While I was booking a doctor’s appointment a couple of months ago, my mom asked me to send a personal WhatsApp message to my doctor right away. She heard a rumor about him going on vacation for a few weeks.
At first, I thought it was unusual. I assumed the receptionists at his clinic were up-to-date on his holiday schedule. I swatted my mum’s advice anyway, went ahead and booked. She pestered me after to text him and ask. Turns out, he was going on vacation, for 3 months.
He returned on the 17th of April, and I thought, “Great, he’s back. Let me schedule my appointment before his calendar fills up.” And I booked for the 18th. To get an appointment the day after he returns… seems kind of unusual, right?
So I confirmed once again, asked if he was going to take a leave, but they denied it; told me to get my documents for the 18th.
Now, it’s the 18th. I booked 5:30 to 7:30 pm out of my calendar, and I was on my way. As I was waddling through the streets, I got lost. I called the clinic for directions and arrived. Then they asked me why I was so early. To which I said that I had an appointment at 6. They swatted me away, told me the doctor was actually coming by 6.
“Was is it the same doctor I booked with?”
“No ma’am, he’s actually starting on the 25th!”
My brain started to seethe. “After all my confirmations, after all my questions, you tell me this now?” But what can a healthcare provider do, besides profusely apologize?
In essence, the call ended, and I, the customer, churned.
What is Customer Churn?
Customer churn (also known as customer attrition) is when a customer dissolves their relationship with a company, after their last service. The full cost of churn includes both lost revenue, and the marketing cost of replacing the lost customers.
The goal of every business is to curb its churn rate.
What is customer churn rate?
The customer churn rate formula is simple: the number of churned customers divided by the total number of customers.
Number of churned customers / Total number of customers
Where number of churned customers is how many people have left your service over the period out of the total number of customers you had during the period.
Why churn is so hard to understand
That looks pretty straightforward, but how exactly we define those two numbers can greatly affect the output? And what if business-dependent external factors wreak havoc on our understanding of what number comes out?
First, we need to account for what we define as a "lost customer:"
The Importance of Predicting Customer Churn
The ability to predict that a particular customer is at a high risk of churning, while there is still time to do something about it, represents a huge additional potential revenue source for every online business. Besides the direct loss of revenue that results from a customer abandoning the business, the costs of initially acquiring that customer may not have already been covered by the customer’s spending to date. (In other words, acquiring that customer may have actually been a losing investment.) Furthermore, it is always more difficult and expensive to acquire a new customer than it is to retain a current paying customer.
How to calculate Churn Rate?
To calculate your churn rate, divide churned customers over a period of time by the number of customers you had at the start of that period. While overly simplistic, this allows you to focus on churn by cohort and analyze the cause — instead of debating between overly complex methods to analyze churn.
But that’s just the start.
Steve Noble, a data specialist at Shopify, outlined 4 basic ways to calculate churn: (1) Simple, (2) Adjusted, (3) Predictive and (4) the method he ultimately settled on. We’ll walk you through these increasingly complex calculations.
You’ll see the same constants throughout these examples:
The simplest way to calculate churn is:
You’re dividing the total number of churned customers over the period by the number of customers you had on the first day of the period.
The Good & The Bad
The main pro' of the simple version of calculating churn is its simplicity. The churn rate formula is easily understandable and quickly calculable. You only need to know 2 quick numbers to figure out your churn rate for the month, and all you need is those two numbers for each month to be able to compare month to month churn.
The problem with this simple churn calculation though is that it has a hard time dealing with significant growth. When you have a lot of growth, both your churn and total customers can go up. If your total customers goes up more, your churn rate will go down, even when you have more customers churning out of your product than the previous month.
If you’re an established company with a significant customer base and stable growth month on month, this isn’t an issue. But if you’re a new company with substantial new customers each month, this can lead to a strange interpretation where you can lose more customers per month, but your rate will get better.
Here is an example illustrating the shortcomings of the Simple Way:
To calculate churn rate, begin with the number of customers at the beginning of August (10,000). In this example you lose 500 (5%) of these customers, but acquire 5,000 new customers throughout the month, of which 125 (2.5%) churn out. This gives you a churn rate of 6.25% for August.
625 / 10,000 = 0.0625
You're then starting September with 14,375 customers. You see exactly the same behavior this month, with 5% (719) of existing users churning, 5,000 new customers joining, and 2.5% (125) of those customers churning. Your simple churn rate for September comes in as 5.87%.
