MINDSPACE Framework: S-Tier Behavioral Designer’s Guide

The shortest sentence in behavioural science is the one most policy teams quote and then quietly ignore: most behaviour change happens automatically, not consciously. Dolan, Hallsworth, Halpern, King and Vlaev wrote it into the opening page of their 2010 report for the UK Cabinet Office, packaged the nine highest-leverage automatic-system levers under the acronym MINDSPACE, and handed governments a checklist for moving citizens without trying to change their minds.

Sixteen years later the checklist is everywhere. It sits behind the tax letters that pulled hundreds of millions of pounds out of late filers, the organ-donation defaults that lifted consent rates across half of Europe, the energy bills that put your neighbour’s usage next to your own, and the pension auto-enrolment systems that now route trillions of dollars in the United States and the United Kingdom. And it sits behind a much longer list of polished interventions that produced a 0.4-percentage-point lift in a pilot, got publicised as a triumph, and quietly evaporated when the team running it moved on.

That gap is the post. MINDSPACE is the right inventory and the wrong stopping point. Read the original 96-page report and you have a vocabulary. Pair it with the Octalysis Framework and you have an operating system for deciding which lever to pull, in what order, for which Core Drive, against which ethical line.

⚡ Speed Run Notes

• MINDSPACE is a checklist, not a theory. Nine automatic-system levers (Messenger, Incentives, Norms, Defaults, Salience, Priming, Affect, Commitments, Ego) Dolan distilled for the UK Cabinet Office in 2010.
• Three of the nine carry most of the documented wins. Defaults, Norms and Incentives drive the field-experiment literature; Priming and Salience are over-represented in famous demos and under-represented in replications.
• Effect size and persistence are the two axes that matter. A 12% lift that fades in a week is worth less than a 3% lift that holds forever. Architectural levers stick; campaign levers fade.
• Every lever maps to an Octalysis Core Drive. Messenger and Norms anchor Social Influence; Defaults, Commitments, Incentives anchor Ownership; Ego anchors Development; Affect anchors Epic Meaning.
• EAST dropped four letters and got more shipped. The Behavioural Insights Team rewrote MINDSPACE as Easy · Attractive · Social · Timely in 2014 because nine levers in a Whitehall meeting turned into zero levers pulled.
• The ethics problem is structural. Every lever works whether the goal is pension auto-enrolment or in-app purchase manipulation of teenagers. The libertarian-paternalism publicity test is the only honest floor.

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 MINDSPACE?

MINDSPACE is a nine-letter checklist for designing interventions that change behaviour without relying on conscious persuasion. Each letter names one of the most evidence-backed automatic-system levers behavioural science had documented by 2010: Messenger, Incentives, Norms, Defaults, Salience, Priming, Affect, Commitments, Ego. The acronym was assembled by Paul Dolan (LSE), Michael Hallsworth, David Halpern, Dominic King and Ivo Vlaev, and published by the UK Cabinet Office and the Institute for Government in February 2010 as MINDSPACE: Influencing Behaviour through Public Policy.

The framework’s claim is narrower than the field often pretends. It is not a theory of behaviour change. It does not predict which lever will work in a given context. It does not tell you in what order to combine levers, or what to do when two of them point in opposite directions. What it does is give a working civil servant a vocabulary that fits on one slide and a comprehension check — if you cannot name which lever your intervention is pulling, you have not designed it yet.

That sounds modest until you watch what happens without it. Before MINDSPACE most UK government behaviour-change work was either (a) information campaigns that assumed people read the leaflet, or (b) financial incentives that assumed people computed the discount. The framework’s real contribution was to make the automatic system — the fast, habitual, context-driven layer Kahneman would later popularise as System 1 — legible to people who had been trained their whole careers in the reflective system.

The Nine MINDSPACE Levers

Each lever is a specific empirical pattern with thirty-plus years of behavioural-science evidence behind it. The framework asks designers to consider all nine on every brief, then choose the two or three with the highest expected effect for the specific context.

M — Messenger

Who delivers the information changes whether the information lands. People give more weight to messages from those who look like them, share their group identity, hold formal authority, or have demonstrated past expertise. The classic field experiment is Garner’s 2005 study where the same persuasive request gained 23% more compliance simply by being signed by an academic instead of a generic respondent. UK tax compliance lifts replicated the pattern at scale: HMRC letters signed by a local official outperformed identical letters signed by the Chief Executive of HMRC, because local identity reads as more trustworthy than central authority.

Messenger effects compound in ways behavioural-insights teams routinely under-weight. The same nurse delivering the same vaccine advice gets one response when she’s the only NHS visitor a family sees in six months, and another response when she’s the third person from a London accent in a Newcastle estate. The lever isn’t “use a credible messenger.” The lever is credibility in the eyes of this specific audience, which is a research problem before it is a design problem.

I — Incentives

This is the lever everyone thinks they understand and most policy designs get wrong. MINDSPACE’s incentive lever isn’t classical economics — it’s the behavioural-economics correction to classical economics. Five sub-mechanisms matter: loss aversion (losses hurt roughly twice as much as equivalent gains feel good), mental accounting (a £100 tax rebate isn’t spent like a £100 paycheck), reference points (every offer is judged against a salient comparison), hyperbolic discounting (future rewards get heavily devalued), and the endowment effect (people demand more to give up what they already own than to acquire it in the first place).

Most government incentive design loses 80% of its potential because it ignores frame. A £200 tax credit framed as a gain converts at a third the rate of the same £200 reframed as a loss being recovered. The Pension Service’s “you could be losing £1,200 per year by not claiming Pension Credit” outperformed every gain-framed version of the same intervention by margins that would look like a coding error to a classical economist.

