You ran the training. The room was engaged, the feedback was good, two people stayed afterwards to ask real questions.
You redesigned the process, carefully, with input from the people who use it.
You set the target and said it out loud with a date attached.
Six months later, the old way is still the way.
The instinct at that point is to assume the effort was too small. Usually the effort was fine. What was missing was a second kind of thinking. Most change work pulls one lever hard and calls that a strategy, and one lever has never been enough on its own.
I have spent 25 years designing, delivering and evaluating change in public health, in research, and inside organizations. The conclusion I keep arriving at is unglamorous. No single discipline explains why important initiatives succeed or stall. Four of them, used deliberately together, get close, because each one answers a question the others cannot hear.
Behavior science asks why a person acts
Behavior science is the study of what actually moves an individual: motivation, confidence, habit, identity, social norms, how we weigh a cost today against a benefit years out.
It explains the things that look irrational from the outside. Why someone who agrees with you completely still does nothing. Why more information rarely closes the gap. Why a person can hold two true positions at once, wanting the change and dreading what it costs them, and stay parked there indefinitely until somebody helps them name the second one.
Its most useful discipline is a rule borrowed from clinical practice: diagnosis comes before advice. A leader who tells a reluctant regional manager to just get the audit done, without knowing what is behind the reluctance, is not doing behavior change. They are running a compliance conversation with a friendly face on it.
If one person is stuck, this is almost always where to start. But organizations rarely stall because of one person's psychology.
Implementation science asks what the system is making difficult
Implementation science grew out of medicine and public health, from a specific frustration: we know what works, and it still doesn't happen. It studies how proven practices actually become routine inside real organizations, and what happens to them on the way.
A hospital can hold the best hand hygiene protocol in the world and still watch infection rates climb, because the protocol was never built around how a nurse moves through a shift. A supplier code of conduct is not implemented because legal signed it. It is implemented when the compliant choice is the easy one inside the workflow procurement already has, rather than an extra task bolted onto a full day.
The shift this field makes is small and load-bearing. Instead of asking why this person won't change, it asks what about the system is making this difficult. Readiness, leadership attention, workflow fit, permission to adapt the thing locally, feedback fast enough to correct a rollout rather than grade it a year later.
Used alone, though, implementation science will build a beautifully engineered process nobody wants to use. And used alone, behavior science builds genuine motivation with nowhere to put it.
Systems change asks where you are in the arc
Even with motivated people and a sound plan, large change stalls, because systems protect the status quo by design.
The Berkana Institute's Two Loops model describes this well. One loop is the dominant system, rising, peaking, then declining. A second loop is the emerging one, seeded early at the edges, growing quietly while the old system still looks unassailable. For a long stretch both exist at once, and most of the difficult work happens in that overlap.
Its value is diagnostic. A leader accelerating adoption of a circular packaging model is not doing the same job as a leader trying to hold together a legacy system that has already peaked, and the strategy, the language and the allies all differ depending on which loop you are standing in and which one you are trying to build. Neither behavior science nor implementation science will tell you that. They tell you how to move a person or a process. This tells you whether the thing you are moving still has a future.
Community practice asks who carries it after you leave
The last piece comes from community-based work, where nobody has formal authority over anyone. Communities do not have chief executives. Progress runs on trust, relationships, shared ownership and local adaptation.
I led a healthy cities research network across 64 countries, and the pattern was consistent enough to be boring. Where local leaders were handed tools, decision frameworks and a peer group, the change kept going after the funding ended. Where the expertise sat with an outside expert, results left when that person did.
People change because they see someone like themselves succeeding, because they helped design the thing, and because a small win travelled through a relationship rather than a memo. That is the shift I spend most of my time helping people make: from being the person who does the change work to being the person who builds the capacity for others to do it.
Am I a behavior scientist or an implementation scientist?
Yes.
And no.
My work sits between several disciplines because after 25 years I became convinced that no one of them is complete, and that treating them as competing schools is a mistake. They are not rivals. They are four lenses on the same problem, and which one you need depends entirely on where your particular thing is stuck.
What it looks like when they work together
Take an organization rolling out AI across every function, which is the version of this most people are living through right now.
Behavior science explains why capable employees feel anxious, unconvinced or quietly threatened, and why reassurance from leadership does not touch it. Implementation science redesigns the workflows, trains the managers and builds the feedback loops, so the new way has somewhere to live once someone is willing. Systems thinking recognizes that AI is rearranging roles, power and status rather than introducing a tool, and that a strategy built for the old arrangement will not survive it. Community practice says adoption will spread through trusted peers and local champions long before it spreads through a communications plan.
Run one of those and you get a familiar outcome. Enthusiasm without infrastructure, or infrastructure without enthusiasm, or both without any read on whether the system underneath is even still standing.
The same logic holds for a health system changing a protocol, a company changing what it buys, or a coalition trying to shift who gets a say in a decision. The content differs. The four questions do not.
There is no framework that covers all of this, and there was never going to be. The work is knowing which question you are actually facing this week.
Where to start
Think about the initiative in front of you that has stopped moving. Which of the four is it really asking? Is one person protecting something, or is the workflow making the right thing expensive? Is the system underneath you rising or declining? And if you stepped away next month, who inside the organization would keep it alive?
If it helps to think that through with something, the Real Change Partner is built on this way of working. It works out which question your situation is asking before it offers anything. Three sessions free at realchangepartner.com.