Churn and SaaS
Churn and SaaS
Join the DZone community and get the full member experience.Join For Free
SaaS companies are faced with copious amounts of advice about pricing, monetization, funnel management and all the different things used to describe the processes involved in attracting, gaining and maintaining customers. One of the big areas that companies think about is that of churn – or how many customers “drop off” and stop using the service. In the SaaS world, companies live and die on their ability to retain customers – churn is a poison pill to them.
Awhile ago I reached out to a series of SaaS accounting vendors to get their different perspectives on what they’re seeing with regards to churn. I thought their comments were broadly applicable to the general SaaS space and so I’ve pasted them below (without specific vendor identification). I’m keen to hear others thoughts on the issue of churn for SaaS.
Here are our thoughts currently: The first issue is what constitutes a sign up. For us it is a full completion of the registration form. However as you know, churn is a tricky thing in the accounting space, and specifically for us with a small business accounting product. The biggest challenge is determining what a lost or lapsed customer looks like, vs a customer who just leaves accounting or bookkeeping until it piles up.
To put a face to that: Imagine that Jane signs up in January and does a whole lot of work. February, March and April go by and she hasn’t come back. We have no way of knowing whether she has abandoned us, or if she’s just busy with other things. Maybe she’ll come back in another month and get caught up, or maybe she’ll only come back once a year.
Coming back only once a year isn’t great for Jane (accounting is way easier if you stay on top of it), and it isn’t great for us either. So we have strategies to encourage Jane to come back more frequently. But if she doesn’t, we’re not prepared to throw her under the bus just yet.
Those facts, combined with the fact that we only launched recently, mean that we haven’t written off any customers yet. Unless a customer has overtly cancelled the account, we’ll continue to protect their data, and to find ways of re-engaging the customer, to either come back and get caught up, or to close the account for good.
I’d always thought a sign up to trial, a conversion to paying, and a lost customer was pretty black and white. It’s only recently I found out various companies don’t count a sign up to trial unless they log in more than twice, don’t include them in churn figures if they only paid for 2 months before quitting, etc. Clearly some vendors are re-writing history and creating self-serving statistics.
Currently we get about 100 new ‘trial users’ a month. Out of that 100, about 50% of them never even really try it out, they log in once to look at it but never enter data and never re-log in, never really sync or integrate with any of our integrations, so no real trials. 10% of them are fake names, company names and information – don’t know really (never paid too much attention) if they are spammers or competitors or what have you. Out of the remaining 40% it depends on the season, during tax season we convert 35 of the 40 to paying subscribers, outside of tax season we convert about 15 of the 40 to paying subscribers.
Out of the paying subscribers we get approximately 50% cancel within 3 months and the other 50% stay on for a year or greater. So out of every 100 trial users we get about 35 paying users temporarily and about 15 long term paying users (during tax season)
Out of tax season for every 100 trial users we get about 15 paying users and about 7 long term paying users
We see monthly cancellation rates of between 1 and 1.5% of the total user base as at that month – it’s an explicit ‘cancel my account’ button that has to be pressed. That’s not quite the same as the monthly churn rate but not far off, I think. On top of that there are ‘suspended’ accounts against expired credit cards that haven’t been updated. The number of ongoing suspended accounts (implicit churn) is somewhat lower than the explicit cancellations.
Obviously we want to make that as low as we can, and we follow up where possible on all of those – from an information point of view, we rarely recover them once they’ve cancelled. We working on a more accurate model of churn – looking back by monthly cohort and calculating the total proportion of customers who signed up with us in, say May 2010, and who have since cancelled. This gives us a better feel for the seasonality of cancellation – are people who join us in April (start of new tax year) more or less likely to stick with us?
In general we’re not hearing about customers leaving us for our competitors but rather moving away from the business, or leaving freelancing, back into employment. That would seem to accord with statistical business survival rates etc: a mean lifetime of 5 years sounds about right (reciprocal of 1.5% churn = lifetime of 66 months) We mostly use that number for thinking about customer lifetime value from the viewpoint of understanding acceptable customer acquisition costs.
Also we’re trying to get a handle on pre-churn alerting – what do users do (or not do) before they cancel? Maybe they stop generating invoices for a month or two, in which case can a well-targeted email or call head off that line of thought?
Published at DZone with permission of Ben Kepes , DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.