Five Lessons from Ranking Behavioral Science Frameworks
I recently re-published my #FrameworkTierList ranking of Behavioral Science frameworks on Substack and promised a follow-up article reflecting on what I learned. This is that reflection.
The series taught me a lot—sometimes clarifying and articulating things I already knew, but also often offering new insights through the dialogue it inspired. The fact that I gained so much—and received such great feedback—makes the series a success in my eyes (even if I have a few more grey hairs in my beard from worrying about offending the wrong person).
I started this series to understand why people have strong feelings about frameworks and why certain ones work in some contexts but not others. It was an exploration, and I never pretended it was the final say. I didn’t begin with a fully developed set of dimensions on which to rate the frameworks; instead, I used the series to help uncover what those dimensions even were. And while the series has been largely successful in that endeavor, it also still feels incomplete because I still feel I haven’t yet identified all the relevant dimensions. Sometimes, a framework just works, and I still don’t fully understand why.
But while I don’t have all the answers, I have gained a few insights that I thought worth sharing.
1. What Even Is a Framework, Anyway?
One takeaway from doing this series is that no one seems to agree on the difference between a framework, a model, and a theory. And those who do claim to know often disagree.
These terms frequently overlap, and even philosophers of science debate their definitions. Under my view, a theory explains phenomena, a model represents or predicts them, and a framework provides a structured approach to solving a problem. But in practice, these distinctions blur. For example, is COM-B a framework for organizing behavior change, or a model for how those domains interact? Reasonable people can—and do—disagree. In my view, it is both.
For the purposes of this series, I ranked behavior change frameworks, and for me, what defines a behavior change framework is whether a tool helps map a behavior change problem space, or map problems to solutions. If it does that, it functions as a behavior change framework, regardless of whether it was explicitly designed that way.
This is pretty unfair of me. Not everything I rated was trying to do the same thing. For example, some argued that EAST was more just a heuristic. But many tools are used as a framework even if they weren’t originally intended for that purpose, and so I focused on tools that people rely on to organize behavior change projects, without worrying about the authors’ intent. So, if someone argues, “This wasn’t meant to be a framework,” they’re right—but if people use it as such, I think it is OK to rank it for its ability to do that job.
Ultimately, my rankings focused on how these tools are used in the field, not how they are labeled. What matters is whether a tool helps organize and guide a behavior change project, regardless of its original intent or theoretical background. Which leads us to the second lesson from this series.
2. It’s Completely Unfair to Rank Behavioral Science Frameworks
I knew this from the beginning, and I probably should have emphasized it more. I’ve said it repeatedly, and I’ll say it again: Every scientifically informed framework has a context for which it is S-Tier. Frameworks like EAST and the Behavioral Drivers Model, for example, aren't designed to organize entire projects but rather serve as tools for specific interventions or tactical references, and are excellent for those purposes.
My ranking criteria focused on how well frameworks help organize a whole behavior change project, and not every framework is meant to do that. Frankly, I am very surprised that I didn’t get more pushback on this throughout the series. Now that I am saying it more explicitly, I wonder whether I will. Does everyone just implicitly know what I mean by that without me having to clarify any further? Does everyone agree that it is fair to rank frameworks for their ability to do this when they weren’t even necessarily designed to do that job?
Besides not even trying to do the same thing, frameworks also just sometimes work better in different contexts. As such a better version of this series would have been to identify WHEN to use a framework rather than simply rating them hierarchically. But this is much harder, and I haven’t yet identified all the dimensions that matter. There are certainly some macro-considerations such as the following:
Project Scope: Some frameworks are for comprehensive strategies, others for single interventions.
Client Capabilities: Frameworks need to match the team's skills—more technical frameworks work for knowledgeable teams, while accessible ones help engage clients.
Domain and Industry: Frameworks are often designed for different domains or industries, and may perform best there.
Audience: Some frameworks are for internal use by behavioral scientists, others are client-facing, or designed for collaboration.
Training and Experience: Backgrounds shape how people engage with frameworks, making some more suited to different individuals depending on their training and familiarity.
But I think there are also various smaller considerations around the particular behavior change problem that also impact how well a framework fits the project, and I haven’t been able to clarify those.
So, I haven’t fully cracked the formula for the "right" framework. Sometimes a framework just clicks in one project, but completely fails in another. Is it how the framework is presented? Something inherent to the framework? Is it about the match between the framework and context we are working in?