N — Norms

People are powerfully influenced by what others do, especially when the others resemble them and the behaviour is observable. The HMRC tax-letter experiments are the cleanest UK example: adding the sentence “Nine out of ten people in your postcode have already paid their tax” to reminder letters lifted on-time payment enough to recover roughly £200 million in additional revenue over the trial period. The lever doesn’t require new information — it requires making existing social information salient.

Two failure modes show up constantly. The first is descriptive-norm backfire: telling people who already pay their taxes that 35% don’t can lower their compliance, because the norm cuts in the wrong direction. The second is reactance — if the norm reads as manipulative, the target audience flips. The Schultz et al 2007 energy-bill experiments showed both: a smiley-face injunctive norm next to the descriptive comparison prevented the “I’m below average, I can use more” boomerang.

D — Defaults

The default option carries enormous behavioural weight because doing nothing is easier than doing something, and because the default is read as the implicit recommendation. Johnson and Goldstein’s 2003 paper on European organ-donation rates is the textbook case: explicit-consent (opt-in) countries averaged 15% consent; presumed-consent (opt-out) countries averaged 98%. Same culture, same religion, same baseline ethical views — the difference is the form.

Pension auto-enrolment carries the same signature. The US 401(k) participation rate jumped from ~50% to ~90% in companies that switched from opt-in to opt-out, with the largest jumps among workers earning under $20,000 a year — precisely the population that needs the savings most. The UK’s 2012 auto-enrolment rollout added roughly 10 million workers to occupational pensions in eight years, by a single change in default.

Defaults are not just lazy choice. They carry a moral-architecture claim: the default is what the designer believes the unmotivated person should land on. That makes default-setting one of the most ethically loaded decisions in policy design, and one of the most under-discussed.

S — Salience

Attention is finite and chronically over-subscribed. Whatever is novel, accessible, simple to process, or personally relevant captures the cognitive foreground; everything else gets metabolised as background noise. The salience lever asks: of all the things competing for your audience’s attention, what one thing do you need them to actually see?

The HMRC studies found that simplifying tax-return language and prominent display of the deadline produced lifts comparable to the social-norm sentence. New York City’s simplified mortgage default-resolution form lifted homeowner response rates from 12% to 24% with no change in substantive content — only in salience and processing fluency.

Salience earns its place in the framework because it’s the lever most designers know about and least design for. They write the leaflet, they add the bold heading, they tick the box. Real salience design treats attention as the scarce resource it is and pays for every gram of it.

P — Priming

Priming claims that sub-conscious cues in the environment shape subsequent behaviour without the actor noticing. The original behavioural-economics literature loaded this lever with theatrical demonstrations: Bargh’s “elderly-words make you walk slower” experiment, North’s “French music in the wine aisle sells French wine” field study, the “clean smell makes you tidier” results from Holland and Hendriks.

The replication crisis hit priming harder than any other MINDSPACE lever. The walking-slower-after-elderly-words finding failed multiple direct replications. Many social-priming effects shrank to non-detectable when registered replications ran with adequate samples. The honest 2026 read: priming is real, but the effect sizes are an order of magnitude smaller than the early literature claimed, and the contexts in which it reliably works are narrower than 2010 reviewers believed. Treat any priming intervention as a low-confidence bet that needs its own field experiment, not a sure win.

A — Affect

Emotional reactions to a stimulus shape decisions before any reasoning kicks in. Slovic’s “affect heuristic” research shows that the same risk feels safer or more dangerous depending on whether the felt-emotion is positive or negative, independent of the underlying probability. Affect drives why people fear flying more than driving, why a charity ad showing one identifiable victim outperforms statistical aggregates by an order of magnitude, and why the “disgust” reaction to a public-health risk swamps the “mild concern” framing of the same risk.

For policy design, affect is the lever that turns dry information into something the limbic system actually responds to. Anti-smoking campaigns that show black lungs in real time outperform statistical mortality figures by margins the rational model cannot explain. The risk of over-reliance is straightforward: affect-driven design at scale is propaganda, and the line between “making the harm vivid” and “manufacturing fear that exceeds the underlying probability” is the ethics question every campaign team should be required to answer in writing.

C — Commitments

People work surprisingly hard to act consistently with their public commitments and reciprocate favours, especially when both are observable to others they care about. Cialdini’s commitment-and-consistency research underpins this lever; the field-trial evidence comes from hotel towel-reuse studies (committing to reuse on check-in lifted compliance dramatically), gym attendance contracts, debt-repayment pledges, and voter-turnout interventions where a verbal commitment to vote lifted actual turnout by several percentage points.

The commitment lever is structurally close to the gamification mechanic of public progress: by stating the goal publicly, the person now has identity at stake on top of the original incentive. The combination is sticky in ways the original incentive alone wasn’t. The honest caveat: commitment-only interventions decay rapidly without paired structural support. A pledge to exercise three times a week without a calendar invite, a gym partner, or an environmental change reverts to baseline inside three months.

E — Ego

The ego lever names a fundamental motivational pattern: people act in ways that make them feel better about themselves. This includes self-image consistency (the person who has already donated to a charity is more likely to donate again, because to do otherwise would contradict their self-concept), self-serving bias (we credit ourselves for successes and external factors for failures), and the labelling effect (telling first-time voters they have been classified as “likely voters” lifted actual turnout by 15 percentage points in Tybout and Yalch’s 1980 study).

