Ask any B2B SaaS team about churn and you'll hear plenty of theories. Everyone has a view on what drives customer loyalty, yet very few measure it rigorously. If any topic deserves a scientific approach, it's this one. A study by ChurnRX puts several deeply held assumptions to the test.
Is a good NPS linked to lower churn? Do customers who raise lots of support tickets stay longer or leave sooner? Does the length of onboarding affect loyalty? You'll find the answers to all these questions below, and you'll come away better equipped to build a reliable customer health score.
#1. Customer satisfaction
You might assume that the happier customers are with your SaaS product, the more loyal they'll be. Not so. ChurnRX tested this belief by comparing average customer lifetime against the NPS (Net Promoter Score) each customer gave. The result: there is no correlation between satisfaction and customer lifetime.
Want to spot churn before it happens? Put NPS aside and look at other methods such as sentiment analysis. NPS is more useful for tracking how satisfaction evolves over time, or for gauging how certain changes (pricing, product, level of support and so on) affect customer morale. What it can't do is predict churn.
Source: ChurnRX - Customer Churn Benchmark Report
#2. Time to onboard
It's widely accepted that onboarding has a direct impact on short-term churn, particularly for vendors offering contracts with no minimum term. But it also has long-term consequences. ChurnRX highlights how central onboarding is to customer loyalty: customers who don't reach a minimum level of usage within their first 10 days churn at 10 times the rate of others during their first year!
In detail, customers with at least minimal usage in their first 10 days show a churn rate of just 4% at 12 months, whilst those who started using the product after day 10 show a churn rate of 40% over the same period. Extend the timeline and the gap widens: after four years, 87% of customers who got started within 10 days are still customers, compared with just 6% of those who were slower off the mark. Hence the importance of avoiding the classic customer onboarding mistakes.
#3. Achieving results
This is one of the most obvious churn factors: your customer isn't getting the results they hoped for from your solution. Put simply, you can predict churn by asking customers whether your product is delivering what they expected.
To measure this, the study distinguishes three levels of performance:
- Minimal results
- Expected results
- Results above expectations
When customer results are strong, retention improves. Customers who achieved results above expectations have a 'half-life' almost four times longer than those with minimal results. Note: 'half-life' is the time it takes for 50% of a cohort's customers to churn. It's a way of expressing customer lifetime.
You might say that's only logical. True, but what comes next may surprise you.
Source: ChurnRX - Customer Churn Benchmark Report
It turns out that tracking results is always better than not tracking them. Customers whose results are measured carry a lower churn risk than those with no measurement in place, whether the results are good or not!
The proof: after four years, retention among customers who measure their results sits at around 70%. Among those who don't, only 20% are still customers. Building that habit early is one of the essential steps to secure the adoption phase.
Source: ChurnRX - Customer Churn Benchmark Report
#4. Issues raised
Do you assume that customers who never report a problem are happy to stay? It's quite the opposite. Support tickets are actually a clear sign of your customers' effort and engagement.
Customers who have opened at least one support ticket have a half-life of 89 months, compared with just 15 months for those who haven't (5.9 times longer). In short, be wary of silent customers before you worry about the ones opening tickets.
As with NPS, we strongly recommend leaving this indicator out of your health score. Common practice isn't necessarily best practice, and the numbers don't lie. If you're rethinking your model, start with 4 key metrics for a customer health score.
Source: ChurnRX - Customer Churn Benchmark Report
#5. Company size
Another surprising finding: smaller customers have a higher churn rate than larger companies. The half-life gap between the cohort of businesses with fewer than 100 employees (19 months) and the cohort with 100 to 500 employees (94 months) is striking.
We've all seen it: smaller organisations have less established processes, so they're more likely to change the tools that support them. They're also more financially fragile and quicker to cut costs when times get tough. Finally, they're more exposed to acquisitions and mergers, which often lead to a change of platform. It's a good reason to factor company size into your customer segmentation.
Source: ChurnRX - Customer Churn Benchmark Report
#6. Discounts
Thinking of offering a prospect a discount to get them over the line? You may not come out ahead. Surprising as it may seem, the companies analysed by ChurnRX that received a significant discount tended to cancel their subscription sooner than others.
Bear in mind that heavy discounts can erode the perceived value of your product and reduce customer engagement. A customer who isn't prepared to pay a fair price is a customer who doesn't see the value of your solution, and a cut-price deal won't change that.
Source: ChurnRX - Customer Churn Benchmark Report
#7. Free trials
In the same vein, offering a free trial so customers can test the product or a specific feature affects retention. Customers may put in less effort to use your product and disengage faster than if they'd been paying from day one.
The study bears this out: customers who had a free trial show a half-life of 13 months, against 16 months for those who didn't.
That said, we think this finding needs some nuance. Free trials tend to suit simpler SaaS products, often sold on a self-service basis. They can't be compared like for like with vendors that assign a CSM to every customer and, unsurprisingly, achieve better retention.
Source: ChurnRX - Customer Churn Benchmark Report
#8. Upsells
Upsell, or rather the lack of it, is one of the most obvious churn factors. Simply put, customers who have increased the value of their contract are more likely to renew than those who never have.
Three years after signing, the cohort of customers with at least one upsell shows a retention rate of 60%. Among customers who have never upsold, it's only around 30%. A customer who stands still is a churn risk in the making.
Source: ChurnRX - Customer Churn Benchmark Report
#9. Downsells
ChurnRX also looked at the impact of downsells on churn. They found that customers who had already reduced the value of their contract stayed longer than those who had never downsold or upsold.
How do you explain this unexpected churn factor? Moving to a lower plan or reducing the number of users is actually a sign of a healthy account: the customer is right-sizing their subscription to their real needs. And if the solution works for them, there's a good chance they'll renew.
Source: ChurnRX - Customer Churn Benchmark Report
#10. Annual billing
You probably offer both a 12-month plan and monthly billing. Keep an eye on customers who choose to pay monthly: they're more likely to churn than those on annual billing. It makes sense, as paying annually signals a customer's commitment to your product, and it gives you a solid base for preparing and maximising your renewals.
The ChurnRX study shows a very strong correlation:
- Half-life of the annual billing cohort: 88 months
- Half-life of the monthly billing cohort: 14 months
Source: ChurnRX - Customer Churn Benchmark Report
Bonus: setup fees
This 11th churn factor might just persuade you to charge setup fees. The study compared churn between a cohort of customers who paid setup fees when the software was implemented and a cohort who weren't charged. Here are the results:
- Two years after subscribing, retention among customers who paid setup fees is around 40%, compared with around 30% for those who didn't.
- The half-life of customers who paid setup fees is 17 months, against 13 months for those who didn't.
Source: ChurnRX - Customer Churn Benchmark Report
ChurnRX has developed a retention analytics platform that helps SaaS companies reduce churn. They also offer consulting and training on retention, and have built a database on customer loyalty.
For each churn factor analysed, they used datasets covering between 53,000 and 156,000 customers. To measure churn, they produced:
- Declining cohort curves
- Average customer lifetime expressed as half-life, i.e. the time it takes for 50% of a cohort's customers to churn
Further reading:
