When it comes to personal health, age and context are critical to understanding the severity of and treatment for issues. What may be a trivial issue for a teenager could be a huge health risk for an elderly adult; just the same, it's critical to approach your B2B customers in a similar way.
Lack of product usage for a brand new customer still in the set up phase is an entirely different issue - perhaps even a non-issue entirely - that it would be for a customer coming up on the verge of renewal. You need a model that accurately signals as such.
There are many possible stages that customers pass through during their lifecycle; a basic 4-stage model might be the journey from (1) onboarding and set up to (2) adoption to (3) maturity to (4) renewal/expansion. You may even have stages mixed with firmographic cohorts. To most effectively understand the health of your customers throughout their journey, any scores you create should ideally be calculated and weighted distinctly for each stage in your customer lifecycle.
Why one customer health score won’t work for all lifecycle stages
You can’t use the same health scores for all lifecycle stages because a customer’s behavior changes as they mature with your product.
A customer will naturally take different actions in your product when they first sign on then they will after using it for a while. For example, you won’t see high-usage rates from a new customer who’s setting up your product. On the other end, you wouldn’t expect a long-time customer to still be enabling many new integrations.
Not only will customer usage differ from stage to stage, but *what* usage matters also differs. For an onboarding customer, we may not yet care about how much usage they are driving, but more simply that they have turned on and set up certain modules or key integrations, while thec onverse may be true for a mature and established customer.
Calculating health scores uniformly without regard to lifecycle stage - or even other aspects, like industry, size, etc - is a recipe for creating an unreliable, non-trustworthy, and ultimately inactionable set of scores and workflows.
A better way of health scoring
Let's take the 4 basic example lifecycle stages we discussed above:
* Onboarding * Adoption * Maturity * Renewal/Expansion
For each of these lifecycles, consider carefully what signals you should be tracking in order to determine if you need to intervene in some way, either positive or negative.
Let's take a look at two examples - onboarding and renewal - to see some of the drastic differences in what data will matter for your scores.
Onboarding health score calculations should be focused less on usage and much more on set up, whether that be accepted invites for seats, specific integrations or feature flags, or setting up of company-specific milestones.
It may also be worth considering the momentum of usage; this early on, while you don't necessarily care about specific product usage thresholds, you should be looking for constant activity from your key stakeholders invovled in set up. In the first few weeks or months of a customer coming on board, any significant lull or lapse in usage is a huge red flag. As they say, first impressions matter most, and if you can't get early adoption then it'll be an uphill battle to success from the outset.
Here are some common ideas to incorporate into your health score for onboarding:
- Net features touched
- Invites sent and invites accepted
- Integrations enabled
- Most recent log in date
- Subjective milestones (onboarding calls completed, meetings, etc)
A good renewal or expansion health score should incorporate not only objective usage data across the customer, but also some sort of subjective criteria relating to sentiment or key stakeholders.
A common challenge we hear about in B2B is that a customer may look healthy in terms of usage, but nonetheless churn 'unexpectedly', most often due to key stakeholder issues; end users may be happy but strategic stakeholders may not be, key champions may leave, etc. As such, it's critical to also incorporate customer sentiment data, such as NPS or CSAT, as well as internal sentiment data, like a probability of renewal field that the CSM is responsible for.
Lastly, it can also be helpful to include momentum metrics on adoption, which can be challenging to set up without a dedicated Customer Success tool. A great driver for upsell or renewal is oftentimes dramatic or rapid increases in usage of features - less so the literal usage itself. Platforms like Vitally can track and incorporate change over time by percentage into health scores, which gives a new level of insight into customer behavior.
- Change in active user percentage or key product feature usage over the last quarter
- Percentage of inactive users
- CSM "call"/probability field
- Key stakeholder usage
- NPS/CSAT/survey data
Don't just 'set it and forget it'
Your product and your target customers will change as you grow; likewise, so will your understanding of what behaviors best predict churn, renewal, and upsell. The health scores you create today will lose relevance as those changes happen.
The best health scores are reviewed on a regular cadence—quarterly for very young SaaS companies and at least yearly for more mature businesses. With that additional investment, you’ll have developed health scores that are like crystal balls—continually predicting what your customers will do next, so your CS team can take action before it happens.