For policy design, the ego lever is the cleanest CD-bridge to identity-based interventions. James Clear’s “identity-based habits” pattern — targeting the “I am a non-smoker” statement rather than the “I am quitting smoking” statement — is an applied-ego intervention dressed in self-help language. The 2014 BIT “you are a registered organ donor” message variant lifted sign-ups by ~3.5% on a base that was already at structural ceiling for any non-identity intervention.

The Cabinet Office Origin Story

The MINDSPACE report was not the first behavioural-economics framework. Thaler and Sunstein’s Nudge had appeared in 2008. Cialdini’s Influence was twenty-five years old. What MINDSPACE was, and what gave it disproportionate adoption inside government, was the first time a UK Cabinet Office body had endorsed an automatic-system framework as the recommended basis for behaviour-change interventions across departments. The report was published with the imprimatur of the Institute for Government and the explicit signature of then-Cabinet Secretary Sir Gus O’Donnell.

That endorsement created the institutional space for the Behavioural Insights Team to be founded inside the Cabinet Office in July 2010, with David Halpern (one of the MINDSPACE authors) as its first director. The BIT’s first major experiments used MINDSPACE’s vocabulary verbatim — the early HMRC tax-letter trials, the Job Centre Plus letter experiments, the organ-donation prompts on the gov.uk driving-licence renewal page were all explicitly framed as “testing the N and S levers” in internal documents from 2010 to 2012.

By 2014 the BIT had spun out as a partly-mutualised social-purpose company, expanded to roughly 200 staff across multiple continents, and rebadged its core toolkit from MINDSPACE to EAST (Easy, Attractive, Social, Timely). The new acronym wasn’t a refutation of the old — the BIT’s own internal memos describe EAST as a deliberate operational compression. Nine levers were too many for a 30-minute briefing with a minister; four were the right cardinality for the political-attention budget. Comprehensive lost to operationalised, and that loss is the framework’s most-quoted institutional lesson.

What Dolan and the BIT Got Right

Three things, and they all sit in the same paragraph of the original report.

First, the framework made automatic-system thinking legible to a non-specialist audience. The civil-service career path does not require any psychology training, yet civil servants design interventions that change the lives of millions of citizens. Before MINDSPACE, the existing options were a 400-page Thaler textbook or a one-paragraph “use nudges” recommendation. The nine-letter acronym sat exactly at the level where a smart non-expert could apply it on Monday without misreading the underlying science.

Second, the framework was diagnostic before it was prescriptive. The report’s own examples don’t prescribe; they label. The Pension Credit letter example walks the reader through which levers the original letter pulled (almost none), which the redesigned letter pulled (Salience, Incentives, Ego), and how to measure the lift. The diagnostic framing protected the framework from the “just add a Norms statement and you’ll get a 30% lift” misuse pattern that destroyed several other behavioural frameworks in commercial settings.

Third, the framework was institutionally seeded with field experimentation. The BIT’s first principle — lifted from MINDSPACE’s closing chapter — was Test, Learn, Adapt: every intervention runs as a randomised controlled trial with a registered hypothesis, a control group, and a pre-specified primary outcome. That commitment is the single biggest reason MINDSPACE produced a defensible evidence base where most behaviour-change frameworks produced sympathetic anecdotes. The Haynes, Service, Goldacre and Halpern Test, Learn, Adapt 2012 paper is the canonical statement of the methodology; every applied MINDSPACE programme worth taking seriously cites it.

Where MINDSPACE Falls Apart

The framework has three structural failure modes. None of them invalidate MINDSPACE as a vocabulary; all of them show up the moment teams try to use it as an operating system.

1. The Automatic-System Bias

MINDSPACE is explicitly a list of automatic-system levers. By construction, it omits the reflective-system levers behavioural science also documents: deliberation prompts, education, structured choice architectures that increase cognitive engagement rather than decrease it, reason-and-evaluate interventions. That omission was deliberate in 2010 — the report’s thesis is that the reflective system is over-represented in policy design. A decade later the omission looks like a load-bearing weakness. The single biggest field experiment in modern behavioural science — Volpp and colleagues’ smoking-cessation trial — succeeded not because it pulled MINDSPACE levers but because it paid $750 commitment-contingent rewards backed by a year-long deliberate-choice protocol. Many of the highest-impact behaviour changes are reflective-system shifts that MINDSPACE’s vocabulary cannot describe.

2. The Persistence Problem

MINDSPACE’s field-experiment evidence base is heavily weighted toward short-horizon outcomes — six weeks for tax letters, three months for default-changes, one year for energy-use comparison reports. The Allcott and Rogers 2014 follow-up on the OPower energy-bill experiments showed the effect partly decays once the intervention is withdrawn, with elasticity falling from ~2% at six months to ~1.4% at four years, but only for households where the report keeps arriving. The pattern repeats across nudges: most MINDSPACE wins are sticky only while the lever is still being pulled. The Allcott-Rogers paper is one of the few that ran long enough to see the decay; many of the headline interventions were never measured past the first quarter, leaving the field openly uncertain about persistence at the policy scale most matters.

3. The Conflation of Tactic and Strategy

A MINDSPACE lever is a tactic. A behaviour-change programme is a strategy. The framework gives no opinion on how to combine tactics, which to sequence first, when the same lever’s effect inverts depending on context, or how to detect when two levers are pulling in opposite directions. The Schultz energy-bill experiment is the canonical illustration: descriptive norms alone backfired for households below the average; injunctive norms (the smiley face) corrected the backfire. A team applying MINDSPACE without that meta-level reasoning will ship the broken intervention and report “norms don’t work for us.” The fix isn’t a different framework; it’s an explicit causal-chain analysis sitting on top of the lever inventory. That’s the gap Octalysis fills, which is the section after next.

