First, let’s take a nice deep breath, because these statistics might suck the air out of the room:
- The average Android app loses 77% of its daily active users within the first three days after the install, and 90% within the first 30 days.
- 25% of installed apps are never used.
- 26% of installed apps are abandoned after the first use.
What should we do, stop building mobile apps? Of course not — Users spend way more time on mobile apps now compared to web apps (90% compared to 10%), so clearly mobile apps have become everyone’s favorite child. But more than ever, we need metrics to gain total clarity on what users think about our app, whether our app is a sustainable business, and whether the numbers back up our conviction that we’re onto something here.
Growth in installs is probably the single most important metric for visibility in the app store. Growth in installs should be measured week over week, and a number I often hear anecdotally to aim for is 5%. The number of installs you get within the first 72 hours of launch is especially important, which is why you want to do everything you possibly can prior to launch (testing, hyping, validating) to increase your chances of a successful launch. An important point about tracking installs is to use an independent service, not one that has conflict of interest such as an ad network. Here’s a great discussion on the topic: https://news.ycombinator.com/item?id=3951582
The app stores care about retention when it comes to rankings. From your perspective, retention metrics basically capture how sticky your users are, and this is probably the single most important metric for your app’s longevity. The stickier they are, the more likely they are to make an in-app purchase, subscribe, click on ads, fall in love with it, and recommend it to friends to become your free marketing. All of these actions carry monetary value for your app. A single sticky user is worth more in lifetime value than infinite users that download a free app and never come back. Here are 3 measures of stickiness:
- Monthly Active Users: Monthly active users captures what percentage of users that installed your app actually used it over the course of the month. Note that the app store uses an opposite metric, uninstalls, to measure this. Monthly active users tends to drop each month, and at the end of 3 months, the industry average is somewhere around 20%. What you should be aiming for is over 25% at the end of 3 months. If you’re at less than 20%, then you need to find out what the cause is and find out if it’s fixable. I believe that product-market fit should never take more than 6 months.
- Usage Frequency: If you’re an app that’s trying to become habit-forming (health trackers, meditation, journaling, other behavior modification) then your app won’t cut it if users visit once a week. It needs to be an every-day or almost-every-day app.
- Time Spent On App: This one’s pretty straightforward.
3. Ratings and Reviews
There is some disagreement on how much these matter versus installs, but this chart from the marketing team at Fisku pretty demonstrably shows that the algorithm was changed on a specific day to value ratings and reviews more in app store rankings.
Getting ratings and reviews from your users is its own science, mostly about optimizing the right time to prompt your users for feedback. Here is a clever way to strategically ask a user while subtly avoiding a poor rating: https://www.smashingmagazine.com/2014/06/a-better-way-to-request-app-ratings/
4. ROI On Paid Marketing
When it comes to knowing if you’re getting a positive return on paid marketing, you need to know two things: CTLV (customer total lifetime value) and CAC (customer acquisition cost). Warning: it’s about to get a tad mathematical up in here:
Customer Total Life Value (CTLV) Calculation:
1. How much revenue is an average user bringing in a month? Divide total revenue per month by monthly active users. If you’re making $1000 a month and you have 250 monthly actives, each user is bringing in $4 a month.
2. How many months is an average user is expected to be active? This requires you to know another metric called “Churn” which is what percentage of users that installed your app stopped using your app this month, aka “churned out”. Taking the inverse of this (1/churn) acts as a proxy for how long each user is expected to stay with your app. So if churn is 0.60, then each user is expected to stay with you for 1/0.60 = 1.67 months.
3. What is the referral value from an average user? This is pretty complex to calculate, because tracking referrals is not clear cut, but if you have something like a social share embed with referral link, tracking something is far better than tracking nothing in this case. Based on the above metrics, each user is bringing in $4 a month and is expected to stay with your app for 1.67 months, making their total lifetime value $6.68 before referral costs. Now say the average user refers 2 other users. That means the referral value from those 2 additional users is 2 x $6.68 = $13.36.
Now the total lifetime value of the referring user becomes ($4 x 1.67) + $13.36 = $20.04! See why you should go to the ends of earth to track referrals?
Customer Acquisition Cost (CAC) Calculation:
This one’s much easier. Just divide your total monthly marketing costs by total number of app installs. If you spent $600 on ads to get 300 installs, your CAC is $2.00.
Now you can do a magic trick. If the CTLV ($20.04) is greater than the CAC ($2.00), it means you’re getting a positive ROI from your marketing spend. In this hypothetical case, I’d double down on ads, since the CTLV is 10x the CAC, and the monthly value from each user by covers the CAC all by itself.
5. Cohort Profitability
A cohort is a slice of your entire userbase based on geography, device, demographics, marketing funnel, or another defining attribute. The reason to track lifetime value from users of a specific cohort is you may find that users from a particular cohort will tend to spend more, which would give you a higher ROI on ads that specifically target that cohort. For example, let’s say your app takes a portion of Breast Cancer Awareness donations, and your ads target multiple social media platforms. Users recruited from Pinterest might end up donating more than users from Twitter (because Pinterest is 85% female.) If this is confirmed, you’d want to spend more on Pinterest advertising than Twitter.
If you find that you’re not hitting the above growth, retention, and ROI metrics, you need to rule out that performance is the cause. Poor performance metrics can ruin your user experience and cause massive churn. Here are some performance metrics to watch out for:
- App crash rate: 1–2% is the average. If you’re exceeding this, figure out why. You’ll churn like a mofo if your app crashes at inopportune moments.
- End-to-end response time: If your app takes more than 3–4 seconds to load, research shows the majority of users will abandon ship. Ideally you should optimize the response time to 1 second or less.
- App loads per period: This is the total number of transactions called within a given period. Unlike the app crash rate and end-to-end response time, there isn’t a specific number you need to stay above or below. Rather, this is about monitoring and making sure that during periods of super-high loads, the app experience doesn’t slow down or die. You know, that story about being lucky enough to get a flood of eyes from press, and then wasting the opportunity because your servers couldn’t handle it.