The costs associated with developing B2B products and acquiring customers are substantial. In fact, they are often significantly higher than the first year’s revenue produced by selling the product. But it’s relatively easy for customers, whose up-front investment costs are low, to terminate a relationship with the seller if they don’t quickly capture the benefits promised to them during the sales process. Low customer-retention rates can soon lead to poor financial performance and negative word of mouth. That’s why it’s important for companies to understand—and care for—the health of their customer relationships.

Today’s software allows companies to track patterns, identify trends, and analyze with granular detail how customers use their products. This data is collected in systems that support marketing, sales, finance, and operations. Account managers and customer-success managers should also use it to determine the health of their accounts and to proactively manage relationships. But many companies aren’t sure how they can best support employees who are overseeing hundreds or thousands of unique relationships. They’re grappling with how to manage and use all their data to improve customer retention. And they’re wondering how they can assess the health of their customer relationships before it’s too late.

One solution that’s growing in popularity is the use of customer-health scores. According to the Technology & Services Industry Association, about 35% of organizations have adopted advanced health-scoring metrics as part of their customer-success programs. We have decades’ worth of experience studying customer-facing roles in academia and in real-world environments. We have led customer-success operations at several firms, and we regularly advise firms on their customer-success practices. On the basis of this experience, and more than 200 interviews with executives, we have arrived at a model that any B2B brand can use to measure and improve customer health. In this article we illustrate several advanced elements of effective customer-health scoring and introduce the customer-health self-assessment framework—a dynamic, metrics-based approach to monitoring B2B relationships.

The Three Dimensions of Customer Health

Monitoring the quality of B2B companies’ relationships with clients is traditionally the job of the customer-success team. The function was pioneered by Salesforce in the early 2000s. Not long after its IPO, the company realized that despite its rapid rate of customer acquisition, new customers were churning—not renewing their subscriptions—at almost the same rate.

At that time salespeople were the company’s primary point of contact for new customers. But once a salesperson signed a customer (and earned the commission), the incentive to provide further support, insights, and proactive guidance was limited. Salesforce created customer-success teams to fill the gap. These professionals monitor things such as how many customer employees are using the purchased software, how much they’re using it, which features they use, and whether they are fully unlocking its value. Customer-success teams do not engage in sales; rather, they manage retention and work directly with customers to encourage optimal product use. (At most companies, the people handling customer success are compensated largely with salary and receive bonuses for meeting longer-term retention targets and growth targets that are measured by leads passed along to sales.) Variations on the general model pioneered by Salesforce have become standard practice.

Customer health takes customer success a step further by focusing more intensely on data and scoring. Customer health combines classic metrics like Net Promoter Score (NPS) with customer engagement data, product usage data, and value assessments. By integrating a wide range of data sources, firms can assemble a more comprehensive view of customer relationships across many stakeholders (understanding how executives, decision-makers, and end users each value a product, for instance). This broad view enables companies to more accurately predict customer actions such as churn or expansion, and it allows for improved diagnostics that account managers can leverage to keep relationships healthy. It also gives companies flexibility. They can customize the weighting of the data depending on what’s most important for their business and their clients.

Although customer-health scoring can be adapted to a company’s needs, some things are universal. First, customer health is forward-looking. In other words, most scores are designed to be leading indicators of churn risk. Second, customer-health scores represent the value a customer is realizing from a product, such as how often it uses the product, the extent to which it capitalizes on the product’s potential benefits, and how the product is helping it meet business goals.

We find it useful to approach customer-health scoring along three dimensions: customer-relationship quality, product usage, and value realization. A simple template for weighting each dimension follows.

Customer-relationship quality (25%). This dimension is commonly captured using NPS and other traditional measures. In B2B markets, annual or quarterly assessments are periodically indexed within the health score as new survey responses are received. In addition to these intention-based metrics, relationship-quality components include referability, which captures the extent to which a customer is willing to advocate for the seller by taking actions, such as recommending the product to others, posting reviews, or agreeing to participate in publicized case studies.

Product usage (25%). Technology allows companies to objectively track a customer’s product use in a range of ways, from simple assessments of software utilization to automated maintenance tracking of industrial equipment. In addition to actual use, these metrics can include product and feature requests and a count of support tickets, which provide behavioral evidence of positive or negative touchpoints. Usage data is an ideal complement to relationship-quality metrics because they can be updated in real time, include various customer stakeholders, and compensate for lags in survey-based reporting.

Value realization (50%). Increasing the value of your products for customers is the fundamental goal of customer-health scoring. What that means in practice can range greatly depending on the type of solution being sold, but customer-health metrics tend to focus on cost savings, efficiency gains, or achievement of more-granular goals that the customer and the firm jointly set at quarterly business reviews.

Adopting a multidimensional customer-health score will allow you to keep a closer watch on your customer relationships. To find yours, take the average of your scores across all three dimensions. Then use the following rubric to determine the health of your relationships.

Great health: Scores of 70 and higher typically require only proactive maintenance, monitoring, and a routine of goal setting and validation. Customers with these scores should be considered potential advocates of your business.

Good health: Scores between 51 and 70 require active attention; you’ll need to develop more product use, engagement, and value with these clients. They are not at high risk of churn, but you want their health scores to trend upward, not downward.

At-risk health: Scores of 50 and below require greater scrutiny by your leadership team, who should assess the opportunity cost of intensive efforts to save the relationship versus spending time nurturing better-scoring customers. In most cases efforts to save accounts should focus on those with a health score in the 30–50 range. Efforts to save relationships scoring below 30 are typically limited to high-value clients or instances where a clear path forward to better health can be established.

Customer Health at BigCommerce

Now let’s examine how one company has developed and refined its customer-health model. BigCommerce is a SaaS-hosted e-commerce company. (One of us, Neal, leads its customer-success team.) Over the past six years it moved from primarily serving the self-service small-business market to focusing on mid-market and enterprise brands. To effectively manage more complex customers, the position of customer-success manager was added to augment customer onboarding and professional services. BigCommerce’s leadership team invested in the function, and over time it staffed a full customer-success department, which developed and adapted a health-scoring system. Its first iterations involved tracking subjective data inputs in spreadsheets, but that evolved into a more objective, metrics-based system. The team and the scores they produce provide managers and leaders with early and actionable signals of opportunities and risks related to the renewal and growth of customer accounts.

When BigCommerce started measuring customer health, it focused on 2,000 accounts. Within two years retention among those clients increased from approximately 95% to 99%. It expanded the process in 2023 to track an additional 1,000 accounts, and the company has seen a similar improvement among those clients. Now BigCommerce is in the process of expanding health scoring to all active accounts.

BigCommerce’s health-scoring system uses 10 focal metrics that it developed and weighted in-house. As the company has refined its scoring and gained experience with the process, five key learnings have emerged, which most B2B businesses could apply. (1) The scoring system must be calibrated to your data. (2) A well-calibrated scoring system is effective at predicting churn. (3) Once validated, the system helps identify areas for concern and action. (4) There are challenges to overcome with health scoring. (5) A flywheel effect occurs as health scoring matures, leading to improved decisions, opportunities, and direction—even beyond retention prediction. Let’s take a closer look at each lesson.

Lesson 1: Calibrate your health metrics. In the customer-health template we introduced earlier, we suggested that customer-relationship quality should be weighted at 25% of the total health score, product usage at 25%, and value realization at 50%. But BigCommerce is an example of why companies need to determine the weighting that best fits their unique situation.

To establish the initial weighting of the various components, BigCommerce relied on the experience of its sales and customer-success teams. Once key predictors of churn were established, analytic models validated and indicated adjustments needed within the weighting scheme. In short, BigCommerce’s initial pass was an educated guess. Over time the data indicated what adjustments might calibrate the weights.

The customer-relationship-quality score anchors BigCommerce’s overall health score. It incorporates NPS, customers’ willingness to provide positive references for the product, and customer engagement, which is an evaluation of the customer’s involvement in meetings with the company’s customer-success manager. Each of these dimensions contributes 5.88% to BigCommerce’s overall health-score calculation. In addition, the subjective evaluation of the relationship from its customer-success manager carries a weight of 11.76%. Finally, BigCommerce tracks a measure of long-term commitment, which reflects the extent to which a customer has signaled an intention to renew, remained neutral, or indicated apprehension about extending its contract with the company. This metric receives the heaviest weight: 35.29%.

With respect to product usage, BigCommerce tracks how platform feature gaps or bugs are affecting the achievement of customer goals (5.88%) and the overall importance of BigCommerce in its clients’ e-commerce ecosystem (5.88%). To get at that number, it asks questions such as, Is the platform used for all the client’s brands? Is the client’s mobile commerce experience also powered by BigCommerce? Does the client use BigCommerce’s data analytics to measure its own customer health?