What’s Really Happening Inside the Brain

The neuroscience evidence supporting the MINDSPACE levers concentrates in three brain systems, and seeing the mapping makes the framework’s strengths and weaknesses much sharper.

The basal ganglia and the dopaminergic loop running through the striatum mediate most habit-driven and default-following behaviour. Defaults work because acting against them requires explicit recruitment of prefrontal control over an already-automated path. fMRI work by Daw and colleagues shows that the cost of deviating from a default rises with cognitive load, which is why default effects amplify under time pressure and decision fatigue. This is the brain system MINDSPACE’s D, P and S levers most directly engage.

The medial prefrontal cortex mediates the self-relevant processing that drives the Ego and Affect levers. Northoff’s decade of work on cortical midline structures shows that information tagged as “about me” gets preferentially encoded and recalled relative to information about generic others, even when the content is identical. That asymmetry is why personalised tax letters outperform generic ones, why “you are a registered voter” sticks better than “voters in your area turn out at 65%,” and why affect-laden self-relevant framings out-engage neutral statistical framings by margins that look implausible to economists.

The anterior cingulate cortex and the social-cognition network including temporoparietal junction and superior temporal sulcus mediate the Messenger, Norms and Commitments levers. The brain’s computational machinery for tracking what others believe, what others have promised, and what others are likely to do is enormously expensive and runs continuously. That is why the lift from a credible messenger or a salient norm is so large relative to the marginal cost of changing the wording — the audience’s brain is already doing the inferential work; the intervention just gives it a different input.

The brain story also clarifies why Priming has fallen so far in credibility. The original priming literature assumed wide cross-domain effects between sub-cortical perceptual cues and high-level behaviour. The neuroscience says these effects exist but are domain-specific, short-lived, and depend on uninterrupted contextual continuity between prime and behaviour — conditions rarely met in real-world policy settings. The 2026 read of the brain evidence is: MINDSPACE’s seven non-priming levers are well-grounded; Priming is real in the lab and unreliable in the field.

MINDSPACE vs Other Behaviour-Change Frameworks

MINDSPACE sits in a crowded field of behaviour-change frameworks. The honest comparison clarifies what it’s for and what it isn’t.

MINDSPACE vs EAST

EAST (Easy, Attractive, Social, Timely) is MINDSPACE’s direct successor, published by the Behavioural Insights Team in 2014. EAST collapses nine levers into four operational principles, deliberately dropping Priming (after the replication crisis), folding Ego into Attractive, folding Commitments into Timely, and absorbing Defaults into Easy. The compression cost the framework comprehensiveness but bought adoption. Inside the BIT’s own commercial work, EAST briefings have an order of magnitude higher action rate than MINDSPACE briefings ever had. For policy work, EAST is the operational doctrine and MINDSPACE is the diagnostic vocabulary you reach for when EAST doesn’t fit.

MINDSPACE vs COM-B

The COM-B model (Capability, Opportunity, Motivation → Behaviour), developed by Michie and colleagues at UCL in 2011, sits at a higher level of abstraction than MINDSPACE. Where MINDSPACE inventories the levers, COM-B asks the prior question: is the target behaviour failing because of a capability gap, an opportunity gap, or a motivation gap? In practice, a well-designed programme runs COM-B as the diagnostic and MINDSPACE (or EAST) as the toolbox — COM-B tells you which gap to close, MINDSPACE tells you which specific levers can close it. Treating them as competing frameworks is a category error.

MINDSPACE vs Nudge Theory

Thaler and Sunstein’s Nudge (2008) provided the conceptual frame that made MINDSPACE possible. Nudge is the “why” (libertarian paternalism, choice architecture, the two-system model of cognition); MINDSPACE is the “how” (here are the nine specific levers most documented to work). A policy team that has read Nudge but not MINDSPACE knows that defaults matter but has no inventory of which other levers belong in the same toolkit; a team that has read MINDSPACE but not Nudge has the inventory but no ethical or theoretical framing for when to use it.

MINDSPACE vs Fogg Behavior Model

BJ Fogg’s B = MAT model (Behaviour = Motivation × Ability × Trigger) is the most direct competitor for the “single-frame, ships on Monday” slot MINDSPACE occupies. Fogg’s frame is tighter and more deployable for product-design contexts (Silicon Valley adopted it almost universally); MINDSPACE’s frame is richer and more defensible in policy contexts (governments adopted it almost universally). The split is not accidental — Fogg optimises for a designer building a single product and a single user journey; MINDSPACE optimises for a team designing across heterogeneous populations and contested ethical terrain.

MINDSPACE in the Real World

Four applied domains have produced the strongest evidence base for MINDSPACE-style interventions. The pattern across all four is the same: the framework is at its strongest when paired with rigorous field experimentation, and at its weakest when reduced to a one-off campaign with no measurement plan.

Government & Tax Compliance

HMRC’s tax-letter experiments are the cleanest applied MINDSPACE record. Hallsworth and colleagues’ 2017 paper, The behavioralist as tax collector, reports a series of randomised trials across roughly 200,000 UK taxpayers. The strongest single intervention combined a descriptive social norm (“Nine out of ten people in your area pay their tax on time”) with an injunctive norm (“You are currently in the very small minority”) and produced lifts that translated to hundreds of millions of pounds in additional revenue at population scale. The strongest single tactic was the personalised salience treatment: writing the recipient’s name and amount owed in a visually emphasised box at the top of the letter. Both interventions cost close to nothing per letter beyond the wording change.