844 / 14,375 = 0.0587
Wait, what happened? You’ve seen the same behavior, 5% of existing customers and 2.5% of new customers churning, in both months, but the outcome is two completely different churn rates. It looks like your churn rate has gone down, but the underlying behavior has remained the same.
Your high growth has distorted your calculation. In August, 125 churned customers are added to the numerator, but the 5,000 new customers that join in August didn't get added to the denominator—which means that the churn rate is artificially high. In the following months, the growth is less proportionally to the existing customer count, so the effect is lessened.
To account for significant monthly growth, we can take the midpoint of the customer count for the month, rather than using its value on the 1st of the month.
Here we’re dividing the number of churned customers by an adjusted average of the number of customers throughout the window.
The Good & The Bad
This approach manages to deal with the growth issue by normalizing changes in total customers over the time window. Now you have a more stable platform to base your churn rate on, with the time window for your total customers the same as your time window for churn.
However, though this approach to churn rate calculation does deal with the growth issue, it fails to scale with different time windows. Using the same calculation and the same data, you’d get very different answers for daily, weekly, monthly, and quarterly churn.
Using the above data again, now with added October:
Now we see the churn rate as the same, even with a different number of customers at the start of the month.
August becomes: 625 / 12,187.5 = 0.0513
September becomes: 844 / 16,453 = 0.0513
October is: 1052 / 20,505 = 0.0513
Quarter: 2521 / 16,239.5 = 0.1552
Bingo! Problem solved. We can all go home for tea and medals.
Not quite so fast. The main problem with this approach is that it makes assumptions about the data. If you calculate this over the course of 3 months you come out with a churn rate of 15.52%. Divide this across the 3 months and you get 5.17%, very close to the individual monthly customer churn rates. So far so good.
But what if you don’t have exactly the same numbers across each month? Let’s make August a bad month for our imaginary B2B SaaS company. This time, it only gets 100 new customers, 2 of which churn out.
The behavior is the same in terms of churn (5% of existing customers and ~2.5% of new customers), and when calculated individually each month shows the same churn rate of 5.13%.
But when calculated as a quarter, you get a 3-month churn rate of 13.72%, which divided across each month is 4.57%.
August: 502 / 9799 = 0.0513
September: 605 / 11,795.5 = 0.0513
October: 825 / 16,080.5 = 0.0513
Quarter: 1932 / 14,084 = 0.1371
Now our monthly churn rates no longer tally with our quarterly churn rate, even though they use the exact same data. This is because we’ve changed the time window we’re calculating. This approach assumes that churn is spread evenly within the time period, with a linear distribution. But churn is never this helpful. A good churn rate ratio should be able to expand or contract well with the length of time it measures, and still deliver comparable results.
Any good churn rate calculation should give some actionable advice. In this example, Shopify has tried to incorporate a predictive element into the equation. They’re trying to determine a weighted average churn rate, so that rate*customers will predict the likely churn rate on any given day.
InactiveCustomers is an array of how many customers active on day i are inactive on day i+n, i.e. one month later. If you have 1000 customers on September 1, you then look forward in time to see how many of those 1000 have churned on October 1. You sum that up, then divide by the sum of total customers in September.
The Good & The Bad
It seems awesome to be able to predict churn. Having a weight that you can multiply with customers to get predicted churn would be great for planning your finances. Who doesn’t want to do that?
Well, you have probably noticed a critical problem with this approach: “...you then look forward in time...”
This requires two months of data to run one month’s calculation. In order to determine your churn rate for this month, you have to wait until the end of next month. That isn’t good for a metric that is supposed to keep you up-to-date on your company’s success. If you have a number of accounts cancel in September, you won’t have this information until October.
The flip of this is that when you do get to the end of October and have a churn rate, it’s now from a month ago. It’s not current. You can no longer report churn rates to your employees for the prior month, you are instead telling them what happened a month ago.
This approach has all the same problems as rolling metrics, and you know you should stay away from those.
Calculations in SaaS marketing are supposed to take all your data and transform it into easily understandable, actionable numbers. This calculation makes your numbers more complicated and less actionable.
Instead of roughly taking the average of the first day and last day of the month as we do with the Adjusted Way, we can take the average of every day in the month to get a more accurate calculation.
You divide your number churned by the average of your customer count between days 1 and n.
The Good & The Bad
This deals with the issues that plague the other variations. You can use it in periods of high growth, and it scales nicely across different time windows. You can also use it in a timely manner, getting an up-to-date churn rate.
