Cohorts

Understand how much money you collected from customers from a specific cohort. You can also see subscription retention in cohorts and compare it using attribution, country, and other filters

Adapty cohorts are designed to answer several important questions:

  1. On what day does a cohort pay off?
  2. How much money does the app earn for a specific cohort?
  3. How much money can I spend to attract a paying customer?
  4. How long does it take to recoup the ad spend?*

📘

Spending analytics

*Now you can see your spending data in Adapty cohorts. To add it please manually get spends from Facebook as described here and send it to our technical support so we could enrich your reports.

Cohorts work with the app data we gather through SDK and store notifications and don't require any additional configuration from your side.

2562

Weekly cohorts with absolute data for weekly products shown by renewals.

2560

Monthly cohorts for products with trials shown by days.

Cohorts by renewals or by days?

You can analyze cohorts by renewals or by days. The control changes the headings of the columns. Consequently, the approach to analysis also changes.

Tracking by days shows you when you actually receive money from the cohorts. And this may be useful for your budgeting and understanding the time of payoff. You may see that in the mode "by days" the blue color in the table cells is more likely in the middle of the lines because users need some time to pay.

Whereas tracking by renewals shows the retention and churn of the cohorts from one payment to another without consideration of the date. So late users who paid with any delay (it can be months) are added to the number of their subscription period. This approach doesn't reflect the situation of calendar earnings but is definitely more convenient to analyze the retention and churn of the cohorts and get insights from their behavior.

Choose your convenient mode or use them both for more conclusions and ideas.

2546

How Adapty builds cohorts

Let's see in the example of cohorts by renewals how the table is formed. To build cohorts, we use two measures: app installations and transactions (purchases). Every row of a cohort represents a specific time interval: from a day to a year. Each row starts with the number of users who installed the app during this interval and activated a subscription or made a lifetime/non-subscription product purchase.

Every next column in the row shows the number of users who renewed a subscription to this period. M3 stands for month 3 and means that subscribers had 3 consecutive renewals to this point, W7 stands for week 7, and Y2 stands for year 2. Sometimes you can see P2 in cohorts. P stands for Period of subscription. Adapty displays instead of W/M/Y when there are multiple products with different renewal periods present in the same cohort.

We use gradient colors to highlight differences in cohort values. The biggest numbers have more saturated colors.

In the image below you can see a typical cohort.

  1. This cohort displays the data only for weekly products (mark #1).
  2. It shows proceeds as absolute values (mark #2).
  3. The time period we're working with is the last 6 months, and every cohort segment is 1 month long (mark #3).
  4. The Total row (mark #4) displays the cumulative value for each period. $29,5K in the first cell of the Total row accumulates the first period (subscription activation) proceeds from all months (June, July, and so on) until the end of the timeframe.
  5. The first column of the June 2022 row (mark #5) shows the first period (subscription activation) proceeds of $677 from the customers who installed the app in June 2022. The second column of the June 2022 row shows week 2 (subscriptions renewed to the 2nd week) proceeds of $476 who installed the app in June 2022.
  6. On the right, you can see Total Proceeds and APPPU (mark #6). You can read more about them a little further in this article.
  7. Above the table on the right there is also a toggle to turn on user acquisition metrics (mark #7). To see the data of these metrics you should send it to us as a file got from Facebook according to this instruction. The metric will be described below in this article.
1226

How to read cohorts' table

You can hover on any cell in the cohort to view detailed metrics for this period.

846

An overview is provided for each cell

The cells with oblique lines in the background are the periods that are not finished yet, so the values in them might be increased.

1416

Not completed periods are normally the last ones

Filters, metrics, cohort segments, and export in CSV

By default, Adapty builds cohorts based on the data from all purchases. It might be useful to filter all the products of the same duration or specific products. You can also use country, store, paywall, segment, and attribution data as a filter.

600 636 646

There are 3 metrics that can be shown in cohorts: Proceeds, Revenue, and Subscriptions. You can either display them as absolute values or as a relative change from the start of the cohort.

560

You can set the date range for cohorts and choose the segment. The segment determines a timespan for each row of the cohort.

842

Finally, on the right of the control panel, there's a button to export cohort data to CSV. You can then open it in Excel, or Google Sheets, or import it into your own analytical system.

966

Total proceeds and average proceeds per paying user (APPPU)

The total proceeds metric helps you to understand how much net money you collected from customers from a specific cohort and calculate ROAS (Return On Ad Spend). For example, if the ad spend for September 2021 was $10000, and the total proceeds for September 2021 cohort are $30000, ROAS=3:1.

The average proceeds per paying user (APPPU) metric helps you to understand how much net money brings you a paying customer on average. You can then use this number as the maximum price you can pay to attract a paying customer.

Revenue vs Proceeds

Both Revenue and Proceeds are money metrics. You can think of Revenue as gross revenue and Proceeds as net revenue. Revenue doesn't account for App Store / Play Store fee, while Proceeds do. Therefore Proceeds are always less than Revenue (15%-30% less to be exact).

Apple and Google take up to 30% of the price paid by the customers as a fee. For the apps included in Small Business Program (i.e. the app makes less than $1m per year), the fee is always 15%. The rest of the apps (>$1m per year) pay 30% by default and 15% for subscriptions that are consecutively renewed for at least a year. This means 53+ renewals for weekly subscriptions, 13+ renewals for monthly subscriptions, and 2+ renewals for annual subscriptions.

Adapty automatically determines the fee for every transaction your customers make and calculates Proceeds based on it.

🚧

Be sure to indicate that your app is included in Small Business Program in Adapty General Settings.

Spending analytics (user acquisition metrics)

  1. Above the cohort table, you see a toggle that turns on user acquisition metrics.

  2. This feature is in beta now. You need to send us your Facebook data to add it to the report. How to get data in the needed format please read here. Other advertising channels are not supported. So you should consider this option if your Facebook acquisition has a significant share in your ads.

  3. If after the uploading of Facebook spending some periods in the selected timeframe are not covered fully and data for them is not complete their cells will be with oblique lines in the background.

  4. When all the data is set, you see spending analytics as the three last columns of your cohort table:

  • Ad spend
  • Paying CAC
  • ROAS
683

Turn on the toggle to see three acquisition parameters

Ad spend - shows how much you spent on user acquisition in Facebook for the specific cohort.

Paying CAC - is an acquisition cost of paying customer. It’s calculated as ad spend / unique paying users. $20000 ad spend / 1000 paying users = $20 Paying CAC. You can see if your app’s economy is balanced by comparing this number to APPPU.

ROAS - (revenue) is a return on ad spend as a percentage. It’s calculated as (total revenue / ad spend) 100. ($60000 revenue / $20000 ad spend) 100 = 300% ROAS. ROAS=100% means that you earned the same amount you spent. ROAS=200% means that you earn twice more as spend.

🚧

ROAS>1000%

ROAS >1000% usually means that something is wrong with the data. Unless you’re extremely good with UA :) Make sure you’ve uploaded the whole ad spend. If a large chunk of your ads is outside Facebook and therefore can’t be accounted for, please let us know so we can add the integration with the ad network you use.