Member Health is a machine learning tool designed to help associations boost member retention by identifying who’s likely to stay and who might be at risk of leaving. By analyzing a wide range of member behaviors and engagement signals, it provides a clear, data-driven picture of renewal likelihood.

At the heart of the system is the Member Health Score a number between 0 and 1 that represents the probability a member will renew. To generate these scores, the machine learning models looks at data points like member engagement, financial history, and long-term behavior, to name a few. The system continuously learns and adapts, updating scores on a regular basis with new information so your insights stay fresh and relevant.

Whether you're planning outreach, launching a retention campaign, or simply prioritizing follow-ups, the Member Health Score helps you do it with confidence and data to back it up.

For associations, member renewals are more than just a metric; they’re the foundation of financial health, future planning, and long-term industry impact. But too often, retention strategies kick in after a member has already walked away.

Member Health flips the script by helping you spot At Risk members early before they churn, so you can take proactive steps to keep them engaged and connected.

Using smart, predictive insights, Member Health gives each member a renewal likelihood status, along with clear explanations about why that status was assigned. These insights help you understand what may be driving disengagement and focus your resources where they’ll make the biggest difference You’ll now have the tools to act with purpose, not guesswork.

The short answer is because it removes the unconscious bias of the staff who develop the scoring matrix and may inadvertently use their own definition of what activities denote "engagement." It also helps remove bias from members who may be answering survey questions targeted to measure engagement, which can be subjective and positively or negatively influenced by their mood in the moment or their feelings about other events happening in their lives that have nothing to do with their member experience. By measuring actual behaviors, we can get beyond those factors and produce more objective and accurate assessments of a member's overall experience and use that data to predict who is in danger of sliding off of the membership roll.

Behind the scenes, Member Health uses a proprietary, powerful machine learning engine that’s designed to learn from your members' data and engagement activities, which, through training and backtesting, will determine whether the data point is a positive or negative influence for each and every member when it comes to predicting their likelihood of renewal.

The model learns patterns in how your members engage, what their history looks like, and how similar members have behaved in the past. It then gives each member a renewal likelihood status, plus a breakdown of the top factors (positive and negative) that influenced that status.

To make sure the model fits the unique patterns of your association (and others), Member Health uses a structure blended of three different model types:

This structure allows Member Health to balance broad insights with personalized accuracy, so you're not getting one-size-fits-all predictions, but tailored results that reflect the diverse nature of your members.

The model is:

Whether you’re managing thousands of members or working with niche segments, the Member Health model delivers insights you can trust—quickly, clearly, and at scale.

You won’t need to worry about micro-managing this data directly. By utilizing your Novi database features in daily operations, Member Health handles the heavy lifting in the background, regularly analyzing and updating information to keep predictions up-to-date and in sync with your member base. This means you get high-confidence renewal statuses and transparent reasoning behind each prediction.

Members with a "Current" or "Grace Period" status will automatically show the following statuses in the Member Health field.

Using the gear icon in your Members list view, you can show or hide the "Member Health" status column.

You will also be able to filter members based on their Member Health status.

(1) Within a member record, you'll see a badge of their current status.

(2) Clicking the status or navigating directly to Engagement > Member Health will give you a Member Health Report unique to that specific member.

This report will give you descriptions of the most impactful Healthy and Warning Signals that are contributing to that status, for that particular member.

You can create groups for easy outreach and give your members some extra attention:

Before Member Health ever makes a prediction about your members, it’s put through a rigorous testing process called backtesting. Think of it like a time machine for data - we roll the clock back a year or two and ask, “What would the model have predicted back then?” Then we compare those predictions to what actually happened.

Because we already know which members renewed and which didn’t, this lets us see how accurate the model really is and fine-tune its predictions based on real-world outcomes while building confidence that the insights you're getting are grounded in reality.

Member Health isn’t a “set it and forget it” system. It’s designed to get better over time with more member data rolling in and with your input playing a key role. These insights get folded back into the model development process to strengthen future predictions.

The result? A smarter, more accurate, and more robust solution that works hand-in-hand with your team, not in isolation. You’re not just using the system - you’re actively improving it.

Have feedback on missing data points or feature enhancements? Send us a message through Intercom (the Blue Bubble) or email us at help@noviams.com.