SaaS Economics Blog

    Why Shopify's churn article doesn't matter for 95% of SaaS companies

    by Brad Coffey

    Shopify recently posted an article on defining churn rate and the 'correct' way to calculate the metric and the limitation of the more standard approach.  Steven Noble is completely right in the importance of churn.  It's a massively important metric for SaaS companies and can kill even the most successful companies and cap potential any revenue potential if not held in check.  However his assertion that many companies get it wrong is incorrect (well at least for the reasons he states)

    Why the Shopify article on churn is wrong for 95% of SaaS Companies

    Where the article fails is in the proposition that most companies are calculating churn wrong because of the impact of churns from new customers within the period.  

    "The problem here is that the [number of churns over period] value is affected by the entire period but the [number of customers at beginning of period] value is a snapshot from the beginning of the period. This might not have much impact if new customers only make up a small percentage of your user base but for a company that's growing this can lead to some major misinterpretations."

    Noble goes on to describe how you have to not only account for the customers at the beginning of the period but also the customers who churn over the period.  He shows some interesting math on the topic - but what he misses is that for most SaaS companies they bill customers monthly (at the shortest interval) so customers only have one opportunity to churn in a month - and similarly new customers can't churn within the first month they sign up. For instance if you have a product that is $100 a month, and a customer signs up and pays you on the 15th of July - that customer can't churn until their next billing date, the 15ht of August.  So that 'churn from new customers' isn't relevant within the month your calculating churn. 

    The real open questions on defining churn

    There are however some more interesting questions surrounding the right way to calculate churn.  Specifically there are two large questions we think about at HubSpot:

    • Should upsell be included in churn?
    • Should customers locked into long term contracts be included in churn?

    For HubSpot we've determined that we should actually calculate both customer churn (excluding upsell) and revenue churn (which includes upsell and is weighted by the recurring revenue from each customer).   Customer churn ensures that you're delivering continued value to all your customers and MRR churn ensures you're getting the economics of the business correct.

    For locked-in customer we've determined that yes - you should include those customers.  The simple reason for this is that if only looked at customers up for contract renewal you could incorrectly move the metric short term by selling more month-to-month deals (at a higher overall churn rate) once at steady state.  For instance if customers who sign 12-month upfront contracts churn naturally at 1% monthly - when the contract was up this month the end of the year they'd have a 1-(.99%)^12 = 12% retention rate in that one month their contracts are up.  If you instead sold month-to-month last month that churned at a higher 2% monthly rate, those month-to-month contracts would also renew this month and blend in with the 12% churn from annual contracts lowering the calculated churn rate that month even though overall they were a worse cohort of customers.

    As illustrated by this topic SaaS metrics are at best inconsistently applied and can be really hard to compare from company to company.  The best thing to do is find an approach you can apply consistently internally and work to push those metrics in the right direction - and worry less about the exact benchmarks you're hearing from friends and VCs. 

    Brad Coffey

    Written by Brad Coffey