IMB Model (Fisher): S-Tier Behavioral Designer’s Guide

The most striking thing about Jeffrey and Will Fisher’s 1992 paper is what it refuses to do. It does not propose a new theory of attitudes. It does not name a new bias. It does not invent a new psychological construct. It walks into the rubble of three decades of HIV-prevention campaigns that had reached billions of dollars in spend and not moved the needle, and asks a question that should have been embarrassing for the field to need: what actually has to be true inside a person for a behavior to happen?

Their answer was almost insulting in its simplicity. Three things. The person needs specific, action-relevant information. The person needs motivation, both personal and social. The person needs behavioral skills: the actual capacity, plus the felt sense of capacity, to do the thing. Take away any one, and the behavior collapses. Have all three in the right doses, and the behavior follows almost automatically. They called it the Information–Motivation–Behavioral Skills Model, or IMB, and they showed in the same paper why every component-by-itself campaign (the pamphlet, the slogan, the demo video) had been doomed before it shipped.

If you design for human behavior, in any domain, this is the model you keep accidentally rediscovering. Every onboarding flow that prides itself on “education” and watches users churn is missing M. Every motivational keynote that sends people home pumped up but unchanged is missing BS. Every elegant product that nobody adopts is missing I. The IMB Model gives you the diagnostic vocabulary to name what is broken before you try to fix it. The Octalysis Framework gives you the design vocabulary to fix the right thing, in the right layer, with the right Core Drive.

Speed Run Notes

• Behavior requires three inputs in the same person at the same moment: specific information, personal+social motivation, and behavioral skills. Drop any one, the behavior collapses.
• Information is necessary but never sufficient. Most failed health and onboarding campaigns ship more facts and assume motivation and skills will sort themselves out. They don’t.
• Behavioral Skills is the proximal determinant. Information and Motivation route through skills to behavior, so the highest-leverage intervention is almost always skill construction, not persuasion.
• Elicitation research is the load-bearing step everyone skips. Measure what your specific population believes, feels, and can do before you design anything. Generic content produces non-results.
• The Octalysis Crosswalk explodes IMB’s three layers into eight Core-Drive design instruments. The model becomes operational the moment you stop treating Motivation as one bucket.
• Compactness is IMB’s superpower and its trap. The index card makes designers feel done before they’ve done the work. The audit below names the design surface the card hides.

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

What Is the IMB Model?

The Information–Motivation–Behavioral Skills Model is the most quietly influential behavior-change theory most designers have never explicitly read. Jeffrey Fisher and William Fisher proposed it in 1992 inside Psychological Bulletin as a unifying framework for AIDS-risk-reduction interventions that, at the time, were funded heavily, designed mostly around information delivery, and producing almost no measurable behavior change. The Fishers reviewed dozens of campaigns, traced the failure pattern, and named the gap with three boxes and two arrows.

The structural claim is the entire model: behavior is the product of three independent inputs that flow through a single bottleneck. Information and Motivation operate in parallel at the upstream layer. Both then route through Behavioral Skills, which sit downstream as the proximal determinant. Behavior is the output. Remove Information, the skills have nothing to operate on. Remove Motivation, the skills sit idle. Remove Behavioral Skills, neither facts nor feelings translate into action. The Fishers’ insight was not that any of these mattered. Every prior theory said that. The contribution was that they had to be measured and intervened on independently, in the same person, before any behavior should be expected.

The model was born in HIV prevention because that domain made the failure mode unmissable. You could give a college freshman a complete and accurate explanation of HIV transmission, ask them every question on a standardized knowledge test, watch them score perfectly, and then watch their condom-use rate the next month look identical to a control group that got nothing. The information was perfect. The behavior was unchanged. Something else was missing, and “more information” was demonstrably not it.

What made IMB different from the social-cognitive and health-belief models already in circulation was its insistence on specificity. Generic information (“HIV is transmitted through bodily fluids”) produces generic non-response. Action-relevant information (“how to bring up condom use with a partner who has not raised it, three sentences you can actually say”) produces behavior. The same principle applies to motivation: generic motivation (“safer sex is important”) does not move people; population-specific motivation grounded in actual elicited beliefs (“most of your peers are now testing every six months and openly disclose status”) does. And the same applies to behavioral skills: generic competence is meaningless; specific objective skill (“you can roll a condom on in the dark while continuing the conversation”) plus specific perceived self-efficacy (“you believe you can do this without ruining the moment”) is what produces behavior.

Three layers. Specificity at each layer. Elicitation before intervention design. The bottleneck principle that puts skills at the proximal end. That is the entire model on an index card. The reason it has survived almost forty years and migrated out of HIV prevention into medication adherence, vaccination uptake, exercise initiation, diet change, smoking cessation, cancer screening, and an increasing share of digital onboarding flows is that those four claims keep being right wherever someone bothers to test them.

