Inspired by recent conversations with Mary Beth McIntyre and Lee Gartley about engagement scoring trends and advanced behavioral analysis.
Executive Summary (Key Takeaways)
- Many organizations over-rely on RFM and wealth indicators because they are standard, commoditized measures – making them easy to use, calculate, or purchase – while underweighting signals of engagement, commitment, and intent.
- In a data maturity cycle, an engagement score is a fundamental step as you begin to integrate constituent activity not in your CRM – providing directional insight into the depth of a constituent relationship.
- A single score weights and blends signals across multiple forms of engagement, but is not outcome-specific in its ability to predict particular behaviors.
- Even when broken into multiple scores, engagement measures remain largely descriptive rather than predictive; however, it may help inform interactions for one-to-one or one-to-few marketing efforts.
- The real opportunity – and in many ways the superpower – emerges when engagement is used as an input into outcome-specific models and as a guide to determining which data is worth consistently capturing.
The Question Behind Engagement Scoring
Spend enough time in nonprofit analytics, and you begin to notice that certain ideas tend to resurface every few years – not because they have been solved, but because they are trying to answer something fundamentally important that we are still trying to crack. Engagement scoring is one of them. At its core, it reflects an ongoing effort to understand how to prioritize relationships in a growing sea of data – what signals actually matter, and how to bring them together in a way that is both practical and actionable.
This is also where engagement scoring fits within a broader data journey. As outlined in the Nonprofit Data Maturity Model, organizations evolve from relying on fragmented, transactional data toward building a more complete, integrated understanding of their constituents. Engagement scoring is a big first step in that progression, bringing behavioral data into focus and beginning to connect signals that were previously isolated across the organization.
Lately, that question often shows up in a very specific form: “Does your CRM provide a standard engagement score?” It is an understandable instinct. If we can take everything we know about a constituent – giving, event participation, digital interaction – and distill it into a single number, we create something that feels both efficient and actionable. A complex relationship becomes something that can be sorted, segmented, and shared across teams.
And in many ways, it is a significant step forward. For organizations that have historically relied on RFM or externally sourced indicators like wealth, an engagement score introduces something new: first-party behavioral data. It begins to reflect how someone actually interacts with your organization, providing directional insight into the depth of that relationship. Even simple engagement measures can expand prospect pools, surfacing individuals who may not stand out based on traditional giving or wealth metrics but who demonstrate real connection through volunteering, advocacy, events, or digital participation.
That progress is real, but it also raises a more important question: what does that number actually help you do? Most engagement scores are descriptive. They summarize past behavior – who has interacted, how often, and across which channels – which provides useful context but does not directly support forward-looking decisions. The real question is not who has been engaged, but what that engagement suggests will happen next.
The Engagement Score in Concept
In concept, engagement scoring works best where the definition of engagement is consistent. Higher education provides a clear example. Through organizations like CASE (the Council for Advancement and Support of Education), institutions share a common framework for defining alumni engagement – giving, volunteering, attending events – making it possible to build structured, comparable models.
That consistency gives the score meaning. It does not capture everything, but it reflects a shared understanding of what engagement looks like within a defined population and lifecycle.
Where It Gets Complicated
Outside of that environment, the model becomes harder to generalize. Most nonprofits operate across multiple relationship types, where engagement can look very different across audiences and channels.
At the same time, engagement scores are typically built using a familiar structure:
Engagement Score = (Recency x Weight) + (Frequency x Weight) + (Interactions x Weight) + (Years Giving x Weight) + (First-Party Data x Weight)
This formulation is an oversimplification, but it is representative. The goal is consistent: combine different types of activity into a single, comparable measure.
And directionally, it works. A well-constructed engagement score can distinguish between more and less active constituents. It is a significant step forward in describing behavioral data.
From there, many organizations take the next logical step: breaking engagement into multiple scores – giving, events, advocacy, volunteering, and digital participation – to preserve more of the underlying signal. That, too, adds value by creating more visibility across the organization into how different types of engagement behave or to make communications more relevant.
But whether combined or segmented, these scores remain descriptive. They reflect the type of relationship someone has with the organization, but not necessarily what that relationship will lead to next. A highly engaged donor does not automatically become a volunteer, and strong advocacy participation does not guarantee event attendance.
The challenge is not that these scores are wrong. It is that they stop just short of the real question organizations are ultimately trying to answer.
From Engagement to Outcomes
Nonprofits are not trying to measure engagement for its own sake. They are trying to drive specific outcomes – giving, upgrading, retention, advocacy, participation – and each of those behaviors is influenced by different signals.
A single engagement or ranked score blends those signals together. Even multiple scores, while more precise, do not fully resolve the issue. They help describe the relationship, but they do not reliably predict what happens next. This is where engagement shifts from being a measure to being a component of something larger.
Engagement as an Input
At ROI, engagement is not treated as an outcome – it is treated as an input. Within MiLo Intelligence predictive models, engagement and behavioral signals such as advocacy, volunteering, event attendance, and many inferred attributes of engagement are evaluated alongside recency, frequency, interaction type, and channel preference in the context of a specific objective.
This is where machine learning becomes particularly valuable. It allows us to identify correlations between engagement behaviors and specific outcomes in a way that was much more difficult previously. Rather than assuming which signals should matter, we can observe which ones actually do – whether that is donor upgrades, conversion to sustainer giving, or reactivating lapsed donors.
The real opportunity – and in many ways the superpower – emerges when engagement is used as an input into outcome-specific models and as a guide to determining which data is worth consistently capturing because it is useful in predicting behaviors. In that context, engagement is not replaced – it is elevated into a more precise and actionable role.
The models begin to reveal which interactions actually matter. As organizations learn from this analytic process which engagement elements consistently support particular outcomes, they can become more strategic and intentional about collecting those signals over time, thereby improving both the quality and usefulness of their data for stewarding constituents.
Importantly, that does not mean trying to force everything into the CRM. Many of the most valuable engagement signals live across different systems – event platforms, advocacy tools, digital channels, and more. This is where a nonprofit constituent data platform becomes especially useful, tying together disparate data sets, resolving identities across systems, and providing a more complete view of each constituent.
That broader, integrated view makes it possible not only to model behavior more effectively, but to build a data strategy around capturing the signals that actually matter.
From Insight to Action
An engagement score can help you understand the depth of a relationship, and for many organizations, it represents an important step toward a more mature use of data. But understanding relationships is only part of the equation.
The real goal is to understand what those relationships lead to – and which interactions signal future behavior. That is where the shift happens: from summarizing activity to predicting outcomes, and from collecting data to intentionally building it.
Ultimately, the question is not whether your organization has an engagement score. It is whether that score is helping you make better decisions – and helping you build the data you will need to make even better ones over time. Do you know which engagement interactions or first-party data are most impactful, and is it accessible for use across the organization? Are you then intentional about which data to collect more of and how best to use it to predict behavior, make communications relevant, be a catalyst for enhanced stewardship, or build internal prospect pipelines for engaged segments that may not have given yet?
Engagement helps you understand the relationship. Using it to predict what comes next and to drive impactful data collection is what accelerates data maturity – and ultimately, better decisions and program growth.
Are you interested in exploring engagement scoring and how it can be used to predict behavior? Let’s Talk!
