If, like many, you tend to compulsively create accounts to try new services online, then you are probably used to these “Your password is weak” alerts you can receive at the sign-up phase.
Often, all you want is to quickly register and you don’t really care about the strength of your account security. I mean, the pizza from this new restaurant can’t wait ?
If you worked in the customer or client success industry, however, you would cherish these health signs! It’s highly possible you would do anything to bring a customer from a “Weak” to a “Strong” customer health score.
Did I say customer health score? Let’s talk about the intricacies of its calculation, how useful it could be to you, and above all why you should favor AI use over human-based, manual calculation methods for such scorings.
Spoiler alert: if you are a CSM, an executive or a software business owner, you might want to pay attention.
What is a customer health score?
A Customer health score is the measure of how likely a client is expected to comply with an event you expect, be it customers renewal, upsales, retention, etc.
Why do Success Teams rely so much on customer health scores?
In B2B companies, and especially in SaaS firms, customer success managers (CSMs) can work with a portfolio of 5 to 200+ customers everyday, and it is often really difficult to keep up with the flow of information produced by a CSM’s customers’ feedback.
Still, CSMs need to understand and act on that information in order to execute the mission they are paid for: help clients use their product, and increase their customers profitability, for both the customer and his company.
And what better way for CSMs to do this than by tracking customer health scores for each account, and reaching out proactively to the poorly scored customers? In order to do this, CSMs need to build such scores. This is where it gets interesting.
Customer health scoring: The manual way
Nowadays, most CSMs manually create their own recipe for predicting customer health scores. How does the calculation usually go?
- First, the team picks a few metrics that can be seen as accurate customer health indicators: how much time a customer took to get fully onboarded, the number of customer logins, the number of active users over time, etc.
- Second, for each of the metrics’ values, a rating between 1-10 is given. 10 would mean that this metric’s current value is seen as a sign of excellent health. On the other side, 1 would mean that this metric’s low observed value is a sign of a very poor customer health.
- Third, a basic average or weighted average of the metrics’ scores is computed.
Often, the CSM or Ops team rely on basic features within their traditional Customer Success software to calculate such averages.
This is a hazardous approach. Let us explain you why.
Customer health scoring with AI: the accurate way
Leverage all of your data
If you rely on health scores for your customer success operations, this means you already have large operations. And large customer datasets. So if you manually pick just a few data points to create an average score, you are giving up on 90+% of the rest of the data.
Instead, you should rely on machine learning. Trained AI models can easily ingest all of your customer data, and learn what caused your customers to spend more, to stay longer on your platform, or on the contrary to end their contract sooner than expected.
How do I set up AI to calculate my health scores?
You have two options here.
1 – You build internally your own machine learning classifiers and use machine learning to obtain customer scoring
2- If you don’t have solid enough data science operations internally, and want to move fast, you should hire an expert.
The solutions offered on the market
For this purpose, we recommend you use a customer success software with AI capabilities.
Known candidates providing expert AI solutions are not numerous. Among them, Churnzero or Loopstr.ai stand up for the precision of their solution. From its ideation, Loopstr has developed top-notch AI-based scoring capabilities for SaaS businesses.
Their robust AI models can provide you with clear, actionable insights. Each account gets a customer health score and a churn score.
On top of this, Loopstr provides automatic, proactive alerts for accounts which need to be checked. This is the kind of feature that CSMs can leverage on.
Accounts most likely to churn
| RANK | ACCOUNT | CHURN SCORE | M/M CHANGE |
| #1 | Epic G. | 92 | +2.6% |
| #2 | Acme | 87 | +1.5% |
| #3 | Hooli | 86 | -4.5% |
| #4 | Initech | 78 | +0.1% |
| #5 | Codehow | 74 | -3.2% |
To conclude, customer health scoring should be an important component of any ambitious customer success team, but this should be done appropriately. When possible, always favor the AI-based approach over the human-based one.
Image credit : Starecat.com