The Three Constructs in Detail

Information — Specific, Action-Relevant, Population-Calibrated

Information in the IMB sense is not what shows up on the back of a leaflet. It is the specific cognitive content a person needs to perform a particular behavior in a particular context. The Fishers separate this construct sharply from two adjacent categories: general knowledge (which correlates poorly with behavior) and scientific accuracy (which is necessary but not sufficient).

The relevant unit is the heuristic. A heuristic in this model is a rule simple enough to deploy under operational pressure. “Check if your refill is due every Sunday night when you pay bills” is a heuristic. “Adherence to antiretroviral therapy requires consistent dosing intervals to maintain therapeutic blood levels” is scientific content that may or may not produce the same behavior. The first compiles into action; the second compiles into a quiz answer. IMB-correct interventions teach the first and treat the second as enabling context, not as the deliverable.

Information also includes the inverse: the specific misinformation and heuristics that block behavior. A campaign to increase HPV vaccination among young adults will run into the false heuristic that “the vaccine is for kids.” A campaign to increase smoke-detector battery checks will run into the false heuristic that “they beep when they need replacing.” Identifying and replacing these false heuristics is information work, and skipping it is the single most common design failure the Fishers documented.

Motivation — Personal × Social, Both Required

The Motivation construct in IMB is deliberately split into two components, and the split matters. Personal motivation is the individual’s evaluative response to the behavior: attitudes, beliefs about outcomes, valuation of those outcomes. Social motivation is the perceived norm: what the relevant reference group is doing, approving of, and visibly rewarding. The model treats these as additive and substitutable up to a point. A person with high personal motivation and zero social motivation can still act, as can the inverse. Interventions that target only one consistently underperform interventions that address both.

This split looks obvious once stated and is constantly violated in practice. Health campaigns lean heavily on personal motivation (“here’s why this matters to you“) and ignore social motivation. Corporate compliance training does the inverse (“everyone in your division has to complete this”) and produces compliance without commitment. Onboarding flows in software products tend to address neither well, substituting feature tours for genuine motivation work.

Within personal motivation, the field has converged on a structure that maps neatly onto the Theory of Planned Behavior: attitude toward the behavior, perceived outcomes, valuation of outcomes. Within social motivation, the parallel structure is descriptive norms (what others do), injunctive norms (what others approve of), and identification with the reference group. A motivation audit that does not separately measure each of these six sub-components is doing intuition, not IMB.

Behavioral Skills — Objective × Perceived, the Real Bottleneck

The Behavioral Skills construct is where most failed interventions actually break, and is the construct least addressed by intuitive design. Skills in IMB are also split into objective skill (can the person actually execute the behavior under realistic conditions?) and perceived self-efficacy (does the person believe they can?). Both are required. Objective competence without felt competence still produces avoidance. Felt competence without actual competence produces failed attempts that crater future motivation.

The skills layer is broader than it sounds. For HIV prevention, behavioral skills include: condom negotiation, condom application, communication with sexual partners about testing history, the social skill of refusing unprotected sex without rupturing the relationship, and the cognitive skill of recognizing in real time that a situation is a high-risk situation. For medication adherence, behavioral skills include: pill-organizer construction, refill management under busy weeks, side-effect interpretation, communication with prescribers about non-adherence, and the metacognitive skill of noticing your own adherence drift before it becomes a missed week. None of these are facts. None of them are feelings. All of them are skills that have to be built through guided practice with feedback, and almost none of them are addressed by an information leaflet or a motivational video.

This is the construct that the Octalysis Framework’s Core Drive 2 (CD2): Development & Accomplishment targets directly, and is the layer where gamification techniques produce the most leverage. Practice loops, scaffolding, deliberate-difficulty progression, feedback timing, and self-efficacy reinforcement are all skill-construction instruments. The reason gamified onboarding works when traditional onboarding does not is that gamification accidentally addresses the BS layer that information-and-motivation-only designs leave empty.

Why Elicitation Research Is the Whole Game

The Fishers’ single most important methodological insistence, and the one most ignored in applied IMB work, is the requirement of elicitation research before intervention design. Elicitation is the systematic measurement of what your specific target population currently believes, feels, and can do across all three layers, conducted with that population, before you write a single line of intervention content.

Elicitation matters because IMB is content-empty by design. The model tells you the three layers must be addressed; it does not tell you what to put in each layer. The “what” is population-specific. The information gaps that block condom use among college freshmen are different from the gaps that block condom use among recently divorced fifty-year-olds. The motivational levers that move medication adherence among newly diagnosed Type 1 diabetics are different from those that move adherence among ten-year-veteran patients. Skipping elicitation means populating the model with generic content that addresses imagined gaps rather than measured gaps, and produces interventions that score well in expert review and produce no behavior change in deployment.

Proper elicitation has a fixed structure:

1. Sample a substantial slice of the target population.
2. Measure Information with a behavior-specific knowledge test, paying attention to the false heuristics that block correct behavior, not just the true heuristics that enable it.
3. Measure Personal Motivation with attitude scales that disaggregate outcome expectancies from outcome valuations.
4. Measure Social Motivation with descriptive-norm and injunctive-norm scales calibrated to the population’s actual reference groups.
5. Measure Behavioral Skills with both objective performance under simulated conditions and perceived self-efficacy on standardized scales.
6. Analyze which gaps in the data predict the target behavior — those are your intervention priorities.

