# Growing app DAUs for dummies — I

Nov 18 · 6 min read

If you are looking at the science behind growing app traffic on your platform, then, the article below, could be an useful read.

Since the beginning of this year, I have this unique task of working towards growing our app DAUs, i.e. daily active users on the app. Five years back, we looked at the same metric, as daily unique visitors on the platform, or traffic on the platform. However, in the world of apps, the measures have changed, but the fundamentals of course, remain the same.

Despite working on this problem for the last 10 months, it has been only recently that I was intrigued about the science of breaking up DAUs into more actionable metrics, in terms of our daily / monthly marketing activities.

The science of app user growth can be broken up into user acquisition & user retention. However, it is definitely not as simple as it sounds to be. The problem becomes very complex when we look at a metric like DAU, which is more difficult to influence than MAU, or monthly active users. Let me explain why.

If an user, say A visits the app on day 1, he is counted as a DAU only on that given day, but the same user is counted as an MAU for the whole month, even if he does not visit the app again. What happens if an user B visits/opens the app on day 2, our MAU number grows by 1, but DAU on the 2nd day still remain at 1. (MAU = 2, DAU = 1). This lands us in a very interesting situation if we want our DAUs to grow, where we would need A to come back again on the app, and hope B also visits that same day, so our DAU grows to 2 on that particular day, unless we have acquired more users already. This now gives rise to the next level of the problem. So, should we acquire more new users or would try to retain the already existing users? Before, we try to get into this complexity, let’s first examine how to formalize the definition of daily active users, when we have thousands or, in some cases millions of users, to look at.

Let i(1), i(2), i(3)…i(R)….i(N) be the volume of daily app installs over 1,2,3,…r,….N days since the app was first released on the app store. For each cohort of users, who are acquired on the rth day, i.e. the set of users who installed the app on that day, which equals iR number of users, there will be corresponding retention on the 1st day, 2nd day, or nth day. What does this mean?

Let’s say, 100k users downloaded the app on a given day. The next day, only 40% of these users will open the app. On the 2nd day, only 20% of the 100k users will open the app. On the 3rd day, only 15% of the 100k users are retained on the app. This goes on diminishing steadily, depending on what is the product proposition. Typically, e-commerce apps have lower retention than social networking apps, for example, as social apps serve a much wider used case, than the need to buy something. However, the same may not be true for a grocery app for example, where the needs would be more frequent compared to high street fashion.

So, these percentages stated above, i.e. 40% is our day 1 retention, 20% is day 2 retention,so on and so forth. For each cohort of users, who installed the app on a given day, these retention percentages of day 1, day 2, ….day N retention can theoretically vary. But, as the used case of the app or the product does not change significantly in a short period, what we tend to observe is that these percentages typically don’t vary a lot during long periods of time intervals. Now, that we have understood what Nth day retention means, let’s look at how we can reconstruct DAUs based on our acquisition numbers and retention percentages.

Let r1i1, r2i1, r3i1….rNi1 be the retention percentages of the i1 cohort of installs. Similarly for the cohort of i2 installs, the retention percentages be r1i2,r2i2,r3i2….rNi2, and so on. However, for simplicity, we can assume these percentages will not significantly vary between i1 set of installs & i2 set of installs. Hence, we can assume an average r1, r2, r3…..rN as the retention percentages over day 1, day 2, day 3…..day N, and so on.

Let’s try to mathematically define DAU.
DAU on a given day = {Sum of installs on all previous days till date} X {the retention percentages of the individual cohort of installs on all those days}

On the Nth day,
DAU (N)= i(N) x r0 + i(N-1) x r1 + i(N-2) x r2 + ………..+ i(0) x rN

or, simply put, using mathematical signs,
DAU (N) = Σ i(N-j) x r(j), where j=0,1,2…..N.
Note: i & r have been defined above.

For a very large N, let’s say an app which has been in the market for more than 5 years, the active base of app installed users could be huge. Often, we may not have the data available for such a long period of time, due to platform migration, account migration, or any other silly reason. First, why is the base of active user base huge? Mostly, because not all users uninstall the app regularly. So, depending on the rate of daily installs and uninstalls, the active base may grow, or diminish, or may find a steady equilibrium. For all practical purposes, for an app which has been in the market for long, has a huge base of users with the app installed, some of them using it often, some not so often. The rate of installs typically tend to balance out the uninstall rate, for a product, which has a good enough used case, and is a steady business, in general.

IF all of the above conditions are met, we can simplify the above equations to
DAU(N) = Σ i(30-j) x r(j) + R X I, where R is the repeat rate of all installs acquired prior to the last 30 days, and I being the volume of all these installs over eternity beyond the last 30 days, while j varies from 0,1,2….30.

This simply means, that the rate of acquisition of app users in the last 30 days and the corresponding retention rates have a significant impact on the DAU today or tomorrow. Depending on the exact volume of installs, and the retention percentages, the impact on daily active users may vary, but the growth in DAUs will be highly sensitive to the volume of installs driven in the short period of time, say, the last 30 days, provided the user retention rates don’t change significantly, to what has been observed typically for the particular app in question.

Thus, in order to grow DAUs, there could be 2 primary means in the short term. The first mean is of course to grow app acquisition or installs in the last 30 days, for example. The second mean is to influence the repeat rate (R) of the already installed base (to grow R X I), I being the active base, minus the last 30 days acquired users.

Does the task for marketing end at acquisition? Definitely not. However, it is an interesting question in itself, how much budget does one spend on acquisition vs re-activation? Can this be more scientific than a gut-based call ?To some extent, it does seem not very prudent to influence the repeat rate through marketing activity, what we should look at is to grow the retention percentages steadily over time, such that in the long run, the repeat rate goes up organically! However, since the universe of app users is limited, we may exhaust acquiring the entire base at one point, and then any marketing dollar spent, will essentially lead to driving repeat users!

Does this sound a little confusing? At least, I am not satisfed myself. I will try to elaborate my thoughts on acquisition vs retention/ re-activation strategy in the next article.

Written by