Energy & Sustainability

OPower (now Oracle Utilities) embedded MINDSPACE’s Norms and Affect levers in residential energy bills across 100+ US utilities, comparing each household’s usage to similar neighbours alongside a smiley/sad face. The Allcott 2011 paper estimates the intervention reduced electricity consumption by ~2% on average, equivalent to a substantial price increase but at near-zero variable cost. The follow-up work showed effect persistence over multiple years (with the partial decay noted above), making OPower one of the longest-horizon nudge studies on record. The Norway-based Nudge Plus programme replicated the effect direction at smaller magnitude in a different cultural context, confirming the lever travels imperfectly across cultures.

Health Behaviour

NHS-aligned BIT trials applied MINDSPACE across organ-donation prompts (Defaults + Norms), missed-GP-appointment reminders (Salience + Commitments), and antibiotic-prescribing letters to GPs (Messenger + Norms). The Hallsworth 2016 GP-letters trial was particularly clean: letters to high-prescribing GPs signed by the Chief Medical Officer and including a peer-comparison sentence reduced unnecessary antibiotic prescriptions by ~3.3% over six months, with the effect concentrated in the highest-prescribing decile. The cost per averted prescription was a few pence; the public-health value (slowing antibiotic resistance) was orders of magnitude higher.

Workplace & UX

The 401(k) auto-enrolment work (Madrian and Shea, Thaler and Benartzi’s Save More Tomorrow) is the clearest workplace application: pure Defaults plus paired Commitments (committing in advance to direct future raises into savings). The same logic applied inside consumer product design produces the entire pattern of cookie-banner defaults, app-permission defaults, subscription-renewal defaults, and one-click checkout choice architectures. The ethical line between “applied MINDSPACE” and “applied sludge” in these contexts is the publicity test the next section names — would the designer be willing to publish the default-setting logic openly to the affected users?

The Elephant in the Room

MINDSPACE is morally neutral, which is the same thing as saying it is morally dangerous. Every lever in the framework works exactly as well whether the goal is pension auto-enrolment or in-app purchase manipulation of teenagers. The Messenger that increases NHS vaccination uptake is mechanically identical to the Messenger that increases authority-laundered crypto fraud. The Default that adds ten million workers to occupational pensions is structurally the same Default that auto-renews a $79.99 subscription nobody remembers signing up for. The skill of doing MINDSPACE well is the same skill regardless of which side of the public-good line the designer sits on.

The field’s usual response to this elephant is to call the unethical applications “sludge” (a coinage by Sunstein himself, given pillar treatment elsewhere in this library) and assume the labelling is the policy. It isn’t. The libertarian-paternalism argument from Thaler and Sunstein offers the only structurally honest test: would the designer be willing to publish, under their own name, to the affected population, the full set of (a) what behaviour change is being targeted, (b) which MINDSPACE levers are being pulled to achieve it, (c) what the funder paying for the intervention stands to gain, and (d) what the affected population stands to gain or lose? An intervention that survives that publication test is libertarian-paternalist. An intervention that depends on opacity to function is sludge regardless of which side of the public-good line it sits on.

The reason this test is the field’s ethical floor (and not a higher bar) is that it imposes no rule about which behaviour changes are permissible. It only imposes a rule about transparency. A pension provider running auto-enrolment passes the publication test trivially; a subscription company running dark-pattern cancellation flows fails it on the first sentence. The test is asymmetric in exactly the right direction: it rules out the cases everyone agrees are unethical without requiring the field to settle the harder cases about which behaviour changes governments should be permitted to engineer in the first place.

Two practical consequences for any team applying MINDSPACE. First, write the publication-test version of the intervention brief on the same day you write the implementation brief, not as a post-launch audit. Second, treat any lever that requires opacity to function as a yellow flag and any intervention that requires opacity across multiple levers as a red flag that should not ship.

How to Apply MINDSPACE with the Octalysis Framework

MINDSPACE tells you which automatic-system lever to pull. Octalysis tells you why pulling it will work, when it will backfire, and what to pair it with. The two frameworks operate at different levels of abstraction and complement each other cleanly.

The MINDSPACE × Octalysis Crosswalk

Each MINDSPACE lever has a primary Core Drive it engages and at least one secondary Core Drive that determines whether the intervention persists or decays.

Messenger → Core Drive 5 (CD5): Social Influence & Relatedness — primary. The lever is the brain’s social-cognition machinery responding to who delivers the signal. Secondary: Core Drive 1 (CD1): Epic Meaning & Calling when the messenger represents a higher cause the audience identifies with. Design implication: if the audience doesn’t share an identity or a cause with the messenger, the lever does nothing regardless of the messenger’s objective credentials.

Incentives → Core Drive 4 (CD4): Ownership & Possession (for endowment-and-loss frames) + Core Drive 8 (CD8): Loss & Avoidance (for loss-framed offers) + Core Drive 6 (CD6): Scarcity & Impatience (for time-bounded offers). Design implication: classical economic incentives without CD4 framing leak 60-80% of their potential lift; the MINDSPACE incentive lever is functionally a Core Drive 4 / 6 / 8 cocktail and should be designed as one.

Norms → Core Drive 5 (CD5): Social Influence & Relatedness — primary. The lever is what others like me are doing. Secondary: Core Drive 2 (CD2): Development & Accomplishment when the norm establishes a benchmark the audience wants to meet or exceed. Design implication: descriptive norms without a CD2 hook produce the “I’m above average, I can use more” boomerang Schultz documented; injunctive norms with a CD2 frame (smiley face = “you are doing well”) lock the lift.