The output of elicitation is not a report. It is an intervention specification: which information to deliver to close which gaps, which motivational beliefs to address, which skills to build. The specification is the entire reason elicitation exists. Skipping this step is not a corner cut. It is a different (and worse) intervention.

What Fisher & Fisher Got Right

The strongest case for IMB is the empirical track record. Fisher, Fisher, Williams, and Malloy’s 1994 first major test produced significant behavior change in college populations on condom use, with effect sizes that consistently outperformed information-only and motivation-only controls. The model has since been applied across HIV prevention populations on six continents, antiretroviral adherence (where the IMB-AT extension by Fisher, Amico, Fisher, and Harman in 2008 is the leading framework), vaccination uptake, smoking cessation, exercise initiation, diet change, cancer screening, oral health behaviors, and an increasing share of digital health-app onboarding. Meta-analyses consistently find IMB-based interventions outperform single-construct comparators.

The structural insight that information is necessary but never sufficient is now so widely accepted across applied behavioral fields that it functions as conventional wisdom. In 1992 it was contested, and IMB is one of the studies most responsible for the shift. The corollary insight that motivation alone does not bridge to behavior without skills is still under-internalized in industry, but the academic case is closed.

The decision to treat Behavioral Skills as the proximal determinant, sitting between the upstream I and M constructs and the downstream behavior, was a quiet structural innovation. It explains why purely informational interventions fail (the bottleneck is empty) and why purely motivational interventions fail (the bottleneck is empty). It predicts where to allocate intervention budget (the skills layer, when in doubt) and why gamified products outperform brochure-equivalents (gamification builds skills almost as a side effect).

The model also got methodological humility right. The Fishers consistently positioned IMB as a framework to be specified per population, not as a finished product. They built elicitation into the protocol, made the model content-empty deliberately, and resisted the temptation to publish a one-size-fits-all curriculum that the field could lazily reuse. The literature is the better for this restraint.

Where IMB Falls Apart

The Specification Burden Is Crushing

The honest critique of IMB is that doing it correctly is expensive. Proper elicitation research costs real money, takes months, and requires methodological training most product teams do not have. Skipping it is what most applied work does, which is why so many self-described IMB interventions produce disappointing results. They use the model’s vocabulary without doing its homework. The Fishers warned against this in print and have been warning against it for thirty years. The warning has not been enough.

The practical consequence is a literature with two populations of studies: small academic interventions that did proper elicitation and produced solid effects, and large applied deployments that skipped elicitation and produced null results. Critics of the model often treat the second population as evidence against IMB. Defenders point out, correctly, that the second population is evidence against running IMB without elicitation, which is a different claim.

The Three-Construct Compression Hides Real Variation

Motivation in particular hides a lot of structure. Personal-versus-social is a useful first cut, but within personal motivation alone there are at least eight distinct motivational systems, what the Octalysis Framework names as the eight Core Drives, and treating them as one undifferentiated bucket loses design resolution. A campaign that activates Core Drive 8 (CD8): Loss & Avoidance (fear of contracting HIV) is doing different psychological work from one that activates Core Drive 1 (CD1): Epic Meaning & Calling (joining the movement to end the epidemic), and the two should be designed, measured, and audited differently. IMB’s compression is parsimonious but lossy.

The same critique applies to skills. Cognitive skill, social skill, motor skill, and metacognitive skill are very different design problems. Lumping them into one BS construct is a useful simplification at the model level but a dangerous oversimplification at the intervention level.

The Static Snapshot Underweights Dynamics

IMB models a moment. It tells you what must be true inside a person for a behavior to happen. It does not tell you how I, M, and BS evolve over time, how they interact dynamically with each other (does building skill increase motivation? does sustained motivation degrade if information becomes stale?), or how to design for behaviors that require multi-month or multi-year persistence. The model’s siblings, particularly the Health Action Process Approach (HAPA) and the Transtheoretical Model, handle these temporal dynamics better, which is one of the reasons many modern interventions use IMB to diagnose the gap and HAPA or TTM to sequence the response.

What’s Really Happening Inside the Brain

The three IMB constructs map onto distinct neural systems with enough clarity that the model’s structural claims have neurobiological support, not just behavioral support.

Information operates primarily in the dorsolateral prefrontal cortex (dlPFC), the seat of explicit, language-mediated knowledge representation and rule-based reasoning. When you teach a heuristic, you are encoding a rule the dlPFC can later retrieve and apply. The reason information alone fails so often is that the dlPFC is computationally expensive and gets bypassed under cognitive load, time pressure, or emotional arousal: exactly the conditions under which most target behaviors actually occur.

