPI is described through quartiles: q1, q2, and q3. For most purposes, you can look at the median of each metric (q2). You can then use the q1 and q3 metrics to inspect the distribution of the metric and find out precisely where you fall.

The CSV is formatted for use with Pivot Tables, for example, but also works really well in SQL databases and SQL-like setups.

  1. Open a new Excel spreadsheet.
  2. Hit File and the Import.
  3. Pick CSV file.
  4. Find the .csv file in your Downloads or other folder.
  5. Hit “Get Data”. In the next dialog, select “Delimited”. The delimiter is a single comma.
  6. You can typically use the default settings for the column data types.
  7. Ensure that your decimal separators are periods and that thousands separators are blank. In certain languages, it’s common that Excel sets localized separators. In the third dialog of the import process, hit “Advanced”, and set the decimal separator to “.” and the thousands separator to “ “.
  8. Finish the steps by clicking finish.

Prefer querying it over SQL? Paul did a writeup a while back on how to load CSV sheets just like these into SQLite.


The Mobile Benchmarks are aggregated averages derived from over 14,000 apps tracked with adjust. Apps that have too little data to qualify fully through the dataset (that is, fewer than 30 days’ of cohorted data in the examined timespan) are removed from the sample, as are apps that can’t be readily assigned to a particular vertical (for example, where they are distributed outside of Google Play or the iTunes App Store.)

Also, we’ve removed verticals that have less than a few dozen apps, in order to keep the averages reliable. The verticals are mappings to the categories in the app stores, very similar to how we did it for the last report. This means that in places we’ve mapped certain verticals together to maintain a sufficiently large sample size.

The insights above are not perfect comparisons. The samples vary somewhat between the two reports. Our focus in producing this dataset was to prepare robust averages that app marketers can use to benchmark their own performance. For insights into trends, it would have been better to use precisely the same sample of apps year over year - but this would have resulted in smaller sample sizes and greater variation, defeating our core benchmarking goal. Additionally, there may be longer-running trends and seasonalities that can’t be reflected by comparing two consecutive reports like this.

It would be very interesting though to look into trends and changes over time with a more robust approach.