SaaS Churn: Why Revenue Churn Isn’t Always the Best Measure
Churn is one of the most important KPIs that SaaS businesses want to improve. But for businesses with net negative revenue churn, although this may seem great, there are issues that we all need to be aware of.
If you are a SaaS business, especially one where you have 3 or 4 bands of subscription (call it, Basic, SMB, Pro and Enterprise), chances are that customers come on board and stay there. Or at least if they move between bands, they don’t do it every other month. Which means that broadly speaking you get revenue churn. Hopefully quite small, but let’s say you have MRR of $100k in month 1: after 12 months, those customers from that month might now be generating $80k. Naturally, you will have replaced that lost $20k with many more new customers and so your overall MRR will have grown, but you are still experiencing revenue churn (in this example of just under 2% per month).
However, let’s say you are a business like ScreenCloud where you charge per ‘seat’ (or per screen in our case), where people do tend to add seats regularly over time, you are likely to find that you have Net Negative Revenue Churn. What this means is that after 12 months, that $100k might be $125k. In other words, although some customers might have churned, they are more than made up for by the existing cohort expanding their accounts (in this example by just under 2% per month).
Does Net Negative Churn mean you have magically found a cure for churn and therefore don’t need to worry about it?
Sadly no. And in fact it makes things a bit more complicated for calculating other things like Lifetime Value (LTV).
Is expansion really expansion at all?
After congratulating yourself for having a product that makes existing customers want more, the first thing you need to check is whether this is actually expansion, or is it just people on-boarding?’What I mean by that is, are people expanding because your service is just so damn good they can’t get enough of it, or is it because it takes 3 or 4 months for the customer to get all their ducks in a row and fully on-board all the seats that they are likely to have?
This matters a lot as illustrated below. Let’s take 3 scenarios:
Normal revenue churn assumes that your customer revenue drops off in a predictable way. So, of those customers paying $100k last month, if you are churning at a rate of 2% per month, after 5 years, the ones left are paying approx $30k per month.
Net negative revenue churn, however, applies the same logic in the opposite direction: ie that your customers expand revenue in a predictable way (let’s say -2% per month), then in 5 years’ time, the remaining customers from that group are now paying ~$336k per month (more about this at the end).
Net negative followed by normal revenue churn means that initial expansion is more about on-boarding: we might see something like growth for 4 months, then a levelling off, before it starts to fall. This would give a 5 year MRR of $35k.
I’ve created a graph below of the three different outcomes. As you can see, the difference is stark if you mistakenly assume that your expanding customers will do so indefinitely.
In effect, what this shows is just that your starting MRR isn’t $100k, it’s really $100k, plus 2% per month for four months:
Customer churn: a better yardstick?
How can you work out if your customers expand over time or it’s just a normal part of your on-boarding process? The detailed way to do this is to look at cohort analysis of retention over time. If your analytics allows you to look at new customers by month and then see how they behave over time, this can give you a good picture. If I take a snapshot of some recent months growth from the first month from our customers, we can see that, with a few exceptions, our customers end up at around 118% after 6 months. Some get there sooner than others, but more or less, they all end up at the same point. I haven’t shown you what happens after 6 months, but for the purposes of this article you could conclude that our expansion is about 18% from month 1 and is a function of our natural on-boarding process which takes up to 6 months.
But given there are some variations and you don’t want to have to wait 6 months to check whether something is out of the ordinary, is there a better bellwether to monitor the health of your churn?
Actually, it might be as simple as just looking at customer churn, rather than revenue churn.
The reason customer churn is often discounted is because it doesn’t account for differences in types of customers. In my very first example, the difference between someone on a Basic and an Enterprise subscription may be enormous. So just looking at customer churn without the context of the revenues they are responsible for, may mean you either miss something very worrying (eg your very large customers are starting to churn even though your small ones are holding firm), or you worry about something that’s not that significant (eg your churn rate is increasing but its not impacting your bigger revenue customers). However, if you tend to charge by the seat and your ARPU doesn’t fluctuate massively between customers, then customer churn could be a better way to think about churn.
If all you are looking at is customers by volume, then you will never get lulled into a false sense of security by only looking at your expanding MRR. Instead, what you will see is a number that will show how successful you are in holding onto your customers or whether they are leaving you faster than previously (even if that change has been traditionally masked by the expansion from existing ones).
A word about Lifetime Value
Lifetime Value is often is often calculated as:
LTV = ARPU/Churn
ARPU = (Total revenue in a period * Gross Margin %)/number of customers ;
Churn = 1-Retention Rate;
Retention Rate = Gross non-new revenue from this period/Gross revenue from previous period.
So, let’s assume that we are looking at monthly numbers, for an ARPU of $100 and a churn of 2% per month our LTV would be:
LTV = $100/.02
But what happens if you have a net negative churn of -2%? What does that do to your LTV?
LTV = $100/-0.2
LTV = -$5,000 (minus)
That cant be right, right? In fact, even without a formula, we can all see that if your customers overall expand their revenue then the LTV is infinite. But again, that doesn’t make sense logically.
However, as a summary, he suggests we look at two things to figure out what is happening:
- Understand that customer churn will eventually over-ride the impact of net negative churn (depending on the rates of customer churn and expansion).
- Discount future cash flow. There is risk associated to future revenue: even if your customers currently behave in a predictable way. This risk might come from changes in the market, competition, technology, economy etc.
So, the formula factors in net negative revenue churn, customer churn and cash discounting. What’s interesting (and possibly slightly scary for most early-stage fast-growing companies), is that Skok suggests that the discount rate that you apply is:
- 10% for public companies
- 15% for private companies with > $10m in ARR
- and a whopping 20–25% for everyone else!
That makes a BIG impact on your LTV. Taking the example above of $100 ARPU with net negative 2% revenue churn and 2% customer churn. The LTV with no discount applied vs a 25% discount would be $9.9k and $3.9k respectively.
The thing about all of this is that nobody has a crystal ball, or at least if they do it doesn’t work as a way of telling the future. So, none of this will ever be 100% accurate especially if you are growing fast. That’s why for me given our net negative revenue churn, although it’s interesting to calculate our LTV if for no other reason than figuring out whether our Customer Acquisition Cost (CAC) is realistic, as a quick and dirty benchmark, customer churn is a great metric. It also focuses everyone’s minds on retaining customers, rather than the false sense of security that comes with net negative revenue churn.
Questions? Give me a shout — firstname.lastname@example.org