Personal motivation routes through the ventromedial prefrontal cortex (vmPFC) and ventral striatum, the valuation system that computes expected utility. The vmPFC integrates outcome expectancies and outcome valuations into a single value signal that drives behavior selection. Social motivation overlaps substantially with the personal-motivation circuit but adds the temporoparietal junction and medial prefrontal cortex, the social-cognition network responsible for representing others’ beliefs and norms. The two motivation pathways converge at the striatum, which is why IMB can treat them as additive at the model level.

Behavioral skills sit in the basal ganglia and motor cortex when they have been compiled into procedural memory through practice. This is the most important neurobiological fact about IMB: skills offloaded to procedural memory operate under conditions where information-and-motivation processing fails. A condom-application skill compiled into procedural memory works in the dark, under emotional arousal, after a few drinks; an explicit “remember to apply the condom” rule does not. The reason the skills layer is the proximal determinant of behavior is that it operates under the conditions where behavior actually happens, while the upstream I and M layers operate under conditions of calm reflection.

This neurobiological structure also explains why perceived self-efficacy is a distinct construct from objective skill. Perceived self-efficacy is itself a representation in the medial prefrontal cortex: a learned belief about one’s own capacity. Bandura’s social cognitive theory is built around this distinction, and the IMB Model inherits it. You can have objective skill in procedural memory and a degraded self-efficacy belief in semantic memory, and the latter will gate behavior even though the former could execute it. The intervention implication is that skill construction must be accompanied by self-efficacy construction; the two are not the same operation.

IMB vs Other Behavior-Change Theories

vs the Health Belief Model

The Health Belief Model (Rosenstock 1974) sits adjacent to IMB and reaches some overlapping conclusions. HBM weights perceived susceptibility and perceived severity (the threat side) against perceived benefits minus perceived barriers (the coping side), with cues to action as the trigger. IMB strips out the threat-appraisal machinery and replaces it with a clean three-construct architecture that puts skills at the proximal end. The two models agree that information is necessary but not sufficient, and disagree on what to add next. HBM adds threat structure; IMB adds skills as the bottleneck. Where the behavior in question is well-understood as a defense against a threat (vaccination, screening), HBM works well. Where the behavior is a positive skill (adherence, exercise, condom use), IMB outperforms.

vs the Theory of Planned Behavior

The Theory of Planned Behavior (Ajzen 1991) shares with IMB the recognition that attitudes plus norms drive intention, and that perceived behavioral control is its own construct. TPB stops at intention as the proximal predictor; IMB pushes through to behavioral skills as the proximal determinant, treating intention as motivation and adding the explicit skills layer between intention and behavior. This is a meaningful difference. The well-documented “intention-behavior gap” that TPB struggles to explain is, in IMB terms, simply the BS layer doing its work. High intention without compiled skill produces unfulfilled intention. IMB does not have the same gap because skills are inside the model, not assumed downstream of it.

vs Social Cognitive Theory

Bandura’s Social Cognitive Theory is IMB’s intellectual parent on the skills side. Self-efficacy as a construct comes from Bandura, and IMB’s Behavioral Skills layer is Bandura’s behavioral capability plus self-efficacy, formalized into a measurable component of a tighter model. SCT is broader and more ambitious; IMB is narrower and more deployable. Where SCT can describe almost any behavior, IMB can be operationalized into an intervention specification with measurable inputs and outputs. The two are complementary: SCT for theoretical breadth, IMB for applied compression.

vs the COM-B / Behavior Change Wheel

COM-B (Michie, Atkins, and West 2011) is the most direct modern competitor. Capability, Opportunity, and Motivation share family resemblance with Information, Behavioral Skills, and Motivation, with Opportunity (physical and social environment) added explicitly. COM-B is broader because it integrates the environment as a first-class construct; IMB is older and more parsimonious. Modern applied work increasingly uses COM-B for diagnosis and intervention mapping at the population level, and IMB for individual-level intervention design where Opportunity can be held constant. They are not in conflict; they operate at different scales.

IMB in the Real World

Antiretroviral Adherence — The IMB-AT Extension

The cleanest applied case is medication adherence to HIV antiretroviral therapy. Fisher, Amico, Fisher, and Harman’s 2008 paper extended IMB to the adherence problem and developed measurement instruments calibrated to that population. The adherence-specific Information construct includes: dosing schedules, food and timing requirements, drug-interaction heuristics, resistance dynamics if doses are skipped, and the specific consequences of even short adherence interruptions. The adherence-specific Motivation construct includes attitudes toward the regimen and its side effects, and norms among the patient’s reference group (other HIV-positive patients, partners, family). The adherence-specific Behavioral Skills construct includes pill-organizer construction, refill management, side-effect interpretation and management, communication with prescribers about adherence difficulties, and the self-efficacy belief that one can sustain adherence under busy or stressful weeks. Interventions built on this specification consistently outperform general adherence-counseling controls in randomized trials. The same architecture has since been ported to adherence problems in oncology, cardiovascular medications, and contraceptive use.

