The Best Way to Measure Retention for B2C SaaS Startups

What is Retention?

Retention is all about how well a business can keep its customers over a set period. In the world of B2C SaaS (Business-to-Consumer Software as a Service), it’s a measure of how many users stick around and/or keep paying for the service after their initial subscription. It’s usually expressed as a percentage, showing how many customers stay subscribed at the end of a period compared to the beginning. High retention rates mean users find ongoing value in the service, while low rates signal potential issues with customer satisfaction or the product’s relevance.

Why is it Important?

Retention is vital for B2C SaaS companies because it directly influences long-term revenue and growth. High retention rates ensure a reliable stream of recurring revenue, which is crucial for sustaining business operations and profitability. Moreover, keeping existing customers is generally more cost-effective than acquiring new ones, as it demands less marketing and sales effort. Strong retention also indicates customer satisfaction and loyalty, leading to positive word-of-mouth referrals and a strong brand reputation. In short, focusing on retention helps SaaS businesses build a stable customer base and achieve sustainable growth.

The best way to calculate it: Cohort based analysis

Note that a picture is worth a thousand words, so read the logic below and then refer to the diagram called "Subscription Retention" around the 1/3rd mark of the following article called - How to Conduct Churn Rate Cohort Analysis to Improve Retention (Article by Chargebee)

1. Pick one metric that shows that the user is retained for your startup. It could be that the user needs to log in 3 times, is still paying for your service, or they must share a post in your social media app. Whatever measurement from your findings that is a good indicator that the user has your product top-of-mind and is using it consistently.

2. Find a cadence of measurement. It could be daily, weekly, or monthly; although monthly is a popular choice because it is a Goldilocks zone where enough time has passed from the previous month to understand if the changes you made and features you built are reflected in the new cohorts one month later. Weekly also works, but daily is too often and only works in very rare cases.

In all the future steps, we will assume a cadence of 1 month, but this could be different for you. You may also download

3. At the end of the month (say June 2024), a new cohort is born. Create a new row for this month and count the number of users that made an account and are retained that month based on your retention metric and cadence. Make a note of these users because you will be tracking these June 2024 users forever.

4. One month later (say July 2024), for the cohort from before, track if those users are still retained and write how many are still retained. As time goes by, every new month will get a new column that tracks how many of the original users from June 2024 are still retained. Ideally, the number will stay the same, but very likely it will go down (e.g. 100 retained users in June, 80 of those users are still retained in July, 50 in August, 40, 38, 35, etc.). So in our example in July we will have 80 out of 100 total users so we will have 80% there.

5. Repeat for next month. Remember that one month later (July 2024) we also need to start tracking a new cohort. So create a new row for this month and count the number of users that have made an account that month. Do the same thing we did in June and remember that this should not take into account any of the users that made an account in June or any month prior.

6. Do this until present day continuously. Keep doing this until the present day, adding a new row and column for every new month, and track how many users sign up in the new row for that new month, and how many users are still retained from your previous cohorts converted to a percentage of the total for when that cohort was born (e.g. what month).

7. Now you have a matrix of sorts that tells you how users are retained in various parts of your product lifecycle. For instance, let’s say you pushed a huge feature in August that created a lot of buzz and fixed a lot of problems. You might see from this retention matrix how effective your feature release was by seeing spikes in retention in August and beyond.

8. To compute churn, you simply take the original users at the start of your cohort measurement (month 0) as the total. Then for month X, your churn will be = 1 - (retained users in month X / total) * 100 (as a percent). So the same matrix that computes retention can also compute your churn.