It is not always clear to me all the dimensions that matter, but doing this series has improved my understanding of why different frameworks work well in different contexts, and why individual researchers or practitioners might prefer one model over another. So that is at least a partial success that gets me a little closer to being able to do a version of this exercise that is more fair to the different strengths of the frameworks.
3. Theory Agnosticism
Frameworks sit at the intersection of theory and practice in Behavioral Science. While they are informed by theories, they are not scientific theories themselves, and this leads to debates over how well a framework represents the underlying science.
I often find myself torn between two perspectives. Frameworks tied closely to a theory can be limiting when that theory has shortcomings or becomes outdated. On the other hand, frameworks detached from theory can feel generic.
For example, Peter Slattery pointed out a frustration he has with COM-B.
Another reason why I have some difficulties with COM-B is because I studied persuasion and motivation in my PhD and know that there are many different ways that these and related concepts are broken down in the literature. To present one way of breaking them down is sort of providing a false structure/theory. Perhaps the most egregious example is social opportunity.
I also know that a lot of theories are critical of dual process models and argue that it is quite an arbitrary way to differentiate the relevant phenomenon. So maybe when you split things up you just create complexity without really providing two clear categories that actually makes sense
I feel similar about the COM-B sub-categories. And relatedly, while I appreciate the Fogg Behavior Model, I don’t fully understand its underlying motivation theory.
So how should one choose frameworks? Should you stick to frameworks that align closely with current science, or (2) accept that all models are wrong, but some are useful?
In my work, I tend to lean toward a third approach: Use frameworks that allow for theory agnosticism. By which I mean frameworks that are flexible enough to allow users to replace or adapt the embedded theory as needed.
No framework can be entirely theory agnostic. Every framework reflects a theory of how change happens, shaping its structure. But some are more adaptable than others. For example, I find the sub-categories of COM-B difficult to work with and intrinsic to the framework. Meanwhile, I don’t have to rely on anyone else’s theory of motivation to use the Fogg Behavior Model, which makes it more adaptive to different scientific theories.
I’m sure some will vehemently disagree with me on this. Ultimately, my preference for theory-agnostic frameworks is precisely that—a preference. If you are new to the field, or are working with people who are, you may prefer more structured theoretical guidance Or perhaps you’ll prefer frameworks that don’t make you feel like a deep understanding of theory is necessary for successful implementation. Either is fine.
I know researchers who want an explicit tie to theory, and that is also fine. I also know people in business who could care less about theoretical rigor and just want something that is intuitive and works. And while that frustrates me, it is probably also fine.
Ultimately, the choice between theory-driven or theory-agnostic frameworks depends on your priorities: adaptability or alignment with current science. For me, flexibility wins, but it’s important to consider the project’s needs and the user’s expertise when making the decision.
4. Layers and depth to create affordances
Most Behavioral Science frameworks operate across multiple layers. They start with a fairly concise surface level—an acronym, a guiding principle, or a simple metaphor—but then have additional layers which add complexity and depth.
An insight from my colleague, Hassan Aleem, led me to think about the utility of these layers in terms of the concept of "affordances." Each layer of a framework affords different ways of thinking, engaging, and communicating. I believe that by thinking through what a framework affords, we can get to the core of what we really care about when evaluating frameworks.
The surface layer of a framework is often the most important because it’s what people rely on most when organizing a behavior change project or communicating with clients. The simplicity of that layer affords easy communication, recall, and structuring.
At this surface layer, many frameworks address similar concepts (e.g., motivation, ability), making it seem like the differences are only cosmetic. But those differences affect how users interact with them in very important ways. A chart can visualize forces at play but may limit flexibility, while phases may organize a journey but become cumbersome if the journey doesn’t match reality.
As we move to deeper layers, more affordances are created, but so are constraints. While deeper layers may offer more precision, they can also be harder to communicate or adapt. The landscape of affordances changes across each layer.
What a framework affords—and what it doesn’t afford—is really core to this whole discussion. The different affordances may explain why frameworks work in some settings but not others, or why one framework may feel intuitive to one person but clunky to another.
To truly understand what makes frameworks succeed or fail, then, is a better understanding of how a framework trades-off different affordances across layers.
For example, Opposing Forces affords flexibility, but lacks an affordance for guiding solutioning, which makes it too vague for clients (in my opinion). 10 Conditions for Change on the other hand is comprehensive with its solutioning, but overwhelming in its comprehensiveness which makes it difficult to adapt to dynamic projects. Both fall short in my ranking because they fail to balance the right affordances across layers.