Sesame Workshop’s Pediatric Health-Habit Interventions

Outside the clinical literature, Sesame Workshop’s pediatric health-habit programs are textbook IMB applications. Children face habit-formation problems — around nutrition, movement, sleep, hand-washing, dental hygiene, and chronic-disease self-management — that punish information-only and motivation-only interventions equally hard. Sesame’s content stack pairs heuristic information (delivered through characters), age-appropriate motivation (peers practicing the behavior, family support norms), and explicit skill construction (rehearsed routines, scaffolded practice with feedback). Sesame’s published evaluations across cardiovascular-risk-reduction and healthy-habits programs show measurable behavior improvements that information-only educational materials had not produced in the same populations.

Duolingo as Accidental IMB

Duolingo did not set out to build an IMB intervention, but its architecture is a near-perfect implementation. The Information layer is the language content itself, delivered in small, specific, action-relevant units (the heuristic of conjugation patterns, not a grammar textbook). The Motivation layer is split between personal (your streak, your XP, your level) and social (your league, your friends’ progress, your placement in leaderboards). The Behavioral Skills layer is built through deliberate practice with immediate feedback, scaffolded difficulty, and explicit self-efficacy construction (the green checkmarks, the “you got 8 in a row” celebrations). The reason Duolingo dramatically outperforms textbook language-learning is that textbooks are I-only interventions; Duolingo is a full IMB stack with strong M and BS layers that traditional materials never built.

Smoking Cessation Apps

The smoking-cessation app market is a useful natural experiment in IMB completeness. The early generation of apps was almost entirely I-and-M: facts about smoking harm, motivational reminders, progress counters. They produced modest effect sizes. The second generation added skill modules: urge-surfing exercises, trigger-identification practice, refusal-language scripts, in-app rehearsal of high-risk situations. Effect sizes improved meaningfully. The pattern is the same shape as the original 1990s HIV-prevention shift: filling in the BS layer turns a marginal intervention into a meaningfully effective one, with no change to the upstream content quality.

The Elephant in the Room

Every applied IMB practitioner eventually runs into the question the model itself cannot answer: whose behavior, in service of whose goal?

IMB is morally neutral technology. The same architecture that increases ART adherence increases credit-card sign-ups. The same skill-construction loops that build condom-use competence build slot-machine fluency. The same population-specific elicitation that designs an effective vaccination campaign designs an effective vape-flavor campaign for sixteen-year-olds. The Fishers’ work has been ported, with no structural modification required, into commercial domains the original authors would almost certainly not have endorsed.

This is the bind every behavioral designer inherits and the reason this Behavioral Framework Library exists in the form it does. There is no clean answer inside IMB. The model is a tool. Tools serve the goals of the person holding them. The defense against weaponized IMB is not a better model. It is the same defense that applies to every behavior-change technology in the library: the publicity test from the Libertarian Paternalism pillar. Would you be willing to publish, on the front page of your newsletter, the elicitation findings, the targeted information, the motivational messaging, and the skill-construction loops you are about to ship? If yes, your intervention is probably defensible. If no, you are using IMB to build something you would not defend in public, which is the classical signature of an intervention you should not be building.

The second bind is subtler. IMB-correct interventions are more effective than IMB-incorrect interventions. This means an unethical IMB-correct intervention does more damage than an unethical naive intervention. The skill of doing IMB well is the same skill, mechanically, whether the goal is admirable or predatory. The field’s ethical anchor cannot be the model. It has to be the designer.

How to Apply IMB with the Octalysis Framework

IMB names three layers. Octalysis names eight Core Drives. The Crosswalk between them is what turns IMB from a diagnostic vocabulary into a design instrument. Each Core Drive has a distinct M-vector (how it motivates) and a distinct BS-vector (which skills it requires and reinforces). When you stop treating Motivation as one undifferentiated bucket and treat it as eight motivational systems, IMB’s design surface expands by roughly an order of magnitude.

Throughout this section, Core Drive N and CDN refer to the same thing — the eight motivational systems Octalysis numbers from 1 to 8. Each section header spells out the full name; inline references use the CDN shorthand for readability.

Core Drive 1 (CD1): Epic Meaning & Calling — The Cause Layer

M-vector: the person is part of something larger than themselves. The target behavior contributes to a cause they identify with. BS-vector: the skill is identity-consistent behavior, acting in alignment with the cause when no one is watching. Information layer: stories of the cause’s progress, named contributions from people like the user, transparent reporting on impact. Used well, CD1 sustains motivation under conditions where personal incentive has fallen off; used poorly, it produces the “saving the world” framing that licenses ethical drift documented in the Behavioral Ethics pillar.

Core Drive 2 (CD2): Development & Accomplishment — The Skill Engine

This is the canonical CD for the BS layer. M-vector: visible progress, demonstrated mastery, the satisfaction of capability growth. BS-vector: this is the BS layer: the practice loops, the scaffolded difficulty, the feedback timing that builds objective skill and self-efficacy simultaneously. Information layer: heuristics packaged as progressive techniques; “level one teaches you to do X, level two teaches you to do Y under condition Z.” Every well-designed IMB intervention will lean heavily on CD2 because it is the design pattern for skill construction.

