The logo for reef.ai has a blue turtle on it

Health Scores are Dead! Leading with Outcome-Based Scoring

Brent Grimes • Jul 11, 2024

If you’re relying on a general health score to run your business, you’re leaving substantial retention and growth revenue on the table. In the world of Customer Success, Account Management, Support, and Services, traditional health scores have long been the go-to metric for assessing customer relationships. We’ve all used them. Sometimes successfully, and other times, with mixed results. These conventional metrics often fall short. Not due to lack of intent, but because their broad strokes fail to paint the detailed picture needed to drive specific actions and outcomes.


Why are Traditional Health Scores Dead?

As a seasoned CS leader, I've grappled with the shortcomings of traditional health scores more times than I can count. While they seem promising on paper, they often struggle with:


  • Accuracy: Are "green" customers *truly* never churning? Are "red" customers *always* at risk?
  • Actionability: A general health score of 62 doesn't inherently tell you what actions to take.
  • Impact Measurement: How do we know if our interventions *actually* made a difference?


While traditional health scores may have a time and place, the industry is quickly evolving to ML-powered Outcome-Based Scoring.


What is Reef’s Outcome-Based Scoring?

At Reef, we've pioneered a more nuanced approach that aligns closely with driving specific revenue outcomes: Outcome-Based Scoring (OBS). This method represents a shift from a generalized score to multiple, highly tuned ML-powered scores tailored to specific challenges and opportunities within customer management from churn mitigation to growth.


Instead of relying on one holistic score that attempts to do too much, we're developing targeted scores backed by data science and tied to specific revenue outcomes. It's like having a suite of specialized tools instead of a one-size-fits-all approach.


What are Some Outcome-Based Scores in Action?

Our data scientists are constantly experimenting with different machine learning models to optimize for each revenue outcome. Let's dive deeper into three examples:


  • Big Deal Finder: Reef analyzes your best customer expansions over the recent past to identify common traits among customers who expanded. Based on these learnings, Reef isolates current customers who share these traits and are likely to produce more growth pipeline that converts at a higher rate and closes at a higher rate for a higher sales price. Reef feeds targeted leads to Sales and CSMs and delivers ABM targets to Marketing.
  • Churn Predictor: While many companies have moved towards churn prediction models, Reef automates recommendation delivery, keeps the model up-to-date, and continuously improves the model over time. It provides CSMs and Renewal Managers early warnings about potential churn, allowing for preemptive action.
  • Product Downgrade Blocker: Some companies don’t have a full churn problem, but struggle with partial churn or downgrades of specific products. Reef can optimize a model to predict these downgrades and alert the right team member 6-9 months in advance of renewal, while there’s still time to change the outcome.

What Makes Reef’s Outcome-Based Scoring Approach Better?

Some of the core benefits of Reef's Outcome-Based Scoring approach include:

  1. The models continuously improve through feedback loops and model stacking
  2. There is a measurable impact of actioning Reef's recommendations
  3. Scores are transparent and optimized for your unique business


How can I continuously Improve Customer Scores

Reef aggregates data from multiple sources - CRM, product usage, support tickets, and more - to create a holistic view of your customer landscape and predict outcomes with increasing accuracy.


With each model, we introduce feedback loops to naturally improve the scores over time. Each model score alerts your team with recommendation and actions which, based on input from your teams, feed back into the model. These inputs - was the recommendation accepted, what actions took place, and what was the resulting revenue impact - are all taken into the model to improve the score’s accuracy and effectiveness over time. 


As we introduce additional models, each model improves the data set, becomes more accurate, and improves the performance of other models.


How can I Quantify Individual and Team Impact?

One of the biggest challenges for CS leaders is quantifying the impact of their team. A compelling benefit of Outcome Based Scoring is the ability to track the revenue impact and ROI on any outcome-based score. Reef benchmarks the performance of any target revenue outcome and:


  1. Creates an optimized score
  2. Sends recommendations to improve the outcome
  3. Measures and reports performance versus the benchmark



Impact and ROI can be viewed for playbooks, individual team members, or the team as a whole. This data-driven approach provides a credible case for headcount requests and demonstrates the value of CS to the organization.


Aren’t there other ML-based Scoring Tools on the Market?

There are tools you can plug into your customer data that will analyze product usage or engagement and feed recommendations. These tools can be useful but they’re quite limited when compared to outcome-based scoring for the following reasons:


  • Avoid “black box” scoring. Reef’s outcome-based scores are transparent. When Reef delivers an outcome-based score, you can see what factors are considered in the score, you can view the accuracy and performance of the score, and you can see the correlation of each factor in the score to the target revenue outcome.
  • Avoid “one size fits all” scores. OBS incorporates the unique aspects of your business. Every Reef customer is unique from their industry to the products they sell to seasonality in their business. Reef models are flexible and can incorporate and test unique product features and even custom data points that you want to evaluate for churn or growth correlation.


✨ BONUS: How can I Prepare for the AI-Driven Future?