On the dimension of value realization, BigCommerce tracks both historical performance and growth (5.88%) and growth potential (5.88%) for customers. To calculate these numbers, the company measures year-over-year changes in gross revenue, total orders, and average order value. It also includes a simple assessment of clients’ payment and invoicing responsiveness, which tends to be a short-term indicator of relationship erosion and is weighted at 11.76%. The financial-health assessment is subjective, and it is typically calculated by customer-success managers on the basis of the likelihood of the customer to add new e-commerce storefronts and the likelihood that the customer will choose BigCommerce for those growth opportunities. Collectively these scores are integrated to calculate customer-health scores that can range from 0 to 100.

Lesson 2: Identify the relationship between health and churn. BigCommerce views the health score as a window into its customer relationships, and it believes that clients with healthy relationships are likely to renew. To assess these relationships ourselves, we analyzed 130 churned customer accounts over a 24-month period and found a strong correlation between health scores and churn. A health score of 30 or below indicates a relationship that is in danger of churning; customers with those scores failed to renew 7% of the time. Customers with health scores of 50 or higher were less than 1% likely to churn.

The average health score for churned accounts at BigCommerce was less than half that typical of accounts that remained with the firm. And increases in a customer’s health score were associated with corresponding increases in customer-revenue generation with the BigCommerce platform and overall revenue for BigCommerce.

Lesson 3: Use the health score to identify areas for concern and action. The adoption of a multidimensional customer-health score enables BigCommerce to track the pulse of its customer relationships and provides account managers with useful insights on how to manage those relationships. For example, if a health score indicates that a customer isn’t using BigCommerce products as often as it once did, the customer-success manager knows to step in to find out why. Or if a customer doesn’t pay its bills as quickly or as often as it once did, a success manager recognizes that as a potential early indicator that sentiment toward BigCommerce has worsened and will investigate.

Lesson 4: Use data to overcome challenges. To succeed, BigCommerce’s customer-success managers needed a tool that would provide quick insights about their most successful customers—one that could sort results by industry and business size. This need led to the creation of BigCommerce’s brand-insights database. Leaders worked in partnership with teams from operations, marketing, and data analytics to build a database that sorts customers into one of three categories: great health, good health, and at risk.

Using cohorts allows account managers to make prescriptive recommendations for customers. They are empowered to speak with specificity from an empirical data set when they offer guidance to their customers. For example, customer-success managers can create cohorts of merchants with excellent or very good health scores by industry and size. To improve lagging health-score issues they can advise how a new product feature that is helping others would benefit a specific client. For BigCommerce this resulted in an uptick of customer-stakeholder engagement, even from stakeholders that historically had kept the company’s team at arm’s length.

Lesson 5: A flywheel effect occurs as health scoring matures. The unintended outcome of using customer-health data is what BigCommerce calls the customer-success (CS) flywheel. Once set in motion, the CS flywheel has continued to expand the value of accurately and consistently measuring customer health.

It was BigCommerce’s marketing department that first identified this dynamic. It studied customer actions to develop a success profile, which uses characteristics of healthy existing relationships to target similar sales prospects. Marketing is now able to direct its efforts toward prospects that have a high likelihood of success with BigCommerce. Moreover, the team can identify and avoid companies that don’t fit the profile, saving effort that could be better directed elsewhere. Using customer-health data, the marketing team is empowered to improve go-to-market strategies such as research on the total addressable market of companies that align with the success profile. It can offer guidance to product development teams regarding new features for adjacent markets that are likely to contain a high number of success-profile prospects. It can also improve the quality of the case studies and other content it creates for customer-advocacy events.

BigCommerce studied the flywheel effect on leads as they moved from marketing to sales. Because customer acquisitions now fit a specific success profile, sales representatives and account managers have more confidence that once a client is onboarded, it will be retained. In the past BigCommerce took on some new clients that weren’t a good fit. The company didn’t want to turn them away but knew that the relationship would struggle over the long term. Today it targets prospects according to their potential alignment with the health predictors of successful clients. It uses prior health-scoring data and analytics to predict customer segments and industries—along with specific characteristics of prospects that are likely to be a good fit with BigCommerce’s products. This process can reduce the reliance on sales teams selling new and yet-to-be-proven products. Instead they can leverage proven success within the existing account base.

The flywheel effect continues during the onboarding of new clients. BigCommerce has enhanced its prescriptive client-onboarding playbooks, which accelerate the rate at which customers begin to realize value. Customer handovers from the onboarding team to account managers are less chaotic, and the need for low-value meetings (to discuss integration, for example) is reduced. Employees can focus more on helping customers succeed and grow on the BigCommerce platform, producing more successful customers in the brand-insights database.