Core Drive 3 (CD3): Empowerment of Creativity & Feedback — The Adaptation Engine

M-vector: the user can shape the behavior to their own context. BS-vector: meta-skills, the capacity to adapt the canonical behavior to one’s own life rather than mechanically copy it. Information layer: design rules rather than scripts; “here is the principle, here are three ways it shows up in practice, choose the one that fits your situation.” CD3-heavy designs trade some adherence consistency for deeper internalization, and are the right choice for behaviors that need to persist across years of life-context change.

Core Drive 4 (CD4): Ownership & Possession — The Identity-Endowment Engine

M-vector: the behavior is part of who the person is, not what they do. BS-vector: identity-consistency skills, the cognitive moves that recognize “I am someone who does this.” Information layer: language that supports identity construction (“you are a non-smoker”) rather than process compliance (“you are quitting smoking”). The classical James Clear “identity-based habits” pattern is a CD4 implementation; the BS layer here is the metacognitive skill of catching identity-inconsistent action before it happens.

Core Drive 5 (CD5): Social Influence & Relatedness — The Norm Engine

This is the canonical CD for the social-motivation half of IMB’s M construct. M-vector: descriptive norms (what peers do), injunctive norms (what peers approve of), and relatedness (the felt warmth of the reference group). BS-vector: social skills, the conversational moves that align personal behavior with public norm, and the disclosure moves that make norm-following visible to others. Information layer: peer-comparison data, named reference groups, accurate norm reporting. CD5 is the most under-used Core Drive in clinical IMB applications and the most over-used in commercial ones; the asymmetry is worth noting.

Core Drive 6 (CD6): Scarcity & Impatience — The Window Engine

M-vector: the behavior must be performed inside a bounded window, after which it becomes unavailable or substantially costlier. BS-vector: timing skills, the meta-skill of recognizing when the window is open and acting inside it without delay. Information layer: heuristics that make the window legible (“if you’ve had a positive test, the seventy-two-hour window for prophylaxis starts now”). CD6 is the Core Drive most prone to ethical abuse because manufactured urgency is the dark-pattern industry’s default tool; honest CD6 design is honest about whether the window is real.

Core Drive 7 (CD7): Unpredictability & Curiosity — The Attention Engine

M-vector: curiosity about what will happen next, what one will learn, who one will become. BS-vector: exploration skills, the social and cognitive moves that let one investigate uncertainty without becoming overwhelmed. Information layer: information that creates a productive knowledge gap rather than dumping a full answer. CD7 is most useful at the upstream end of an intervention sequence (pulling people into engagement) and least useful at the steady-state end, where consistency matters more than novelty.

Core Drive 8 (CD8): Loss & Avoidance — The Warning Engine

M-vector: avoidance of a specific, personally-relevant loss. BS-vector: threat-recognition skills, the cognitive move that classifies a current situation as a risk situation in time to act. Information layer: heuristics that name specific risk patterns (“if your partner has not had a test in the last six months, this is a risk situation”). CD8 is the most psychologically expensive Core Drive to sustain and the most prone to backfiring through reactance or fatalism if the perceived threat exceeds perceived coping capacity, the same failure pattern named by the Extended Parallel Process Model. Use sparingly, calibrate to perceived efficacy, and never as the sole motivational driver.

The Six-Step IMB × Octalysis Audit

The Crosswalk above gives you the vocabulary. The audit below gives you the procedure. Run this on any intervention before shipping.

Step 1 — Elicit before you design. Sample the target population. Measure current I, M, and BS levels. Identify the specific gaps that predict the target behavior. If you skip this step, every later step is theater. Most failed IMB deployments are failures at Step 1 dressed up as failures elsewhere.

Step 2 — Map the gaps to Core Drives. For each measured gap, name which of the eight Core Drives most cleanly addresses it. A gap in social motivation is a CD5 gap. A gap in skills practice is a CD2 gap. A gap in identity-consistency motivation is a CD4 gap. The mapping is rarely one-to-one; expect each gap to suggest two or three Core Drives.

Step 3 — Design at the Core-Drive level, not the IMB-construct level. Build a specific design element for each Core Drive you decided to recruit. For CD5, the design element might be a peer-comparison dashboard with accurate descriptive-norm data. For CD2, it might be a scaffolded practice sequence with feedback timing tuned to the skill being built. Each design element should specify its I, M, and BS contributions explicitly.

Step 4 — Sequence the Core Drives across the user journey. CD7 and CD1 pull people in. CD2 and CD3 hold them through the skill-construction phase. CD4 and CD5 support steady-state behavior. CD6 and CD8 belong at decision moments where urgency or risk are real. The Octalysis four-phase architecture (Discovery, Onboarding, Scaffolding, Endgame) is the canonical sequencing structure; map your design elements to the appropriate phase.