Meanwhile, the Make It Toolkit, which I ranked highest in my tier list, strikes a better balance. It affords targeted solutioning with its 15 strategies, while also offering flexibility, as not all strategies need to be used in every situation. It’s also easy to communicate, using general terms that non-researchers can understand (e.g., “make it easy”), and the strategies are at a similar level of analysis, making it straight forward to map them to metrics and tactics.
The main point is this: the affordances of Make It match my needs, which is why I ranked it highly. This alignment may be personal, based on the types of problems I work on, or due to my approach to Behavioral Science. It wouldn't surprise me if others disagreed, and found that the affordances and constraints didn’t work for them.
Developing a complete list of framework affordances and their trade-offs is a little too much for an article like this, but some that come top of mind include; flexibility and adaptation, connection between problem and solution, connection to theory, comprehensiveness, clarity in communication, easy recall, actionable solutioning, process guidance.
The best frameworks effectively manage these affordances at each layer. Higher levels afford flexibility, communication, and recall, while the deeper layers afford problem-informed solutioning. This balance between the surface and deep layers is what makes a framework worthy of S-Tier in my ranking.
5. The Human Element
Are frameworks valuable in themselves, or is it the user who gives them power? I see both sides.
On one hand, expertise allows practitioners to work around a framework's shortcomings. An experienced user can adapt a framework to fit within its limitations by leveraging their understanding of the underlying principles. On the other hand, a well-designed framework elevates a novice, providing a structured approach to thinking and problem-solving, effectively training them to act like an expert.
If expertise is the key factor, ranking frameworks might not matter much, as success depends on the implementer. But if frameworks are meant to elevate novices, rankings become more significant.
In reality, both perspectives hold a little but of the truth. However, this doesn’t imply there’s one "best" framework for every situation. Each of us brings our own expertise, training, and personal quirks, meaning frameworks will work differently depending on the individual and the specific problem. What resonates with one person may feel constraining or unintuitive to another.
For this reason, a low ranking should not be taken as a blanket judgment against a framework. Frameworks cater to different ways of thinking, and their utility depends on context and the person using them. I thoroughly believe that forcing uniformity would stifle diversity and impede progress in a field that needs multiple approaches so that it can grow and progress. Even frameworks I use less often may be the right choice for specific projects or individuals.
As such, I don't believe in creating the "best" Behavioral Science framework. Any framework I design would simply add to the list. The ideal approach, in my view, is to customize frameworks to fit the context and user. A customized framework allows experts to combine the strengths of existing models while tailoring them to specific needs. This synergy between the user's expertise and the framework's design often leads to the best outcomes. If I had reviewed “customized frameworks”, it would have received S-Tier.
Ultimately, the best outcomes arise when skilled practitioners use well-designed—or customized—frameworks to help teams map the relationship between problem and solution. Frameworks and expertise work best in tandem. A good framework enhances expertise by providing structure, while expertise adds the adaptability needed for practical application. Without expertise, even the best framework may feel like an empty checklist. Without a framework, even an expert might miss key considerations or waste time reinventing the wheel.
Conclusion
What I’ve taken away from this series is not a definitive meta-framework or a universal answer to which framework is “best.” Instead, I’ve gained a deeper understanding of the factors that make frameworks valuable in mapping out a behavior change problem.
Every framework has a context in which it is S-Tier. There is no one-size-fits-all framework, just as there is no one-size-fits-all intervention. The diversity of frameworks reflects the diversity of problems, people, and processes in Behavioral Science. Instead of seeking a perfect framework or forcing uniformity, we should focus on creating or adapting frameworks to fit the unique needs of each situation.
At their core, frameworks help teams align themselves to a problem by mapping problems to potential solutions, and organizing and clarifying their thinking. They are neither all-powerful or irrelevant. Their structure is important, but also sometimes fairly arbitrary. They work best in the hands of a skilled practitioner, but a skilled practitioner can also make them accessible and effective for someone with little experience.
This series wasn’t about finding the ultimate answer, but about exploring the question. It was thinking in public. I made mistakes, and my conclusions are based on experience, not empirical experiments. I hope it has been a useful exercise for the Behavioral Science community and a helpful starting point for those outside it.
While I may not have uncovered a definitive answer, I’ve gained clarity on the questions we should be asking. If nothing else, I hope this series sparks further reflection, dialogue, and experimentation—because the value of thinking about frameworks is not just about deciding which to use, but also in clarifying what it even means to do Behavioral Science—an important question unto itself.