Following on from my post about NetSuite I got to wondering whether churn matters. It’s a topic that often gets vendors struggling for the right expression to explain what is happening in the customer base. Few want a detailed conversation on the topic and most prefer to stick with talking about top line revenue growth and growth in the annual subscription rate.
In the early days of building a service, vendors want to grab as many accounts as they possibly can. That is natural as they vie for attention. But along the way, they’re going to lose some of those accounts. There are many reasons for this including:
- Misalignment between cost and value delivered
- Insufficient functionality
- Lack of individually required functionality
- Perceived poor performance
- Perceived poor service
- Perceived poor support
- Perceived poor response times
- Unacceptable downtime
- Security breaches
- Billing problems
- Unresolved software bugs
- Unexpected charges
- Unexpected or unacceptable price rises
- Gone out of business
- Service no longer required
It’s a long list to which I am sure others can add. Regardless of which factor(s) apply, it leads to churn. In the SaaS/cloud world, churn is exacerbated by the relative ease with which customers can switch. Don’t like a service? Stop paying and move on. That is fundamentally different to the on premise world where an upfront commitment is often accompanied by hardware and other software upgrades. It is that initial capital commitment plus implementation services that keep buyers locked into an otherwise sub-optimal solution.
From what I can tell, in the early days of service provision, most SaaS/cloud vendors encourage uptake by a group of enthusiasts prepared to sing the service praises. That helps establish a position and tempts others to try the service. But at some point, that occurs around 2-3 years in, attrition hits hard. Buyers who were tempted to try and buy a service start walking away. Again from what I can gather, the degree to which this occurs varies but can easily be in the 25-35% range. That sounds horrific but right now, the market seems sufficiently large that attrition rates at those levels do not matter. There is enough choice in the market for disgruntled customers to go elsewhere. At least that seems mostly true. I am not for instance hearing about services struggling because of churn. Similarly, I am not hearing about customers experiencing real pain as a result of switching, other than occasionally having problems getting hold of data.
Most interesting, I am not hearing vendors talk about winning from a competitor’s service to their own. I suspect that’s because for every customer they win, it is just as likely they lose. There are exceptions. The further up the food chain one goes, the more likely a service will talk about competitor wins. I have seen this with NetSuite, SAP and Salesforce. But even they have their own churn challenges.
NetSuite pointed out to me that their customer churn levels in Q4 2010 were the lowest in the company’s history. Well, they didn’t exactly say that. What they said was retention was the best they’d seen. That’s a slightly different metric and that attrition was half the rate it was in 2009 and, if I heard them correctly, 20% of that experienced in 2008. The trouble is we’re not certain what those rates looked like in 2008-9. But as I thought more about this, it makes sense. NetSuite is the most mature of the vendors offering ERP style packages but then it did change its business model in the 2008 timeframe. That would have automatically led to higher rates of attrition, even though the business grew revenue. 2009 was a tough year for everyone as the recession bit hard. Should we be concerned? In their case I think the answer is a qualified no.
If as happened with NetSuite, the emphasis changes from one kind of customer to another, then it is inevitable that some will fall away. It is how that process is managed that matters. In their case, it was ham fisted at first though things definitely improved in late 2009 and 2010. Do those same qualifiers apply to others?
SAP had a tough time in 2009-10 for several reasons. The economy impacted them to a certain extent but it was the need to re-engineer Business ByDesign and improve performance that had the sharpest impact. Those factors have largely gone away so 2011 will be a telling year. If they can dramatically cut churn in 2011 then we will have another data point upon which to assess their strength in the market. In Salesforce case, customers found that adding functionality made what seemed an inexpensive service much more costly than they bargained for. Those factors have not gone away but Salesforce has sufficient market momentum for it not to matter as they try and win larger accounts where the incremental cost of additional services can be offset by negotiated discounts that are not available to the smaller buyer.
Most of the small business services are being expanded functionally though the services themselves are not changing in terms of scope or addressable market segment. That means their churn rates deserve closer attention. If a vendor is improving functions but is still seeing high levels of churn then it is important to understand why this is happening.
Today, I anticipate that many of the vendors you hear mentioned in these pages will be looking at churn of around 20-25%. That is the natural outcome of services that are still maturing. It is possible that we’re at the peak of churn rates and that going forward we should be thinking about rates declining to around 10-15%. How long that takes will depend on many factors. Basically, vendors will need to reverse whichever of the factors I’ve outlined above are most impacting their business.
Of all the issues I mostly commonly hear about, lack of functionality or delays in delivery followed by service issues are the most common sources of switching.
Why should any of this matter?
- Vendor viability is always one of my due diligence points. Does the vendor have enough in the tank to stay in business? In the alternative, what will my customer’s position be if the vendor is acquired? Similarly, I want to know to what the vendor attributes churn and what they are doing about it. I also want to know the historical churn rates and how it is being managed. Vendors who fight shy of providing this information should be treated with extra caution. Why?