Step 5 — Build deliberate self-efficacy reinforcement into the skills layer. The BS construct is not satisfied by skill-building alone; perceived self-efficacy must be addressed explicitly. Every practice loop should produce visible evidence of growing competence. Every scaffold should be removable as the user proves they no longer need it. Every successful action should be recognized in a form the user can later retrieve when self-efficacy is wavering.

Step 6 — Run the publicity test. Before shipping, write out (in plain language) the elicitation findings, the targeted information, the motivational levers, and the skill-construction loops. Ask whether you would publish the full design specification under your own name. If you would not, identify which element fails the test and either redesign or kill it. This step exists because IMB-correct interventions are more effective than naive ones, which makes the ethical stakes proportionally higher.

Practical Steps for Designers and Operators

1. Audit your last intervention for missing layers. Take the most recent campaign, onboarding flow, or behavior-change product you shipped. For each of I, M, and BS, list the specific elements that addressed it. If any layer is empty or near-empty, you have just identified the most likely reason it underperformed.
2. Run a one-week elicitation sprint before your next intervention design. Even a small qualitative sprint (fifteen target-population interviews structured around the IMB constructs) will catch the gaps your team has been imagining wrong. The asymmetry is enormous: a week of elicitation can save a quarter of misaimed intervention work.
3. Stop measuring information delivery as a primary success metric. “Users read X percent of the onboarding content” is not a behavior-change metric; it is an I-layer compliance metric. Replace it with measures of actual behavior performance, ideally under realistic conditions.
4. Build a skill-construction module into every intervention. If your design contains no deliberate practice loop, no scaffolded difficulty, and no explicit self-efficacy reinforcement, you have shipped an I-and-M-only intervention and should expect the matching results. The BS layer is rarely free; budget for it deliberately.
5. Use the Octalysis Crosswalk to disaggregate the M construct. Stop asking “is this user motivated?” Start asking “which Core Drives is this user motivated by, in what mix?” The eight-construct view will identify motivational levers your three-construct view was averaging out.
6. Apply the publicity test to every shipped intervention. Make it a literal team ritual. The cost is fifteen minutes; the protection is real.
7. Re-elicit on a calendar. Populations drift. The information gaps, motivational levers, and skill levels that defined the design two years ago are not the same today. Build a periodic re-elicitation cadence into the intervention’s operating model.

IMB Was the Beginning, Not the End

The Fishers built a model that has aged well precisely because it stayed small. Three constructs, two arrows, one bottleneck. The compactness has carried it from HIV prevention into a substantial fraction of modern behavioral-design practice. The compactness is also why so many applications of the model produce disappointing results. The index card hides the work that has to happen behind it.

The Octalysis Crosswalk does not replace IMB. It populates it. The three layers stay; the Motivation layer gets eight Core Drives instead of one bucket; the Behavioral Skills layer gets the Core-Drive-specific skill instruments that build the actual capacities behavior requires. The audit gives you the procedure. The publicity test gives you the ethical anchor. Together, these turn IMB from a diagnostic vocabulary into something a working designer can ship with.

If you only take one thing from this guide, take this: the bottleneck is almost always behavioral skills. When you find yourself reaching for more information or more motivation as a fix, ask first whether the skill exists in the population you are designing for. Most of the time, it does not, and most of the time, that is the actual problem.

If you want to go further: the Octalysis Framework pillar walks through the eight Core Drives in full, the Behavioral Framework Library assembles every model IMB sits alongside, and Actionable Gamification goes deep on the Core Drive design instruments referenced throughout this guide.

Frequently Asked Questions

What is the IMB Model in one sentence?

The Information–Motivation–Behavioral Skills Model is a behavior-change framework that says behavior requires three inputs in the same person at the same time (specific information, personal-plus-social motivation, and objective-plus-perceived skill), with Behavioral Skills as the proximal determinant.

Who developed it and when?

Jeffrey D. Fisher and William A. Fisher introduced the model in 1992 in Psychological Bulletin 111(3), 455–474, as a unifying framework for AIDS-risk-reduction interventions. The 2008 extension to adherence (IMB-AT) by Fisher, Amico, Fisher, and Harman is the most-cited follow-up.

How is IMB different from the Health Belief Model?

HBM weights perceived threat (susceptibility × severity) against perceived coping (benefits minus barriers) and treats cues to action as triggers. IMB strips out the threat-appraisal machinery and puts Behavioral Skills at the proximal end as the bottleneck. HBM is stronger for vaccination and screening behaviors; IMB is stronger for skilled behaviors like adherence and condom use.

How is IMB different from the Theory of Planned Behavior?

TPB stops at intention as the proximal predictor; IMB adds a Behavioral Skills layer between intention and behavior. The well-documented intention–behavior gap that TPB struggles to explain is, in IMB terms, simply the BS layer doing its work. Skills sit inside IMB, not downstream of it.

Why is elicitation research so important?

IMB is content-empty by design — it tells you the three layers must be addressed but does not tell you what to put in each layer. The content is population-specific. Skipping elicitation means filling the model with generic content that addresses imagined gaps rather than measured gaps, which is the single most common cause of disappointing IMB results.

