While every business and industry is different, there is a number of metrics that I am looking at closely at every subscription business. Not so much for benchmarking, but to understand any seasonalities as well as the baseline for upcoming experiments. This article provides an overview of these metrics.
This metric is calculated by
Install-to-Paid Conversion Rate = # of users who subscribed / # of users who installed the app * 100
For mobile apps that offer free trials (and according to the state of subscription apps report by RevenueCat more than 60% of all apps offer some sort of trial), I am looking at the following two conversion rates instead so that I see the conversion rates after install and after starting a trial:
I know that people like to compare and look at numbers, so here are some numbers mentioned by RevenueCat, but be advised that these differ a lot between industries and countries (you can read more about it in the linked report at the bottom of the post):
The higher your Install-to-Trial and Trial-to-Paid conversion rates, the higher your monetization of new users. As a marketer, your goal is to find the sweet spot of acquiring as many of the right users as possible so that the number of new users keeps growing without harming your conversion rates. As a product/growth manager, you want to increase the conversion rates without harming your acquisition costs, e.g. by personalizing your onboarding funnel or experimenting with pricing and packaging.
In order to optimize for higher ARPU and LTV as well as lower CAC, you will want to segment the data by acquisition channel and other user properties such as location or platform. If you are offering multiple ways to start a free trial from within the app, then you also need to segment based on the source where the trial was started to understand how well each funnel works. You can read thefull RevenueCat report here.
Retention rate is the most important metric to care for if you are playing the long game of sustainable growth. Today I want to focus particularly on subscription retention, i.e. looking at what % of my subscribers renew their subscription WoW, MoM or YoY. In the customer journey, this is the right metric to look at after you have established your baselines for Install-to-Paid conversion rates that I talked about in my post last week (see link at bottom of the post).
Subscription retention is a crucial component for your customer’s LTV. The higher your subscription retention, the higher your ARPPU and LTV. If you are more familiar looking at subscription churn, then that works too since churn can be understood as the inverse of retention, i.e. the lower your subscription churn, the higher your ARPPU and LTV.
Subscription retention differs based on the subscription period (weekly vs. monthly vs. yearly). Here are the median subscription renewal rates of the first renewal based on the state of subscription report from RevenueCat (see link to full report at the end of this post):
While the median renewal rates go down with longer subscription periods, it is important to call out that longer subscription periods still have a lower overall churn compared over the same period of time, e.g. after the 1st month of subscription, the subscription retention for a monthly subscription will be higher than the subscription retention for a weekly subscription.
The in-app subscription management benchmarks report from Qonversion (see link to the report at the bottom of this post) states that 42% of monthly subscribers cancel their subscription during the first month opposed to 78% of weekly subscribers that cancel during the first month. Be aware that the numbers from the two reports don’t align because they are looking at different platforms, industries, time periods, etc. but the overall message is still right.
Another thing to consider is that subscription churn is the highest during the first few days of a subscription, followed by another (but much smaller) peak at the end of your subscription period. Churn per cohort reduces with each subscription renewal. Read full Qonversion report here.
Let me also provide some guidance on how to understand the status of your subscription retention and on how to improve it.
Once you have established an understanding of your conversion rates and retention rates, you can use those to predict the lifetime value of your customers.
For the predicted LTV to work, you are using historical (or estimated) data that predicts future subscription renewals (subscription retention). As such, there is always a level of uncertainty in such models due to potential changes in how users interact with your app over time. Also, historic data is not always available wherefore you may fall back to conservative benchmarks from similar industries.
The most straightforward way for me has been to calculate the LTV as average revenue per (paying) user at certain points in time, such as after 1 month or 12 months assuming monthly and yearly renewal dates. I use 1 month and 12 months because the majority of users churns right after subscribing or close to the renewal date. Note that this period should be extended by the period of your free trial; for instance you would look at lifetime value after 2 months and 13 months if your app is providing a 1 month free trial. Following this methodology, your LTV after 1 month, assuming you are offering only a monthly pro plan option, can be calculated as follows:
LTV (1 month) = Average Revenue Per User after 1st month (=initial Subscription Revenue + 1mo Subscription Plan Renewal Revenue)
This LTV after x months will never provide you with the actual lifetime value over the complete lifetime of your users as you don’t know yet how long your customers are going to stay with you (if you do, then you can multiply the average revenue per user with the customer lifetime), but as your app is maturing, you will be able to predict further into the future and understand the LTV after 2 or 3 years. Make sure that you do not make unrealistic assumptions and start conservatively if you are just starting your business (the image of this post shows a predicted LTV where the retention rate stays pretty much stable after month 3 which is not the most conservative assumption). And do not forget to update your assumptions as you learn more about your business.
