So let’s say you’ve got a fair amount of churn in your SaaS customer base (you’re not alone!). You decide to tackle the problem in what seems the obviously correct way: proactively, using AI to identify the most at-risk customers and then retain as many as possible with a rock-solid customer success workflow. But have you ever stepped back and thought hard about what your goals are from this project? It turns out that simply saving as many customers as possible isn’t necessarily the best goal.
If you start with the business urgencies that compelled you to look into your churn rate, you’ll find a more meaningful metric that you should be optimizing for. In the case of customer retention, one good metric I recommend is “net dollars saved”. This measures the revenue you saved from customers that would have otherwise churned, minus the cost of the effort used to try and save customers.
It’s not easy estimating the revenue saved from a customer you brought back from the point of churn (if I pay Netflix $10 per month, what am I worth to Netflix as a subscriber?), and we’ll tackle this very important problem in a subsequent post. For Enterprise SaaS, even if you start with a simple rule of thumb, e.g. a customer is worth 2.5 times their annual contract value, you’ve gone much further than simply trying to get your churn rate down at all cost.
You’ve got an estimate of the revenue you can save if you retain different types of customers, and suppose you’ve also done a thorough job of assigning costs to the efforts required to save them. Now what do you do with that? Your business objective of maximizing “net dollars saved” leads to a shift in the technical objectives — It’s no longer a prediction problem of classifying customers as “Churn Risk” or not, it’s an optimization problem. The ingredients to this “net dollars saved” optimization problem are relatively straightforward, and we talked about them above — estimated revenue saved and cost of effort required. But it gives rise to some interesting sub-problems.
You’ve achieved clarity on your business objective, and these have been communicated to the Churn AI. Sticking with the “net dollars saved” objective, what will the AI’s considerations be as it tries to figure out who the best customers are to try and save and how you should prioritize your efforts?
Is your high-risk customer even worth saving? On the surface, this sounds like a taboo. We grew up believing “the customer is king”, so they’re all worth saving. Well, not always. From an accounting point of view, suppose a given customer is paying you $100/month, but has made a habit of calling your service team several times each month and raising excessive support tickets. All of this ends up costing you a big chunk of the revenue you earn from that customer (if not more than), so your CFO might be grateful if that customer leaves.
Even if a customer is very important to you, you might discover the risk so late or there might be other extenuating circumstances that might make that customer almost impossible to save. You might spend a great deal of time and effort trying to save them, only to be disappointed.
Even if a customer is very important to you, you might discover the risk so late or there might be other extenuating circumstances that might make that customer almost impossible to save. You might spend a great deal of time and In addition to high maintenance and savability, another filter you’d want to consider is the monthly revenue of the customer. For low-revenue customers, you might want to employ a different, low-touch method (likely tech touch, e.g. e-mail).
So now that your objective and approach are clear, what should you expect as output from your Churn AI solution? It’s not just a list of at-risk customers sorted by their risk. It’s a set of recommendations that takes into account all the considerations above, and gives you a prioritized list of customers to reach out to with a clear stopping point. You can follow this prescribed list in order and rest assured that you’re not wasting your efforts on unsavable or high-maintenance customers and that you’re right-sizing your offers to save customers.
