Endurance – A New Way To Measure Churn
Being in the Hosting industry for over 15 years, an IPO of any Hosting company always intrigues me. You can always learn more information and get new insights into their business from their Prospectus & Roadshow presentations. I closely watched Endurance presentations, and their claim of “1% Churn” got my attention. They have repeated that claim on their Q4-2013 earnings call while defining the MRR retention at 99%. They have kept that statement in their Q1-2014 results. Endurance International Group Holdings, Inc. (Nasdaq:EIGI) is a leading provider of cloud-based platform solutions designed to help small and medium-sized businesses succeed online
Out of the over 50+ companies that were bought over the past years, there are a few which were players in the Hosting market for many years and were struggling with churn throughout those years. Included among them are: HostCentric, Hostmonster, Bizland, homestead, domain.com, Ehost, ApolloHosting.
I wondered what the Endurance team did in order to reduce the industry-inherited churn in these companies.
Before I jump into my analysis of endurance and churn numbers, I would like to refresh your memory about the churn of SaaS companies in general and Hosting ones in particular. See for example, my previous articles. As demonstrated, churn is always an issue in any subscription-based businesses. There are however situations in which Churn is to be expected and is more or less bearable, while in others it is like a cancer to the SaaS organism.
Endurance 1% Churn Model
Before I begin, I would like to present the following disclaimer: as I did not speak to the endurance team nor to analyst which covers them, the below analysis , is based on my own 15 years of experience in the industry and on the material they have chosen to share with their investors during their roadshow. Therefore, I might be completely wrong….
Let’s assume that on January 2013, a SaaS company had 100 new customers joining one of their Hosting brands. Assuming each customer would sign on for a subscription account of $10/Month plan, we would see an MRR (Monthly Recurring Revenue) of: 100 customers X $10/mo. = $1,000 MRR.
Recognized revenues in a SaaS company refer to revenues for services actually being rendered to the customers. And in a Hosting company, in which the service is a monthly hosting service, over the course of each month the customer got the service he paid for, and the company can recognize the resulting revenues for that month.
Assuming all customers pay either on a monthly base or upfront for a yearly plan, their monthly recognized revenues would be equal to the MRR (for monthly paying customers they would recognize all of their monthly payment, while for yearly plans they would recognize 1/12 of their payment for each month of service). $1,000 MRR = $1,000 Recognized revenues.
However, in our case, January 2013 was the 1st month those 100 customers subscribed to the service. In such a case, those customers would join throughout the month, as the sales & marketing activities are spread over the month.
For the sake of this calculation we assume an even spread over the month. Therefore, on average, each customer would get a service of 15 days out of the month. In such a case, they would recognize only 50% of his monthly payment (while deferring the remaining amount to the following month). Therefore, their collection would be: 100 customers X $10 each = $1,000
However, their recognized revenues would only be 50% of that, as they delivered service for 50% of the month, on average: $1,000 X 50% = $500 Recognized Revenues
Impact on Recognized Revenues
On the following month, assuming they would keep all of their customers, the service will be rendered for a full month and therefore: $1,000 MRR = $1,000 Recognized Revenues
This means that the contribution of the customers who joined on January 2013 to their aggregated recognized revenues would be doubled during February 2013, and would remain the same for the following months, assuming none of the customers would leave the service.
Now, in order to demonstrate a bit, let’s look at several months outlook while pinpointing the cohorts:
We can see here, that our numbers are growing on the total user base, while our cohorts are being churned. Clearly, when time passes and the amount of users joining will be equal to the amount of those who are leaving, we will be in the classical “churn trap.” However, all those are not new to us, but let’s look at the Rec. Rev. Model Endurance are using:
Recognized Revenue Model
- They would recognize only 50% of the MRR the new customers generate (on an average 15 days a month).
- 20 customers canceled that month, but on average they did got a service for 15 days on an average.
- Same goes here with 10 customers leaving.
What we could see here, is that by using the Rev. Rec. guidelines we are “flattening” the impact of the churn on the revenues (not on the MRR), by differing revenues of new customers to the following month, and by recognizing revenues & differed revenues of customers who are leaving that month.
I learned few things from this exercise:
- The Endurance guys used in their analysis a model which is not the industry common practice, while using industry terminology.
- Their businesses suffers from the same issues other ‘hosters’ suffer, and most probably at the same level.
- Revenues recognition guidelines may need to be modified for the use of SaaS subscriptions companies, as MRR would better reflect their upcoming financial outlook.
In summary, Churn has always been a key performance indicator of a Hosting company. No hosting company can avoid it, and they all continue to struggle to minimize it.
However, the churn should be dealt with product, operational and service manners and not by the finance department…..
Roy Saar is a partner at Mangrove Capital Partners, a bold but patient venture capital firm helping innovative entrepreneurs start and grow global, disruptive companies. was involved in the launch of Wix. Roy was also the founder of Sphera Technologies (sold to Parallels in 2007), which was one of the very first software platforms for SaaS providers. Roy seats on the boards of: WIX, WalkMe & RFcell