February 6, 2016

Reducing SaaS customer churn – Part II (Metrics and reasons for churn)

In the first article of this three-part post, “Reducing SaaS customer churn – Part I (The problem)”, we looked at why customer churn – and its related metrics, customer acquisition costs (CAC), monthly recurring revenue (MRR) and customer life-time value (LTV) – are so critical to SaaS business success. We also saw how churn can vary based on where one sits on the SaaS model scale, which ranges from “no-touch” web self-service to “high-touch” feet-on-the-ground enterprise sales.

In this second post we’ll review the various types of churn metrics, try and assess the current levels of churn and finally examine the main reasons why some SaaS customers don’t renew their subscriptions.


Let’s start by defining churn, which can be measured in multiple ways. For example, you can measure only the attrition of your existing customer base, ie without taking into account new customers. Or you can combine them and measure the net. This can then be applied either to the number of customers or to their revenue. If you then also throw in the time factor, you can muddy the waters even further, as explained in this article called “Defining churn rate (no really, this actually requires an entire blog post!)”

So, just as for politicians giving you the latest unemployment figures, it makes sense to know what lies behind the numbers before drawing any conclusions. Not surprisingly, both politicians and software vendors will try and present the lowest numbers possible.

For the purposes of this article, we’ll use what I’ve seen as the most common definition of churn, which is the number of lost customers/(number of existing customers + number of new customers) over a given time period.

So, for example, if a company loses 20 customers out of 100, but also acquires another 100 customers over the same period, then the churn rate is 20/(100+100), or 10%. (If you were measuring only attrition, then the rate would be 20/100 or 20%.)

Finally, note that the inverse of churn is retention, which would be 90% in our example above.


SaaS vendors are unlikely to be totally upfront when it comes to communicating their churn rates. To be entirely fair, this is understandable. After all, we would expect a new market like SaaS to experience growing pains before maturing. We therefore need to turn to sources other than vendors for answers.

Joel York, in his article “SaaS Benchmarks, Acquisition Cost and Churn Challenges”, quotes an average churn rate of 13%, with significant variation based on scale and maturity, ranging from a median of 8.5% for private SaaS companies with more than $10m in revenue to 20% for SaaS companies with less than $10m in revenue. A handful of start-ups even reported churn rates as high as 60%!

Dennis Howlett, in his article “Does 25% churn matter in the SaaS apps world?” talks about numbers from 25-35%. Daniel Druker, from SaaS vendor Intacct, in a well-qualified comment at the end of Dennis Howlett’s article, mentions rates ranging from 5-20% depending on the types of churn (revenue vs customer) and the functional area (ERP vs CRM). Note, though, that even if Dennis and Daniel both use the term churn, they are in reality talking about attrition (ie lost customers only) rather than churn (the net of both lost and new customers).

So, despite the different types of metrics and the many qualifiers, these authoritative sources are quoting current churn rates of at least 10% and attrition rates of at least 20%.

Should this be a source of concern?

  • From a business profitability perspective, “yes”: David Stok, in this  “SaaS Business Models presentation from Matrix Partners” (slide #41), says that 1-2.5% monthly churn is acceptable – though he doesn’t specify if this is an average which is applicable across the SaaS model scale, from low-touch self-service to high-touch enterprise sales.
  • From a prospect and customer perspective, “yes – but”: it all depends on why customers are dropping out. Is it because of things they experienced after going live that would have been difficult to know about before signing up, eg poor service, a security breach or billing problems? Or did they just not do their homework beforehand and then “discover” problems with features, complexity or an overall poor business case? In other words, was the customer almost an accident waiting to happen?


Dennis Howlett, in his article mentioned earlier, “Does 25% churn matter in the SaaS apps world?”, gives a long list of possible reasons for churn, which I won’t reproduce here.

It is important to note that reasons for churn are equally valid in the on-premises world, except that the consequence is not churn, because of vendor lock-in, but customer dissatisfaction or a lack of adoption which, at the extreme, can actually transform the software into shelfware.

My own views on the subject, based on my experience on both sides of the fence, ie on the customer side (evaluating, purchasing, implementing and running enterprise software) and the vendor side (running pre-sales and services in EMEA) is as follows.

For near-commodity products with negligible switching costs, churn can be linked to the usual suspects like price, SLA performance (response times, downtime, service and support) and product quality (missing features or bugs). This is especially true for no-touch, low complexity web self-service products at the low-end of the SaaS model scale.

However, for light-touch and high-touch products at the middle and the top end of the scale, this is hardly ever true. When any of these factors are invoked, the chances are it is a politically and commercially convenient cover for one or more of the “real” reasons (below).

In the real world – ie the one in which a combination of financial commitment, organizational politics and the need to save face prevent certain things from being said out loud – churn drivers can be most often found in a combination of the following reasons:

  • A lack of organizational readiness for the product. Thinking that any company can embrace, for example, CRM, ERP or PPM at departmental or enterprise level and expect results is like thinking anyone can run the New York Marathon. Depending on your physical readiness (in terms of weight, training and endurance), you could be months, or even years, away from entering. Similarly, a minimum organizational readiness (in terms of customers/employees, processes and data) must be in place to ensure adoption.
  • A misalignment between cost and value, the result of an incomplete or inaccurate business case which overstated the benefits and understated the costs and risks. This leads to “surprises” 6-12 months down the line which end up turning the original business case into a dismal case. The contributing factors can be many, and are usually found in the areas of costs and benefits, due diligence and project management. For examples of where the slip-ups can occur, see the high-level SaaS self-assessment survey which I recommend to companies as a first step before embarking on a SaaS project.
  • A reorganization which sees the departure of the original business sponsor. Companies don’t start projects; people do. And when these visionaries move on, the chances are that their successors are less willing to go bat for funding and articulate the commitment needed for the long haul.

Needless to say, these real reasons are hardly ever communicated to the vendor (except off the record – or over a beer …), so the face-saving fall-back is to invoke the usual suspects of price, performance and quality.

Now vendors might rub their hands in glee and say “See, the real reasons are all on the customer side – we’re clean!” Well, not quite. By accident or design, vendors can contribute to customers not doing their homework. During the sales cycle, for example, marketing and sales pitches can downplay these real-world factors – all of which are well known to any B-to-B software vendor – and end up creating customers with a high churn risk profile even before their project kicks off.

And the current pre-IPO high-growth phase of many SaaS vendors, combined with the ease of pitching directly to unwary business execs (with limited or no IT input), increases the chances of this happening.

In our third and final post, Reducing SaaS customer churn – Part III (Levers for reducing churn), we’ll look at how vendors can reduce customer churn by actively addressing these real-world factors right from the funnel stage (thus identifying potential problems much earlier on), and subsequently throughout the life cycle via an active customer engagement strategy that favours customer advocacy.


Speak Your Mind