Defaults → Core Drive 4 (CD4): Ownership & Possession — primary. The status-quo bias underlying default effects is structurally an endowment effect: the existing arrangement is already “mine” in the cognitive sense. Secondary: Core Drive 8 (CD8): Loss & Avoidance — opting out of the default registers as a loss. Design implication: defaults persist because they engage two of the strongest Core Drives simultaneously, which is why their effect sizes dwarf almost every other lever in the inventory.

Salience → Core Drive 3 (CD3): Empowerment of Creativity & Feedback — primary, in the sense that salience design is the “feedback signal” half of CD3 made visible. Secondary: Core Drive 7 (CD7): Unpredictability & Curiosity for novelty-based salience treatments. Design implication: salience-only interventions are CD3-thin and don’t persist; salience paired with a CD2 progress signal or a CD5 social comparison does.

Priming → Core Drive 7 (CD7): Unpredictability & Curiosity — primary, in the upstream-pull-in sense. The honest design implication is the inverse of the brief: priming is the weakest evidenced lever in the inventory, so treat any priming intervention as needing its own field experiment before scaling, and never build a programme that depends on priming as its primary mechanism.

Affect → Core Drive 1 (CD1): Epic Meaning & Calling — primary, when the affect is recruited toward a cause larger than the individual. Secondary: Core Drive 8 (CD8): Loss & Avoidance for fear-appeal affect treatments. Design implication: positive-affect treatments tied to identifiable beneficiaries (CD1) outperform statistical aggregates by margins so large the design choice often dominates everything else in the intervention.

Commitments → Core Drive 4 (CD4): Ownership & Possession (identity consistency) + Core Drive 5 (CD5): Social Influence & Relatedness (when public). Design implication: a private commitment is a thin CD4 hook that decays; a public commitment to an in-group recruits CD5 and persists. The hotel-towel field experiments worked because the commitment was to the room, made visible to the next guest, and tied to identity rather than to a cash incentive.

Ego → Core Drive 2 (CD2): Development & Accomplishment — primary, in the “self-as-competent” sense. Secondary: Core Drive 4 (CD4): Ownership & Possession for identity-based interventions. Design implication: the “you are a registered voter” label works because it creates an identity the person now owns (CD4) and an implicit competence claim they want to live up to (CD2). The labelling effect is the cleanest demonstration in the MINDSPACE inventory of a Core Drive cocktail outperforming any single-Core-Drive intervention.

All 8 Octalysis Core Drives are reachable through the MINDSPACE inventory if the Crosswalk is run carefully. CD3 (Empowerment) shows up through Salience; CD7 (Unpredictability) shows up through Priming and novelty-based salience; CD6 (Scarcity) shows up through time-bounded Incentives. The teams that ship MINDSPACE-correct interventions and still get flat results are almost always missing one of these three coverage gaps.

The MINDSPACE × Octalysis 6-Step Audit

The Crosswalk above is the static mapping. The dynamic part — the one that actually produces shippable interventions — is a 6-step audit you run on every brief before any design work begins.

Step 1: Elicit before designing. Run a structured belief-elicitation with 8–12 target users before drafting any intervention. This is the step every applied team skips, and the failure-mode trace of every behavioural-insights programme that flat-lined runs back to this skip. You cannot design Norms without knowing which reference group the target audience belongs to; you cannot design Messenger without knowing whose credibility translates in this specific context. Twelve 30-minute conversations are not negotiable.

Step 2: Map the highest-leverage MINDSPACE levers to Core Drives. Use the Crosswalk to identify the primary CD each candidate lever engages. Score each (lever × CD) pair on (a) base-rate evidence the lever moves the behaviour in this domain, (b) cultural fit with the target audience, (c) institutional fit with the implementing team. The output is a ranked shortlist of 2–4 (lever, CD) pairs.

Step 3: Design at the Core Drive level, not the lever level. This is the translation step most CMO/policy teams stall on. A “Defaults intervention” is a tactic; a “CD4 Ownership intervention that uses Defaults as its primary mechanic and pairs with Commitments for persistence” is a design. The CD-level framing forces explicit treatment of decay, ethics, and secondary effects that lever-level design ignores.

Step 4: Sequence across the four Octalysis Experience Phases. The Discovery phase elicits the audience’s pre-existing beliefs (this is where Step 1 lives). The Onboarding phase corrects the most consequential misconceptions and establishes the initial framing. The Scaffolding phase builds the structural support that makes the new behaviour sustainable. The Endgame phase maintains the behaviour after the original intervention is withdrawn. Most MINDSPACE wins die in the Scaffolding-to-Endgame transition because the intervention was designed for Onboarding only.

Step 5: Build the persistence channel deliberately. The persistence problem is a design problem, not an external constraint. Defaults persist because the architecture changed; Commitments persist if the social context that witnessed the commitment is still present; Ego-based interventions persist as long as the identity is salient. For every lever-CD pair from Step 2, write the explicit persistence mechanism on the brief. If you can’t name it in one sentence, the intervention will fade.

Step 6: Run the publicity test before shipping. Write the public-facing version of the intervention brief: what behaviour change is being targeted, which levers are being pulled, who is paying for it, and what the affected population gains or loses. If the team is unwilling to publish that brief under its own name to the affected population, the intervention is sludge and should not ship in its current form. Redesign until it passes.

Practical Steps for Applying MINDSPACE

If you are a behaviour-change practitioner, policy designer, UX lead, or product manager looking at MINDSPACE for the first time, the following six steps will get you from “I read the report” to “I shipped an intervention that worked.”

