Simple Customer Health Scoring: Start with Risk

Customer health scoring doesn't have to be complex and overwhelming. If you want something that truly works, just start with the obvious and known customer risks


Joseph Loria

4/19/20232 min read

Why is it that customer health scoring is so often discussed as if it were some highly complex construct? It’s as if experts hiding deep underground were crunching nightmarish multivariate equations to determine such obscure measurements.

My consistent experience has been twofold:

  1. At its basic level, health scoring is simply knowing the common, major risks in your customer relationships;

  2. Whether or not they are tracked, those risk factors are typically well-known by your post-sale team members.

And that’s all you really need to start some highly predictive health scoring.

A Scaleup Story

I showed up at a company that had grown dramatically in its early years only to flatten off. Despite selling net-new customers at a good clip, all that incremental revenue was being canceled out by churn. I started by asking my team, “So tell me, what are the typical signs that things are going to go South?” Pretty immediately they came up with traits they had noticed: customers dragging their feet on implementation, or straying from their monthly cadence, or putting a certain type of role as the application owner, or suddenly changing leadership, etc.

So the team empirically knew the risks, but those risks were being treated as tasks and reminders versus harbingers of impending churn.

There’s a big difference between a to-do item on a list and a blazing fire emoji sitting beside your high-ARR customer.

Call out your risks, track them, and use them as a health score.

This scaleup ended up growing double digits, and that growth started by attacking risk.

A Startup Story

I was at a startup that went from zero to a few hundred customers before selling for a 10X multiple. In the early years, we didn’t have any risk data, so we tracked what we thought would predict success. Later, with enough churn data, we went back to see which of the attributes we were tracking correlated to churn.

Short story? None did.

We had customers with high-risk factors who stayed, and those with low-risk who churned. What the heck?

We went back to the drawing board, mapping out additional product analytics and org attributes to find what correlated. And eventually, we hit on six attributes that predicted both churn and expansion. We then tracked those as a health score and even shared that health score with the customer. In short, we became able to predict risks months before they manifested in churn.

By actively managing risk, you can create a health score that predicts churn.

Keep it Simple

Customer health scores don’t need to be complex, and quite frankly shouldn't be. The more variables and factors you have, the less likely you are to see the major problems.

Think about what is required for your customer to experience value. Is it specific product usage? Are there certain customer cultural issues that cause problems, maybe specific roles and responsibilities you find necessary?

Write these things down. Then ask, are they happening? If not, then you have risks.

Track those attributes, create a scorecard for your top 20 customers, then hold a meeting to discuss the health of your top 20. Now you have a simple start for health scoring that will drive real value.