But there are always going to be variations in your numbers that a single calculation can’t account for. Newer customers churning at a higher rate the older customers, differences in cohorts, in plans, in size of account. None of these are captured in this formula, and by using it, companies could have a false sense of security that the number they get each day, week, month, or quarter is the whole story of their churn.
How to reduce customer churn rate?
No need to panic, there are things that you can do right here and right now to combat customer churn.
To get you started, here are 6 ways you can reduce customer churn.
Yes, this may sound obvious, but let’s stress it once again: you have to simply find out why customers decided to leave.
The easiest way to do this is to talk to the customer.
And by “talk”, I mean really talk: getting your customers on the phone is the best option. This way you can demonstrate that you genuinely care, and you can find out what went wrong instantly.
FACT: 68% of customers leave because they think a company doesn’t care about them.
Don’t go down the lazy path by sending the customers exit surveys; just call them up and ask why they left. This will give you the immediate, voice of customer feedback on whether or not your product solves the customers’ problems or causes trouble
In fact, communicating with the customers does miracles in analyzing churn. And you need to be actively using all channels for that: phone, e-mail, website, live chat, and social media. The valuable feedback on how well you serve your customers is just a phone call, an e-mail or a survey away. As simple as that.
Another way to prevent churn is to actively engage your customers with your product.
Often referred to as relationship marketing, give your customers reasons to keep coming back by showing them the day-to-day value of using your products, by making your products, services, offers, etc. a part of their daily workflow.
So, how can you do it?
For starters, provide ample and versatile content about the key functional benefits of your product and offer regular news updates, such as announcements of deals, special offers or upcoming upgrades.
Again, you should engage with your customers on all channels.
According to a recent report from Marketo, the most efficient customer engagement channels for B2B companies to reach out to their existing customer base is through email marketing.
But as for when you should contact your customers, start by analyzing the customers’ journey.
The customer journey gives you a clear picture of all customer interactions across all channels, devices and touchpoints throughout every stage of the customer lifecycle, and be present with the right content at the right place and time. Consider automating engagements using chatbots to reach your customers where they're the most active.
This churn-prevention trick naturally flows from the point above.
You have to provide enough good quality educational or support materials, which will help increase retention and reduce churn. Offer free trainings, webinars, video tutorials, and product demos – whatever it takes to make your customers feel comfortable and informed.
In other words, you have to not only to give them the tools that work, but also offer the training on how to use these tools at a maximum profit. In this way you will demonstrate the full potential of your products and services, and ensure that customers have a successful onboarding and implementation.
The best way to avoid churn is to prevent it from happening in the first place, right?
There is always a group of customers that is more likely to leave than others – so it’s in your best interests to know who is balancing on that dangerous edge. This way you can reach out to them in time to make them stay.
Identifying at-risk customers is one of the most popular churn tactics for B2B companies. In fact, 35% of B2B of organizations have used this tactic to successfully reduce customer churn.
The good news? Spotting the ‘at-risk’ group of customers is easy. Find out which customers have not been contacted for a while. Or maybe they asked for something like a price list, a quote or just more information, and you forgot to follow up on that?
Knowing all this will help you become more proactive in preventing churn.
Also, after analyzing the reasons for churn, you become aware of certain actions, or maybe the lack of actions, that your churned customers made. This knowledge can help you foresee if someone, who is behaving similarly, is likely to leave your company soon.
As sneaky as this might sound, you’d better separate the most valuable customers from the rest and go an extra mile to make sure that at least they are getting what they have signed up for.
Why? Well, let’s be honest, these are the customers you want to keep the most. Valuable customers have to be taken extra care of because they bring in the biggest revenue.
A history of your interaction with the customers can show how deeply they are involved at each stage, whether they had any problems with the product, and whether these issues were dealt with.
So, what you can do is segment your customers into groups of profitability, readiness to leave, and their likelihood to positively respond to your offer to stay. In this way you can better predict customer churn.
Another advisable tip is to offer incentives, such as discounts and special offers, to those customers who were identified as likely to defect.
Not sure how effective this tactic is?
Offering incentives and discount offers is widely regarded as the most effective tactic in reducing churn.
But! Beware that you have correctly evaluated whether offering an incentive is beneficial for you. That means you have to be sure that the costs of your retention program do not outweigh the profits to be gained from the customers you intend to save.
Bottom line – you should not be wasting money on customers who are not likely to bring you substantial revenue.