- Regardless of service cost, buyers are making a substantial commitment to learning and using a service. They are placing trust in the vendor and should be able to rely on their promises. We know for instance that in the on-premise world, vendors talk about feature release on an annual basis whereas in the services world we’re likely looking at 3-4 releases per annum. That places a burden of urgency on the vendor which I am more likely to rely upon than with an on-premise vendor. This plays back to my series on roadmaps.
What do you think? Does churn matter to you as a buyer?
Comments on this entry are closed.
Get into the conversation
Hi Dennis, interesting.
At Xero our numbers are public so we can authoritatively say that we see very low churn.
We are this year seeing up uptick in new customers from other SaaS vendors as the industry quickly matures and the different investment levels are becoming more obvious. We monitor that pretty closely and will report on that later in the year.
There is still a massive gap in costs between micro to small business software and midmarket software. As the small end of town adds features they will eat into the mid market as mid sized companies use simple commodity priced accounting paired with a specific line of business applications. But the numbers of churn would so far be tiny.
25% sounds very high and out of kilter from typical accounting software churn. Accounting software is very sticky. I suspect that 25% churn numbers are more related to a revenue/customer recognition issue where customers are counted before they are fully committed.
In enterprise software there is always a long gap between cutting the invoice, implementation and payment. Though in SaaS at least you can’t ship to the other side of the warehouse.
My two cents,
Rod
@rod – yes, revrec issues do muddy the waters so some clarity is needed. But the *general* pattern I am describing does seem common though not universal. 25% in the early period is something I have seen in a lot of places. You are right about stickiness, but I do wonder at what point that kicks in – 1st month? 3 months in? Some later period?
We haven’t seen anything like those churn rates in 4 years, even in the early days. But we take a conservative approach to customer recognition and with SaaS usually the customer can try extensively before they buy so by the time we count the customer they are usually committed.
Interesting.
OK- look forward to your latest results !!
Churn for SMBs is very different from churn for enterprise. At an SMB level it’s very simple, once a product is up and running day-to0-day, the stickiness is already there. With a trial period most users (and Rod can confirm this) will have a good poke around with demo data and from that make a decision. Thereafter they’re hooked. In my discussions with SMB accounting app vendors, churn rates (once signed up and actually using the product) are in the single digits…
Enterprise buying decision however have an element of politics and vested interests and, by extension, the concept of throwing out sunk costs is less problematic to them than to SMBs
My 2 cents
Hi Dennis, interesting.
At Xero our numbers are public so we can authoritatively say that we see very low churn.
We are this year seeing up uptick in new customers from other SaaS vendors as the industry quickly matures and the different investment levels are becoming more obvious. We monitor that pretty closely and will report on that later in the year.
There is still a massive gap in costs between micro to small business software and midmarket software. As the small end of town adds features they will eat into the mid market as mid sized companies use simple commodity priced accounting paired with a specific line of business applications. But the numbers of churn would so far be tiny.
25% sounds very high and out of kilter from typical accounting software churn. Accounting software is very sticky. I suspect that 25% churn numbers are more related to a revenue/customer recognition issue where customers are counted before they are fully committed.
In enterprise software there is always a long gap between cutting the invoice, implementation and payment. Though in SaaS at least you can’t ship to the other side of the warehouse.
My two cents,
Rod
@rod – yes, revrec issues do muddy the waters so some clarity is needed. But the *general* pattern I am describing does seem common though not universal. 25% in the early period is something I have seen in a lot of places. You are right about stickiness, but I do wonder at what point that kicks in – 1st month? 3 months in? Some later period?
We haven’t seen anything like those churn rates in 4 years, even in the early days. But we take a conservative approach to customer recognition and with SaaS usually the customer can try extensively before they buy so by the time we count the customer they are usually committed.
Interesting.
OK- look forward to your latest results !!
Churn for SMBs is very different from churn for enterprise. At an SMB level it’s very simple, once a product is up and running day-to0-day, the stickiness is already there. With a trial period most users (and Rod can confirm this) will have a good poke around with demo data and from that make a decision. Thereafter they’re hooked. In my discussions with SMB accounting app vendors, churn rates (once signed up and actually using the product) are in the single digits…
Enterprise buying decision however have an element of politics and vested interests and, by extension, the concept of throwing out sunk costs is less problematic to them than to SMBs
My 2 cents
Taken from a vendor’s perspective churn is measured against profitability.
From a client’s perpective churn is wasted opportunity, negative ROI, stolen dreams, destroyed careers.
In the headlong rush for growth, software companies ignore (sometimes explicitly) developing strong implementation capabilities either themselves or through partners. SaaS or on-premise doesn’t matter. The implementation activities are braodly the same.
What is different is for SaaS the software cost to get started is low, so it feels like it is easy to get out.
Taken from a vendor’s perspective churn is measured against profitability.
From a client’s perpective churn is wasted opportunity, negative ROI, stolen dreams, destroyed careers.
In the headlong rush for growth, software companies ignore (sometimes explicitly) developing strong implementation capabilities either themselves or through partners. SaaS or on-premise doesn’t matter. The implementation activities are braodly the same.
What is different is for SaaS the software cost to get started is low, so it feels like it is easy to get out.
There is more than one kind of churn.
