Customer Success Management solutions are all the rage these days, and so are myths about churn.
The Venture Capital community loves them, all the experts are saying you need one if you run a recurring revenue business, and each vendor stands at the ready to explain why their solution is the best one.
The one thing they all have in common is the promise to help you reduce churn. As is often the case with any maturing marketplace, the vendors all start to sound very much alike, especially when one message starts to win consistently, but there are some very different views and approaches on how to best accomplish that elusive goal of reducing churn.
So let’s start by debunking some myths that can sound very appealing but may lead you down the wrong path.
Myth #1 – It’s all about product usage
I wrote a blog a week ago about how important it is to watch and track how your customers use your products. I also mentioned that it’s not the Holy Grail. Let me explain. The bottom line driver of retention (and churn) is value – is your customer getting the value out of your product that they expected, or need, to justify continuing to pay you for it? And unless you have a provable ROI built into your product, you are left trying to proxy the value by other means.
All attempts at this proxy are just different ways of measuring customer health. The myth about product usage says that it is the ONLY aspect of customer health. If you take a step back and think about customer health, you won’t think long before you realize that there are a myriad of factors that make up overall customer health. Let me just name a few:
- Support history
- Renewal history
- Contract growth
- Community interaction
- Invoice history
- Survey results
- Executive relationship
I could easily name 10 more but I’ve made my point. And, if there are that many aspects of customer health so easily named, is it really possible that only one matters? I leave that question for you to answer for yourself.
I’d be remiss if I didn’t mention that there’s a secondary factor in whether a customer renews or churns – the overall customer experience. No matter what value you provide, if the customer experience is not a good one, you still run some risk, especially if there are other solutions out there. Incorrect invoices, poor response times, a bumpy onboarding process.
We know what they are and we know they have an impact on the customer experience and, ultimately, retention. I dare you to try to retain a customer by just tracking and responding to product usage (or any other factor), while ignoring the rest of the customer experience.
Myth #2 – You must find leading indicators
The problem here is that this term is misleading and misused. Every element of customer health is, or could be, both a lagging indicator and a leading indicator. They are all lagging indictors because the detection of them, by definition, means that they’ve already happened, thus lagging. They could all be leading indicators because they might tell you that a customer is at risk and might have a higher chance of churning.
Those who believe that product usage is the beginning and the end of the story would have you believe that it is somehow superior (usually because it’s the one thing they can help you measure). Product usage is no different than all of the other factors of customer health I listed above – they are simply facts. Logins went down 22% last month.
Three people at Customer A gave you a customer sat score of 5 or lower. Customer B called Support five times in the last three days. Facts. Nothing more, nothing less. The $64M question is what those facts can tell you. You can argue that the drop in logins is more important than the survey scores and you may be right. You might even be able to prove it.
But don’t buy into the myth that it’s a leading indicator. It’s simply telling you about something that has already happened. It’s completely up to you how to diagnose what it means and determine how to deal with it. Unless you are onsite looking over the shoulder of your buyer while they browse competitors’ websites, you can only interpret the facts at your fingertips and build a case for which ones matter and which ones don’t.
Myth #3 – If you just have enough data/analytics, you’ll figure it out
The problem we are attempting to resolve here – finding and fixing at-risk customers – is not just a simple BI problem. More data will not solve the problem. More ways of slicing that data, or drilling into it, will not solve the problem. You may find some gold nuggets in the mounds of data. In fact, it’s likely you will. It’s not hard to find the low-hanging fruit.
If your product requires logging in frequently to use it effectively and you have customers who haven’t logged in, in a month, that’s probably a metric you should track and flag. But what happens when you are using that piece of data effectively and have removed that problem from your list? What’s next? There may be two or three or 10 more pieces of low-hanging fruit but, at some point, you will run out. The human mind has limitations which is why it invented the computer.
The time will come when you will need a computer’s help to assist you in determining which red flags to pay attention to. Enter the world of data science. Without going into great details, data science will help you uncover indicators that your mind simply can’t grasp.
Let’s use an example to illustrate this:
- 17% of all customers who don’t login within two weeks end up churning
- 8.5% of all customers who open more than 10 Support cases in 30 days churn
- 13.3% of all customers who give you a sat score below 5 end up churning
Now, if your overall churn rate is 20%, none of those facts are really very helpful. You are probably looking for bigger nuggets than that. But, what if you knew the following:
42.3% of all customers who don’t login within two weeks AND opened more than 10 Support cases in any 30-day period AND gave you a customer sat score below 5 ended up churning.
The facts are all the same. The difference is, with the help of a computer and the right algorithms and, VERY IMPORTANTLY, the right process, a new fact appears. One that a human mind cannot possibly have any intuition about. The combination of variables makes that impossible. The best part is that this new fact is one that you would want to act upon, and quickly. Think about how fast you would react to a report that came to you on a Monday morning saying “32 of the following 70 customers will definitely churn if you don’t act immediately”.
I’m guessing that by, say, Tuesday, someone would have called all 70 of those customers. If that flag went up at the exact time when all three criteria were met, you’d probably have a very good chance of retaining these customers if you acted quickly. And note this please – all of those indicators are lagging. They’ve already happened. But acting quickly can make the process feel proactive both to your customers and to your Customer Success team.
The point here is obvious – you can start down this path with some good data and good intuition. I strongly encourage you to start there today. But be aware that you are barely scratching the surface when you do that and that you are laying the groundwork for finding a solution (also a very good thing) that will help you get deep into the guts of all your customer health data and find the pieces that really matter.
Single-metric triggers can get you started but they won’t take you to the finish line. You need a way to analyze the several dozen customer-health related variables in combination. Impossible to do without true data science, not just something casually and commonly referred to as “predictive analytics”. Saying the words, no matter how loudly, does not make them true.
Be wise as you listen to the myriad of messages. Is there really only one factor in measuring customer health? Probably not, right? And if analytics was the solution to this problem, why hasn’t it been solved yet? After all, there are dozens of analytics solutions out there. The answer is obvious – it’s not just about product usage and mere analytics is not enough. This is a very complex and extremely difficult problem to solve.It requires data integration, analytics, workflow, and data science to name just a few pieces of functionality you’ll need in a real solution.