Building Customer Health That Executives Can Actually Use

Customer health has become one of the most common ideas in recurring-revenue companies and one of the least consistently useful. Most companies have some version of it. Accounts are green, yellow, or red, and the score may include product usage, support activity, survey data, relationship status, and manual inputs from the account team.

The problem starts when the score becomes more important than the decision it is supposed to support. A supposedly healthy customer churns, another account marked red renews without difficulty, and a large account suddenly turns yellow two weeks before the renewal. Eventually people stop trusting the model and start relying on experience, relationships, and manual inspection again.

At that point, the company may still have a customer health system, but it does not really have customer intelligence.

CUSTOMER HEALTH SHOULD HELP SOMEONE MAKE A DECISION

A lot of customer health programs start with the wrong question. The company asks what should go into the score before deciding what decisions the score is supposed to improve, which usually leads to more inputs, more weighting, and more complexity without much improvement in decision quality.

Executives do not need to know whether an account is 72 percent healthy. They need to know:

• Which revenue is becoming less durable
• What changed
• How early the change became visible
• Whether the risk is recoverable
• What action is required
• Who owns that action
• Which accounts are ready to expand
• Where executive attention should be spent
• How much confidence leadership should have in the renewal forecast

If customer health does not make those questions easier to answer, the organization may be collecting a lot of information without improving the way it manages revenue. The purpose of a health system should not be to classify customers. It should improve the quality and speed of decisions.

HEALTH IS NOT A SINGLE CONDITION

One reason health scores become unreliable is that companies try to compress very different types of risk into one number. A customer can be healthy from a product perspective and unhealthy commercially. Another can have strong executive sponsorship but weak adoption, while a third may be renewing consistently but becoming increasingly expensive to support.

Those are not the same conditions, so they should not be treated as though they are.

I prefer to think of customer health as a combination of signals across several dimensions:

• Product adoption and usage
• Value realization
• Stakeholder strength
• Executive sponsorship
• Support friction
• Implementation quality
• Renewal readiness
• Expansion readiness
• Commercial behavior
• Cost-to-serve

The value comes from understanding how these signals interact. A customer with declining usage and strong executive sponsorship may require a very different response than a customer with strong usage but no budget owner. A customer with healthy adoption and rising support demand may have an economic problem instead of a retention problem.

A single score can hide all of that context.

THE SCORE SHOULD EXPLAIN ITSELF

One of the fastest ways to lose executive confidence in customer health is to present a score that no one can explain. If a CEO asks why a major account is red and the answer requires someone to reverse-engineer the model, the system is too complicated.

Executives do not need the mathematics behind every signal. They do need to understand the reason the account changed.

A useful health view should make the story relatively clear:

• Adoption fell materially over the last 60 days
• The executive sponsor left the company
• Several high-severity support issues remain unresolved
• An implementation milestone is significantly late
• The customer reduced active users
• The business review confirmed measurable ROI
• Another business unit entered evaluation
• A renewal decision is expected next month

That kind of information is useful because it connects the signal to business context. The score can still exist underneath it, but the score should support the decision rather than become the decision.

TREND IS OFTEN MORE IMPORTANT THAN THE ABSOLUTE NUMBER

A customer with 80 percent adoption may appear healthier than one with 60 percent adoption, but direction often matters more than the absolute number. If the first customer was at 95 percent three months ago and the second was at 35 percent, they are telling very different stories.

Executives need visibility into change, not just current state. I care about questions such as:

• Is adoption improving or declining
• Is support demand getting better or worse
• Is stakeholder engagement broadening or narrowing
• Is time-to-value improving across new cohorts
• Are customers consuming more or less of the platform
• Is cost-to-serve rising faster than ARR
• Are risk signals appearing earlier or later than they used to

A static health score can easily hide momentum. A useful health system makes direction visible because recurring revenue rarely deteriorates all at once. It usually weakens through a series of smaller changes that become obvious only after enough of them accumulate.

THE BIGGEST RISK IS OFTEN IN THE SIGNALS THAT DISAGREE

Some of the most useful information appears when the signals do not line up. A customer is green but usage is falling. Another account has strong adoption but no executive engagement. A large customer has a strong CSM relationship while unresolved support issues continue to rise.

Those contradictions are worth investigating because they often mean one of three things:

• The model is missing an important variable
• Different data sources are telling different parts of the story
• The account team knows something the system does not

That is not a reason to abandon customer health. It is a reason to use the disagreement as a prompt for better analysis. The goal should not be to eliminate human judgment, but to make that judgment more informed and more consistent.

When the model and the account team disagree, the result should be a question, not an automatic override.

CUSTOMER HEALTH IS ALSO A DATA QUALITY TEST

Customer health tends to expose how fragmented customer data really is. Product usage may live in one platform, Support in another, CRM may contain commercial history, Services tracks implementation, Customer Success manages stakeholder information, and Finance owns contract and payment data.

Then the company tries to build one health score from all of it.

That is usually the moment when everyone discovers that definitions are inconsistent, data is incomplete, IDs do not match, and different teams use the same words to mean different things. It can be frustrating, but it is also useful because customer health forces the company to decide what it actually believes.

The organization has to define things like:

• What counts as adoption
• What qualifies as an active user
• What makes a support escalation meaningful
• What makes a stakeholder active
• When an implementation becomes late
• What qualifies as renewal risk
• Which behaviors actually correlate with retention or expansion

A strong health system is partly an analytics project, but it is also an operating discipline project. The real value comes from creating shared definitions that different teams can use to make decisions about the same customer.