If you want to go for a more sophisticated approach, you can build your own ML models that may consider multiple parameters (e.g. key actions performed) or use any of the many SaaS providers that predict the LTV. In any case, just like with conversion and retention rates, you will want to segment your users based on the predicted LTV to understand which segments are the most profitable ones on the short- and/or long-run so that you can optimize your product and acquisition towards those users. You should also look at the cohorted data to understand improvements over time and seasonalities.
Predicting the LTV allows you to understand the CAC payback period as well as the profitability of your business.
If you want to scale your business, you need to be able to scale the input to your growth loops. For that to happen, you need a fast CAC payback period. The faster your CAC is (re)paid by your acquired customers, the faster you can continue spending on new customers.
The payback period is achieved as soon as you have been repaid the exact amount it took you to acquire the customer. It is calculated as
CAC payback period = Customer acquisition cost / Gross profit (Monthly revenue x Margin in %)
While looking at the LTV:CAC ratio provides an idea of the profitability of the business (rule of thumb: you want to establish a ratio where LTV is 3 times higher than CAC), the CAC payback period metric allows you to understand how fast you can reinvest into the business. If your payback period is significantly shorter than that of your competitors, then you can outrun them over time (assuming that your payback period stays somewhat stable while scaling acquisition cost).
Another advantage of the CAC payback period is that it is a metric that you can measure quickly while understanding your LTV:CAC ratio will take some time to establish as you don’t know the duration your customers are going to stay with you as you are getting started. When calculating the LTV:CAC ratio, what I like to do is to calculate it using the LTV at a certain point in time, e.g. LTV after 12 months which tells me the LTV:CAC ratio after 12 months.
When looking at the LTV:CAC ratio, I am considering any revenue and any acquisition costs to calculate the ratio. For the CAC payback period, I have seen it often being applied only for paid acquisition cost, but what I would suggest is to segment it, so that you get an understanding per acquisition channel and even acquisition campaign, and can then scale the ones with the highest payback period.
I recently wrote about viral growth loops (see also here) and mentioned that the k-factor, also known as referral coefficient, can be used to measure the impact of one particular type of viral growth loops: the referral loop. So what is the k-factor?
The most common form of a referral program is that an app grants some sort of reward either to the referrer and/or recipient. This reward can be of financial nature (e.g. a few $ in your wallet) but could also grant you other benefits such as more GB online storage. The success of such a referral program is measured by the k-factor which represents the rate at which new users are joining the app through referrals. The formula for calculating the k-factor is:
k-factor = (Number of invitations sent by each customer) × (Conversion rate of invitations into new customers)
For instance, if each customer sends out 5 invitations and 20% of those invitations result in new customers, the k-factor would be:k-factor = 5 (invitations per customer) × 0.20 (conversion rate) = 1
Apps achieving a k-factor greater than 1 are experiencing viral growth since each customer is bringing in more than one new customer. Conversely, a k-factor below 1 means that the product's growth is linear or stagnant as each customer is referring fewer than one new customer. This k-factor can be quite tricky to measure, specifically if your customers are retaining on your platform for a long period of time and inviting users at different times.
Driving decent growth through referral programs has become tougher over the past years as users have developed some sort of referral program fatigue with every new app offering referral programs. If you are thinking about introducing a new referral program, make sure that your program is providing sufficient value to the sender and/or recipient to make it worthwhile to share. The best referral programs are those that feel inherent to the app where you like to share the app with your friends because it will make the use of the app better for all parties involved.
Also don’t be afraid of experimenting with your referral program (e.g. providing an incentive only to the sender or only to the recipient) until you get it right. The upside of an established referral program is that, once implemented, it is a cheap and consistent way of acquiring new users even if your k-factor stays far below 1.
If you are interested in some more benchmarks, then take a look at my free financial growth model that has a separate section providing an overview of the most important benchmarks of subscription-based apps. Should you use this without thinking? No. The benchmarks tab is here to make life easier when you are starting from scratch. Many people, myself included, look for benchmarks when building financial models or having to estimate or forecast certain parameters such as conversion rates from install-to-paid subscription. Benchmarks essentially provide a shortcut. A shortcut that will not always be correct, but it unblocks you and allows you to move forward for the time being. You can find the free model here.
TBD
TBD