1. Define one specific target behaviour with a measurable lift. Not “more pension savings.” Specifically: “the percentage of new hires who remain enrolled in the 401(k) ninety days after onboarding rises from 47% to 70%.” The MINDSPACE checklist is useless without a measurable target; with one, it is a structured ideation tool.
2. Elicit the current automatic-system path. Run 8–12 user interviews with audio recordings. Ask: what happens at the decision point today? What does the person see, hear, feel? What is the default? Who, if anyone, is messaging them? You cannot pull a lever you haven’t located.
3. Score each MINDSPACE lever on (a) plausibility for this behaviour, (b) cultural fit, (c) team capability. A 1–5 scale on each axis, ranked. Cut anything below 9/15 from the shortlist. You will end with 2–4 levers, never nine.
4. Run the Octalysis Crosswalk on each remaining lever. Identify primary and secondary Core Drives. Add the explicit persistence mechanism. Note any ethics-test red flags. The output is a shortlist of (lever, CD, persistence, ethics) tuples ready for prototyping.
5. Build the minimum testable intervention and run an RCT. Two arms minimum: status quo and the strongest single-lever version. If feasible, add a paired-lever arm. Pre-register the primary outcome and the analysis plan. Run it for at least one decay cycle — six weeks for short-cycle behaviours, six months for annual ones.
6. Publish the result, including the failures. The single biggest improvement to the global behavioural-insights evidence base would be public, structured publication of interventions that didn’t work. The BIT’s commitment to this is the gold standard; replicate it.

MINDSPACE Was the Beginning, Not the End

MINDSPACE earned its place in the canon because it was the first behaviour-change framework that working civil servants could apply on Monday morning. It made automatic-system thinking legible, gave the field a shared vocabulary, and seeded institutional homes for the methodology that produced most of the field’s defensible evidence base.

It is not the last framework anyone should read. EAST compressed it for operational ship rate. COM-B sits above it diagnostically. Octalysis sits beneath it as the why-engine and the design-time discipline. The skilled practitioner uses all four, plus the underlying field-experiment literature, plus a publicity test for every brief. The unskilled practitioner reads MINDSPACE once and ships a Norms statement on a leaflet.

If you build interventions that influence other people’s behaviour at scale, the question MINDSPACE leaves you with is not which lever to pull. It is the older, harder question: would you publish the design brief to the people it’s being applied to? Answer that honestly on every project and you have the ethical floor MINDSPACE itself doesn’t supply.

Frequently Asked Questions About MINDSPACE

What does MINDSPACE stand for?

MINDSPACE is an acronym for nine behavioural-economics levers documented to influence behaviour through automatic, non-conscious channels: Messenger, Incentives, Norms, Defaults, Salience, Priming, Affect, Commitments, Ego. The framework was published in 2010 by Paul Dolan, Michael Hallsworth, David Halpern, Dominic King and Ivo Vlaev for the UK Cabinet Office and the Institute for Government.

Who created the MINDSPACE framework?

The framework was assembled by Paul Dolan (London School of Economics), Michael Hallsworth, David Halpern, Dominic King and Ivo Vlaev. It was commissioned by the UK Cabinet Office during the late Brown government and published in February 2010 under the joint imprint of the Cabinet Office and the Institute for Government. David Halpern went on to direct the Behavioural Insights Team founded later that year, where MINDSPACE became the operational toolkit until it was succeeded by EAST in 2014.

What is the difference between MINDSPACE and EAST?

EAST (Easy, Attractive, Social, Timely) is MINDSPACE’s direct operational successor, published by the Behavioural Insights Team in 2014. EAST collapsed MINDSPACE’s nine levers into four principles to improve adoption inside time-constrained policy meetings. The compression dropped Priming (after the replication crisis), folded Ego into Attractive, and absorbed Defaults into Easy. MINDSPACE remains the richer diagnostic vocabulary; EAST is the operational doctrine most BIT briefings now use.

What is the strongest evidence base for MINDSPACE?

The HMRC tax-letter experiments (Hallsworth et al, 2017), the OPower energy-bill experiments (Allcott, 2011; Allcott and Rogers, 2014), pension auto-enrolment (Madrian and Shea, 2001; Thaler and Benartzi, 2004), and the European organ-donation defaults comparison (Johnson and Goldstein, 2003) are the four cleanest large-scale field-experiment records. Each isolates one or two MINDSPACE levers, runs at population scale, and reports measurable lifts on outcomes that matter.

Which MINDSPACE lever has the weakest evidence?

Priming. The original behavioural-economics literature heavily featured demonstrative priming experiments (Bargh’s elderly-words study, North’s wine-aisle music study) which subsequently failed direct replication or shrank to non-detectable effect sizes under registered-replication protocols. Priming effects are real but smaller, more context-dependent, and less reliable than the 2010 MINDSPACE report assumed. Treat any priming-based intervention as a low-confidence bet requiring its own field experiment.

Is MINDSPACE the same as nudging?

Closely related but not identical. Thaler and Sunstein’s Nudge (2008) is the theoretical and ethical frame for choice-architecture interventions; MINDSPACE is the practical toolkit of nine specific levers documented to work within that frame. Every MINDSPACE intervention is a nudge in the Thaler-Sunstein sense, but Nudge as a concept covers a broader design space including sludge audits, libertarian-paternalist meta-rules, and choice-architecture domains the original MINDSPACE inventory doesn’t name.

How does MINDSPACE relate to gamification?