The product and engineering teams, too, benefit from the CS flywheel. When brands that fit the ideal customer profile are onboarded, the product and engineering teams field fewer escalations. That helps keep product plans on track with fewer distractions and less pressure to deliver ad-hoc, unplanned features that usually reduce quality.

Understanding your company’s CS flywheel empowers you to not be overly dependent on new and unproved features. For SaaS companies, the success profile and resulting flywheel effect allow for more-deliberate development of new features according to the likelihood of monetizable and successful adoption—not as a reaction to market whims.

Although you may choose to assess and improve customer-health scores differently than BigCommerce does, its journey and scoring rubric provide an excellent example of how you can use a customer-health score to mitigate unhealthy relationships and improve customer value across the organization.

How to Calculate Your Customer-Health Score

Our interviews uncovered a wide variety of ways that companies score customer health. The German technology company Siemens, for example, follows a simple three-pronged approach to health tracking: It monitors relational health, financial health, and product-use health (in line with our framework). Although Siemens uses proprietary measures to calculate its scores, companies typically measure relational health using the number of individual communications they have with a customer in a given period, NPS, and quarterly business reviews. For financial health, they typically look at the frequency and timeliness of bill payments and any other accounts-receivable patterns that may indicate positive or negative financial health. They measure product-use health by calculating the number of active users for an account and how often everyone uses the software’s key features, among other measures. By focusing on three aggregated metrics, Siemens’s teams gain a comprehensive view of the health of their customer relationships. And they get good visibility into which areas to emphasize during planning meetings.

Axon, maker of Taser brand body cameras and technology solutions for the public-safety sector, takes a different approach: It prioritizes product-usage data when monitoring the health of its accounts. The company’s products themselves report back on usage and charging during and following shifts with public-safety customers. Using this data, account managers can proactively work with customers to identify opportunities to strengthen compliance, optimize operations, and improve maintenance of their solutions. In doing so, they become indispensable partners, which fosters retention and business growth.

We’ve already described another weighting template: assigning 25% weight to your customer-relationship-quality metrics, 25% to product-usage metrics, and 50% to value-realization metrics. Whichever model you decide to follow, we recommend starting the process by leveraging your existing data. Billing and payment data, for example, can help you determine which customers inconsistently meet their financial obligations—a warning sign of churn. Existing survey data like NPS can be used to develop a baseline of attitudes toward the company. Existing data on product usage (such as number of log-ins and depth and breadth of usage) can serve as a representation of customer engagement.

Next, develop a simple dashboard that houses these metrics and effectively allows others in the company to gain high-level insights from them. In many cases an early dashboard can be updated manually by account managers. Once the dashboard is in place, you can begin to add outcome-focused metrics to your survey and product-usage data. Net recurring revenue, contract renewals, and support requests are all examples of these more-advanced metrics.

At this stage you can expand and include cohort analysis to capture average churn rates in a specific category. This data should help address pain points to increase retention and identify revenue-expansion opportunities. Once these broader measures are incorporated into the dashboard, you can develop initial models to predict churn.

By this point you should have confidence in your metrics, and the value of customer-health investments should be established. Resources can be allocated to develop and upgrade the data infrastructure and overall data-science capabilities. That will require at least some investment in business-intelligence tools. You’ll also need to expand your data-warehousing capacity and your data-science teams.

Your health model should now update automatically. Everything you’ve learned about your customers should be used to automate and personalize customer interactions when possible. But it’s important to note that automated personalization should supplement—not replace—actions taken by account managers.

As we’ve seen, the relative importance of the various customer-health inputs will vary by industry and firm. Executive judgment, with input from sales and service executives, is a good place to start. In addition, discussions with customers can help determine the importance of various aspects of the score. Such an integrated approach tends to result in reasonably well calibrated weighting schemes and assists in gaining buy-in across the organization.

Customer-health scoring helps B2B companies build stronger relationships with their clients. Armed with increasingly nimble and accurate health metrics, employees at companies like BigCommerce, Siemens, and Axon are now focused primarily on their customers’ goals. As a result, they increase customer loyalty and improve word-of-mouth recommendations.

As customer-health scoring matures, new uses for the deep insights it provides will spark additional innovations. The customer-success flywheel effect will drive sustainable and positive outcomes for both buyers and sellers.

Author: Bryan Hochstein, Clay Voorhees, Ross Johnson, Neal McCoy and Vijay Mehrotra