Customer churn measures what percentage of the customer base leaves every year. So if you start off with 100 clients and 10 leave during the year, you’ve got 10% customer churn aka 90% customer retention. A rule of thumb is the top tier of cloud computing companies have less than 10% customer churn per year.
Revenue churn measures how revenue from the existing customer base changes every year. So if you had $100 in subscriptions from a set of clients last year, and this year you get $95 from the same clients, you have 5% revenue churn. The rule of thumb here is that the best cloud computing companies have negative revenue churn – because of additional subscribers and add-ons, a number along the lines of negative 10% revenue churn is a good benchmark. Some companies will not include add-ons in the revenue churn number, in that case 0% revenue churn is a good benchmark of a well run cloud company.
Either variety of churn is nearly ALWAYS bad for cloud companies. All have relatively low cost of providing service (high operating margins) coupled with relatively high cost of new customer acquisition. Said another way, as compared to enterprise software it costs a lot to bring on a new client, but keeping them on costs vanishingly close to nothing and delivers fantastic ongoing margins. The two main operating metrics all cloud companies track are Net CMRR and revenue churn – the former is more an indicator of business growth, the latter is more if an indicator of customer satisfaction.
In the ERP space – All of the research I have done says that the average time for a business to switch out their ERP applications is around once every 10 years. So if an ERP vendor has greater than 10% annual customer churn you know there is some kind of issue going on.
In CRM as I understand it the typical replacement cycle is more like a 5 years, so you would expect closer to 20% annual customer churn as the norm.
As I hope I described above, it’s very important you ask what a vendor means by churn – it can be misleading if you don’t understand which type of churn they are documenting and how they are calculating it.
To get to your question – churn should absolutely matter to the buyer – it is the clearest indication of satisfaction for the companies using the system – are they voting with their dollars to stay or go. Buyers should also know that vendors are extremely motivated to minimize churn for all the reasons above – so in my experience a high number is a real red flag.
There is more than one kind of churn.
Customer churn measures what percentage of the customer base leaves every year. So if you start off with 100 clients and 10 leave during the year, you’ve got 10% customer churn aka 90% customer retention. A rule of thumb is the top tier of cloud computing companies have less than 10% customer churn per year.
Revenue churn measures how revenue from the existing customer base changes every year. So if you had $100 in subscriptions from a set of clients last year, and this year you get $95 from the same clients, you have 5% revenue churn. The rule of thumb here is that the best cloud computing companies have negative revenue churn – because of additional subscribers and add-ons, a number along the lines of negative 10% revenue churn is a good benchmark. Some companies will not include add-ons in the revenue churn number, in that case 0% revenue churn is a good benchmark of a well run cloud company.
Either variety of churn is nearly ALWAYS bad for cloud companies. All have relatively low cost of providing service (high operating margins) coupled with relatively high cost of new customer acquisition. Said another way, as compared to enterprise software it costs a lot to bring on a new client, but keeping them on costs vanishingly close to nothing and delivers fantastic ongoing margins. The two main operating metrics all cloud companies track are Net CMRR and revenue churn – the former is more an indicator of business growth, the latter is more if an indicator of customer satisfaction.
In the ERP space – All of the research I have done says that the average time for a business to switch out their ERP applications is around once every 10 years. So if an ERP vendor has greater than 10% annual customer churn you know there is some kind of issue going on.
In CRM as I understand it the typical replacement cycle is more like a 5 years, so you would expect closer to 20% annual customer churn as the norm.
As I hope I described above, it’s very important you ask what a vendor means by churn – it can be misleading if you don’t understand which type of churn they are documenting and how they are calculating it.
To get to your question – churn should absolutely matter to the buyer – it is the clearest indication of satisfaction for the companies using the system – are they voting with their dollars to stay or go. Buyers should also know that vendors are extremely motivated to minimize churn for all the reasons above – so in my experience a high number is a real red flag.
Dennis, To your point: “Few want a detailed conversation on the topic and most prefer to stick with talking about top line revenue growth and growth in the annual subscription rate.” I did an investigation of churn for public SAAS companies and blogged about it here: http://www.apptegic.com/blog/2011/8/11/attrition-and-retention-rates-for-public-subscription-compan.html Even for public companies, your comment holds true!
Retention/Churn are crucial to monitor and affect. I think the best way to do that is to measure customer engagement for each customer, understand what makes for a successful customer who doesn’t churn and what factors influence churn, and then address them in the product/business but also by reaching out to customers.
@ben – if I relied solely on what vendors say then all would look rosy in the garden. But it doesn’t. Xero is ONE data point and I am happy to hear how they measure. As to enterprise, I don’t understand what you’re saying re: ‘vested interests.’ That plays orthogonally against what I am seeing and the reasons for buyers going to the cloud…
@ben – if I relied solely on what vendors say then all would look rosy in the garden. But it doesn’t. Xero is ONE data point and I am happy to hear how they measure. As to enterprise, I don’t understand what you’re saying re: ‘vested interests.’ That plays orthogonally against what I am seeing and the reasons for buyers going to the cloud…
How do you assess churn without relying on what vendors say?
How do you assess churn without relying on what vendors say?
Get into the conversation