Why are behavioral skills the proximal determinant?

Skills compiled into procedural memory operate under cognitive load, time pressure, and emotional arousal: the conditions under which target behaviors actually occur. Information and motivation operate under conditions of calm reflection, which is not where behavior is performed. The neurobiology of skill consolidation (basal ganglia, motor cortex) is structurally different from the neurobiology of explicit knowledge (dlPFC) and valuation (vmPFC, striatum), which is why skills are the bottleneck.

What is the difference between objective skill and perceived self-efficacy?

Objective skill is the actual capacity to perform the behavior under realistic conditions. Perceived self-efficacy is the learned belief about one’s own capacity. Both are required and both are addressed by the BS construct. Skill without efficacy produces avoidance; efficacy without skill produces failed attempts that crater future motivation.

How does IMB integrate with the Octalysis Framework?

The Octalysis Crosswalk maps each of IMB’s three layers onto the eight Core Drives. Motivation in particular benefits from the disaggregation, with the designer working across eight distinct motivational systems instead of one motivation bucket, each with its own M-vector and BS-vector. The six-step audit (elicit, map to Core Drives, design at the Core-Drive level, sequence across phases, reinforce self-efficacy, run the publicity test) is the operational procedure.

What domains has IMB been applied to outside HIV prevention?

Antiretroviral medication adherence, vaccination uptake, smoking cessation, exercise initiation, diet change, cancer screening, oral health behaviors, contraceptive use, blood-glucose self-monitoring among diabetics, and an increasing share of digital health-app onboarding and behavior-change product design.

Is IMB still considered current?

Yes. Meta-analyses continue to show IMB-based interventions outperforming single-construct comparators across health-behavior domains. The model has been extended (IMB-AT for adherence, sIMB-CIM for care-initiation and maintenance, with multiple adaptations for digital behavior support) but the core architecture has remained stable since 1992, which is itself evidence of structural soundness.

References

1. Fisher, J. D., & Fisher, W. A. (1992). Changing AIDS-risk behavior. Psychological Bulletin, 111(3), 455–474. — The original model paper.
2. Fisher, J. D., Fisher, W. A., Williams, S. S., & Malloy, T. E. (1994). Empirical tests of an information–motivation–behavioral skills model of AIDS-preventive behavior with gay men and heterosexual university students. Health Psychology, 13(3), 238–250.
3. Fisher, J. D., Fisher, W. A., Misovich, S. J., Kimble, D. L., & Malloy, T. E. (1996). Changing AIDS risk behavior: Effects of an intervention emphasizing AIDS risk reduction information, motivation, and behavioral skills in a college student population. Health Psychology, 15(2), 114–123.
4. Fisher, W. A., Fisher, J. D., & Harman, J. (2003). The Information–Motivation–Behavioral Skills Model: A general social psychological approach to understanding and promoting health behavior. In J. Suls & K. A. Wallston (Eds.), Social Psychological Foundations of Health and Illness (pp. 82–106). Blackwell.
5. Fisher, J. D., Amico, K. R., Fisher, W. A., & Harman, J. J. (2008). The information–motivation–behavioral skills model of antiretroviral adherence and its applications. Current HIV/AIDS Reports, 5(4), 193–203. — The canonical IMB-AT extension.
6. Amico, K. R. (2011). A situated-information motivation behavioral skills model of care initiation and maintenance (sIMB-CIM): An IMB model based approach to understanding and intervening in engagement in care for chronic medical conditions. Journal of Health Psychology, 16(7), 1071–1081.
7. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211.
8. Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice-Hall. — The conceptual parent of the BS construct.
9. Bandura, A. (1997). Self-Efficacy: The Exercise of Control. Freeman.
10. Rosenstock, I. M. (1974). The health belief model and preventive health behavior. Health Education Monographs, 2(4), 354–386.
11. Michie, S., Atkins, L., & West, R. (2014). The Behaviour Change Wheel: A Guide to Designing Interventions. Silverback Publishing. — COM-B canonical reference.
12. Schwarzer, R. (2008). Modeling health behavior change: How to predict and modify the adoption and maintenance of health behaviors. Applied Psychology, 57(1), 1–29. — HAPA, the dynamic complement to IMB.
13. Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59(4), 329–349.
14. Sheeran, P. (2002). Intention–behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12(1), 1–36. — The intention–behavior gap meta-analysis.
15. Chou, Y. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media. — The canonical Octalysis reference.

Related Reading

• Health Belief Model (the threat-appraisal sibling pillar).
• Theory of Planned Behavior (the attitudes-and-norms framework IMB extends).
• COM-B / Behavior Change Wheel (the modern environmental sibling).
• Health Action Process Approach (HAPA) (the dynamic sequencing complement).
• EPPM (the fear-appeal calibration model for CD8-heavy interventions).
• Libertarian Paternalism (the publicity-test ethical anchor).
• The Behavioral Framework Library (every behavioral model in one place).

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?