EXECUTIVES NEED A PORTFOLIO VIEW, NOT THOUSANDS OF ACCOUNT REVIEWS

As the customer base grows, leadership cannot inspect customers one at a time. The health system needs to help executives understand where risk and opportunity are concentrating across the portfolio.

That requires looking beyond individual red accounts and asking broader questions:

• Which customer segments are weakening
• Which products are creating more support friction
• Which cohorts have poor time-to-value
• Where adoption is deteriorating before renewal
• Which segments have strong NRR but weak GRR
• Where expansion is concentrated
• Which high-value accounts have weak stakeholder coverage
• Which customers are becoming more expensive to retain

This is where customer health becomes much more than a Customer Success tool. It becomes an executive operating tool because the CEO can see where revenue quality is changing, the CFO can see where service economics are weakening, Product can identify adoption friction, and Sales can better understand where expansion is credible.

That is far more useful than a dashboard filled with red, yellow, and green circles.

HEALTH SHOULD CONNECT TO RENEWAL FORECASTING

One of the clearest tests of whether customer health is working is whether it improves forecast quality. If a company has a sophisticated health model but renewal surprises remain common, then something important is still missing.

A useful health system should make risk visible earlier than the commercial conversation. It should not wait until the customer says they may leave because, at that point, the system is reporting a fact rather than identifying a risk.

The questions I would use to test the model are straightforward:

• Does it change the forecast earlier
• Does it reduce late surprises
• Does it distinguish real risk from noise
• Does it help teams act before the renewal window becomes critical
• Does it improve executive confidence in the installed-base forecast

In operating environments I have led, connecting product telemetry, adoption signals, support friction, stakeholder engagement, and lifecycle data helped surface renewal risk earlier and improve forecast accuracy above 95 percent. The value was not the score itself. The value was giving the organization more time to act and giving leadership more confidence in the revenue view.

HEALTH SHOULD ALSO SHOW EXPANSION READINESS

Customer health is often built mainly around churn prevention, which makes sense because risk is expensive and visible. A stronger system should also help identify where the relationship is becoming more valuable.

Expansion has signals too:

• Increased product adoption
• Strong value realization
• Additional teams requesting access
• Broader stakeholder engagement
• Higher product dependency
• Successful executive business reviews
• New use cases
• Advocacy activity
• Customer-led referrals
• Increased consumption

That does not mean every healthy customer should receive an upsell call. It means the company should know where the conditions for expansion actually exist.

There is a difference between selling more and expanding because customer value has increased. The second tends to create more durable revenue.

DO NOT BUILD CUSTOMER HEALTH FOR CUSTOMER SUCCESS ALONE

One of the biggest design mistakes is when Customer Success builds the health model, Customer Success reviews it, Customer Success updates it, and Customer Success owns every action. Then the company wonders why the model has little influence outside the function.

Customer health should reflect the entire customer relationship, which means Product, Support, Services, Sales, Finance, and Customer Success all contribute useful signal. It also means the output has to be useful across the business.

For example:

• Product should see adoption and friction patterns
• Support should know which issues create revenue risk
• Finance should see changing renewal confidence
• Sales should understand expansion readiness
• Services should know when implementation issues are affecting value realization
• Executives should know where intervention can actually change the outcome

When customer health becomes shared operating intelligence, it begins to influence behavior across the company instead of remaining a Customer Success dashboard.

AI CAN MAKE CUSTOMER HEALTH BETTER, BUT IT CAN ALSO MAKE IT MORE COMPLICATED

AI and predictive analytics can improve customer health because machines are better than people at processing large numbers of signals consistently. Models can identify adoption decay, changes in support behavior, shifts in engagement, cohort patterns, and combinations of signals that may be difficult to see manually.

That can be valuable, especially at scale, but more sophistication does not automatically create better decisions. If leadership cannot understand why the system believes a customer is at risk, trust will disappear quickly. If the model generates too many alerts, teams will stop paying attention.

AI should make the important signal easier to see, not make the operating model harder to understand. I prefer systems that improve human judgment by helping teams see patterns earlier, process information faster, and focus attention where it has the highest economic value.

MEASURE WHETHER THE HEALTH MODEL WORKS

Customer health should be measured like any other operating system. The question is not whether the dashboard exists. The question is whether the model improves business outcomes.

I would look at:

• How early risk is identified before renewal
• Forecast accuracy
• Churn and contraction among previously healthy accounts
• Recovery rate for identified risk
• Expansion conversion for high-readiness accounts
• Accuracy of health classifications
• Frequency of manual overrides
• Time between signal change and action
• Adoption trends by health category
• Cost-to-serve by health segment

If green customers regularly churn, the model has a problem. If nearly every red customer renews, the model may be too sensitive. If CSMs constantly override the score, the system and field reality are not aligned.

Those are not reasons to abandon the model. They are reasons to improve it.

THE EXECUTIVE TEST IS SIMPLE

There is a fairly simple way to determine whether customer health is useful. Put it in front of the CEO, CFO, CRO, and CCO and ask whether it helps them decide where to spend time, money, and executive attention.

A useful system should help leadership answer:

• Which revenue is weakening
• Why it is weakening
• Which risk is recoverable
• Where expansion is becoming more likely
• Where customer economics are deteriorating
• What changed since the last review
• Where executive intervention may change the outcome

If the system cannot help answer those questions, the problem is probably not that the dashboard needs another metric. The model has not yet been connected tightly enough to the decisions the business needs to make.

That is the standard I would use for customer health. It should not exist simply to tell us whether customers are red, yellow, or green. It should help leadership understand what is changing inside the revenue base, why it matters, and what the business should do next.

© 2026. Pat Ferdig. All rights reserved.

Post-sale revenue control · Customer Success · Support · Services · Renewals · Operations