Every MINDSPACE lever maps cleanly to one or more Octalysis Core Drives, which means a well-designed gamification system is functionally an applied MINDSPACE intervention with stronger persistence mechanisms built in. The MINDSPACE Crosswalk in this post details the mapping; in short, Norms and Messenger anchor in CD5 (Social Influence), Defaults and Commitments and Incentives anchor in CD4 (Ownership), Ego in CD2 (Development), Salience in CD3 (Empowerment), Priming in CD7 (Unpredictability), Affect in CD1 (Epic Meaning).

Can MINDSPACE be used unethically?

Yes, and trivially. Every lever in the framework works whether the target behaviour is in the user’s interest or against it. Manipulative dark patterns in commercial UX, subscription-traps with default auto-renewal, and authority-laundered crypto scams all use the same MINDSPACE levers as pension auto-enrolment or tax-compliance letters. The libertarian-paternalist publicity test — would the designer publish the intervention brief openly to the affected population? — is the field’s minimal ethical floor.

How long do MINDSPACE effects last?

It depends on the lever. Architectural levers (Defaults, Commitments backed by structural change, Ego-based identity interventions) tend to persist for years or until the architecture changes. Campaign levers (Salience, Affect, Priming) typically decay within weeks or months once the intervention is withdrawn. The Allcott-Rogers 2014 follow-up on OPower energy bills found a partial decay even for repeated-treatment interventions, with the effect dropping from ~2% at six months to ~1.4% at four years. Plan for decay as a design constraint, not as a surprise.

What should I read after MINDSPACE?

The Behavioural Insights Team’s EAST framework paper for the operational successor; Test, Learn, Adapt (Haynes et al, 2012) for the methodology; Cialdini’s Influence for the persuasion-theory foundations; Thaler and Sunstein’s Nudge for the choice-architecture frame; Michie’s COM-B and Behaviour Change Wheel work for the diagnostic level above MINDSPACE; and Yu-kai Chou’s Actionable Gamification for the Octalysis design discipline that complements MINDSPACE at the design-time level.

References

1. Dolan, P., Hallsworth, M., Halpern, D., King, D., & Vlaev, I. (2010). MINDSPACE: Influencing Behaviour through Public Policy. London: Institute for Government and Cabinet Office.
2. Dolan, P., Hallsworth, M., Halpern, D., King, D., Metcalfe, R., & Vlaev, I. (2012). Influencing behaviour: The mindspace way. Journal of Economic Psychology, 33(1), 264–277.
3. Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions about Health, Wealth, and Happiness. New Haven: Yale University Press.
4. Hallsworth, M., List, J. A., Metcalfe, R. D., & Vlaev, I. (2017). The behavioralist as tax collector: Using natural field experiments to enhance tax compliance. Journal of Public Economics, 148, 14–31.
5. Hallsworth, M., Chadborn, T., Sallis, A., et al. (2016). Provision of social norm feedback to high prescribers of antibiotics in general practice: A pragmatic national randomised controlled trial. The Lancet, 387(10029), 1743–1752.
6. Service, O., Hallsworth, M., Halpern, D., Algate, F., Gallagher, R., Nguyen, S., Ruda, S., Sanders, M., et al. (2014). EAST: Four Simple Ways to Apply Behavioural Insights. London: Behavioural Insights Team.
7. Haynes, L., Service, O., Goldacre, B., & Torgerson, D. (2012). Test, Learn, Adapt: Developing Public Policy with Randomised Controlled Trials. London: Cabinet Office Behavioural Insights Team.
8. Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95(9–10), 1082–1095.
9. Allcott, H., & Rogers, T. (2014). The short-run and long-run effects of behavioral interventions: Experimental evidence from energy conservation. American Economic Review, 104(10), 3003–3037.
10. Johnson, E. J., & Goldstein, D. G. (2003). Do defaults save lives? Science, 302(5649), 1338–1339.
11. Madrian, B. C., & Shea, D. F. (2001). The power of suggestion: Inertia in 401(k) participation and savings behavior. Quarterly Journal of Economics, 116(4), 1149–1187.
12. Thaler, R. H., & Benartzi, S. (2004). Save more tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164–S187.
13. Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2007). The constructive, destructive, and reconstructive power of social norms. Psychological Science, 18(5), 429–434.
14. Cialdini, R. B. (2008). Influence: Science and Practice (5th ed.). Boston: Pearson.
15. Michie, S., van Stralen, M. M., & West, R. (2011). The behaviour change wheel: A new method for characterising and designing behaviour change interventions. Implementation Science, 6(42).
16. Fogg, B. J. (2009). A behavior model for persuasive design. Proceedings of the 4th International Conference on Persuasive Technology, 40.
17. Kahneman, D. (2011). Thinking, Fast and Slow. New York: Farrar, Straus and Giroux.
18. Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2007). The affect heuristic. European Journal of Operational Research, 177(3), 1333–1352.
19. Chou, Y.-K. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media.
20. Sunstein, C. R. (2022). Sludge: What Stops Us from Getting Things Done and What to Do about It. Cambridge: MIT Press.

Related Reading

• The Octalysis Framework — the 8 Core Drives MINDSPACE’s nine levers map into.
• Sludge (Sunstein, 2022) — the dark mirror of MINDSPACE’s nudge inventory.
• Libertarian Paternalism (Thaler & Sunstein) — the publicity test that should sit on every MINDSPACE brief.
• COM-B and the Behaviour Change Wheel (Michie et al) — the diagnostic frame that sits above MINDSPACE.
• Nudge Theory (Thaler & Sunstein) — the conceptual foundation MINDSPACE operationalised.
• The Behavioral Framework Library — every framework in this series, indexed.

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