While outcome-based scoring delivers immediate value through machine learning, it also prepares your data for the next frontier: Enterprise Generative AI. 


By consolidating, aggregating, and normalizing your data plus creating robust history on all Reef-connected data, we're setting the stage for a future where AI can provide profound insights into your customer data in seconds.


Imagine asking, "I have a $1.5M gap next quarter. Who are the best customers to target and the best actions to close that gap?" and getting an immediate, actionable, data-driven answer. That's the future we're building towards by structuring your data into a longitudinal format ready for generative AI use cases.


Getting Started with Outcome-Based Scoring

An advantage of outcome-based scores is that you can start small and grow as you go. Simply identify a revenue outcome you want to improve, start with a single score, measure incremental revenue and ROI, and then add more as you see proof. Every model pays for itself many times over. Reef’s outcome-based scores support over 30 use-cases across multiple GTM functions.

Customer Success Sales Support Services Marketing
Customer Success Product cross-sell Support subscription churn prevention Service subscription churn prevention Product cross-sell ABM targets
Product churn prevention Product upsell Support downgrade prevention Service downgrade prevention Large deal ABM targets
Revenue churn prevention Big deal pipeline Support product churn prevention Service add-on cross-sell Consumption ABM targets
Product downgrade prevention Cross-sell with renewal Support add-on cross-sell Service upgrade with renewal
Revenue downgrade prevention Upsell with renewal Support upgrade with renewal Services experience revenue growth correlation
Competitive churn prevention Consumption prediction Support experience revenue churn correlation Services experience revenue churn correlation
Consumption prediction Support experience revenue growth correlation
Product engagement Support escalation churn correlation

If you're curious to learn more, let's connect! 🪸

Ready to dive in?

Brent Grimes, Founder and CEO of Reef.ai, on the Unchurned Podcast hosted by Update.AI
26 Sep, 2024
Brent Grimes, Founder and CEO at Reef.ai joins the hosts ⁠⁠Kristi Faltorusso⁠⁠, ⁠⁠Jon Johnson⁠⁠ & ⁠⁠Josh Schachter on the Unchurned Podcast⁠⁠. Brent shares insights on outcome-based scoring for customer health and the power of leveraging data for predictive modeling in customer success.
By Brent Grimes 08 Jul, 2024
When you found a startup, you get to worry about a lot of things. One thing that was not on my radar was what to call Reef.ai employees. Now that we’re here, it’s actually a pretty big decision. It creates identity, community, and has a direct influence on company culture. And once you commit, it’s really hard to change course. At Reef.ai, we had some unique challenges. We’re an AI/ML SaaS business based in Hawaii and we wanted something that captured our culture. The most obvious names had alternative meanings that made them interesting but not ideal choices. (See “Reefers” and “Polyps”) Then we came across “corallites”. What are corallites, you ask? If you’ve ever touched a coral reef, you probably noticed that beneath the anemones, polyps, urchins, and other sea life, there is a hard, stone-like framing that gives the reef it’s strength. These are corallites. Corallites are created by polyps and become a safe home where they can escape from danger. We like this analogy. Reef customers create the need for Reef employees and we provide a safe place where their customer data can help them grow. A reef starts small, often a single corallite. Then, other corallites form and multiply. The reef grows bigger and bigger, not just supporting polyps but creating an entire ecosystem of life around the reef. 5 reasons why we like corallites as a name for Reef.ai employees: 🪸 Corallites grow and connect to become the reef. A single corallite is the original building block of a reef. Additional corallites connect and begin to form a reef. They all build and grow together until there is a rich, thriving ecosystem. This is the story of Reef.ai. 🪸 Corallites sounds a lot like correlates. We’re nerdy, we ♥️ data, scoring is a Reef core value prop, and correlation is a big part of ML scoring. It just fits. 🪸 Corallites are strong and adaptable. A reef’s environment is always changing, and a reef has to quickly adapt to thrive and grow. When we hire new corallites, adaptability is one of the key traits we look for. 🪸 Corallites are secure. Corallites protect life in a reef from predators. Connecting, improving, and creating data is at the heart of what Reef.ai does for our customers. We take keeping it secure seriously. 🪸 Corallites are fun. Or, so we’ve heard. :) So now, we’re not just Reef.ai employees, we’re Corallites! Interested in becoming a Corallite? Follow us on LinkedIn and DM us. We're hiring.
By Brent Grimes 08 Nov, 2023
The healthscore serves as a common foundation for customer success strategies, yet it is often misinterpreted and mishandled due to a lack of recognition regarding the challenges of effective customer scoring. Many individuals view a healthscore as a numerical or categorical (red, yellow, or green) assessment resulting from calculations involving various inputs, weighted according to their perceived impact on customer well-being. While this perspective isn't entirely inaccurate, it only scratches the surface. A customer score and its underlying scoring model are akin to the visible tip of an iceberg; there exists a substantial amount of work beneath the surface required to develop an accurate score that can be continually enhanced.
Share by: