Switch How to Change Things When

1. 1Name a specific destination, not a vague aspiration. Translate 'eat healthier' or 'reduce purchasing costs' into one crystal-clear behavior the Rider can picture (Booth-Butterfield and Reger did this by targeting the single act of switching to 1% milk).
2. 2Script the critical moves. Instead of listing everything that could change, isolate the two or three behaviors that matter most and spell them out step by step—like Pronovost's five-item line-infection checklist (wash hands, clean skin with antiseptic, etc.).
3. 3Find bright spots and point the Rider at them. Rather than analyzing every problem, locate an existing instance where the desired outcome already occurs and direct effort toward replicating it—this stops analysis paralysis before it starts.
4. 4Make the ask specific and behavioral. Convert intentions into measurable actions: Kelman challenged agencies to 'double your use of government credit cards over the next year' rather than 'improve procurement.'
5. 5Shrink the requested behavior so it reads as unambiguous and small. Winsten asked for exactly 'five seconds' of dialogue—so simple no writer felt their independence challenged and no one had to interpret what compliance meant.
6. 6Give the Rider an ironclad reference he can't dispute. Back the scripted move with black-and-white evidence (Pronovost's steps were backed by solid medical research), so the analytical mind stops debating and starts executing.
7. 7Formalize commitment to lock the direction in. Kelman asked agencies to sign a written 'Pledge'—turning a defined destination into a public, trackable behavior.
8. 8Build in a checklist for complex or high-stakes moments to prevent the Rider's blind spots and overconfidence (Cisco uses one to vet acquisitions: will key engineers relocate, can we sell to their customers, what's the support plan).

Michigan ICUs faced preventable 'line infections' that cost money and lives. Dr. Peter Pronovost of Johns Hopkins refused to treat this as an analysis problem—there was no new science needed. Instead he directed the Rider with a five-part checklist of straightforward, black-and-white steps: doctors should wash their hands before inserting a line, the patient's skin should be cleaned with antiseptic at the insertion point, and so on. The steps were backed by solid medical research, so a doctor could ignore the checklist but could not dispute it. Over eighteen months of use, line infections were nearly eliminated, saving hospitals an estimated $175 million and about fifteen hundred lives. The power came not from new knowledge but from removing ambiguity—showing people the ironclad right way to act.

• If the desired behavior can be described more than one way (like 'eat healthier'), then it isn't directed yet—narrow it to a single scriptable action before proceeding.
• If experts have written position papers but nothing has changed (as in Sternin's Vietnam), then stop analyzing problems and point the Rider at an existing bright spot.
• If there is an ironclad right way backed by evidence, then build a checklist so the Rider can't dispute it; if there isn't, still use a checklist to guard against blind spots and overconfidence.
• If the change is big enough to drain self-control, then shrink the requested action (Winsten's 'five seconds') so the Rider stays fresh.
• If the ask is abstract, then attach a measurable target and a public pledge (Kelman's 'double credit-card use') to make compliance unambiguous.

• Directing the Rider with a goal instead of a behavior—'be healthier' or 'cut costs' leaves the Rider to spin its wheels choosing where to start. Tell: people nod, agree, and then nothing happens.
• Front-loading analysis. Like the Vietnam experts, you catalog every problem and write plans, exhausting effort on sizing up rather than acting. Tell: mounting research documents, zero behavior change.
• Making the scripted move too big. A large change saps self-control (the radish-eaters quit in eight minutes). Tell: early enthusiasm collapses into 'it's too hard, we're no good at this.'
• Assuming resistance when the real issue is lack of clarity. You push harder on motivation while the Rider is actually just confused about direction. Tell: people who claim to support the change can't say what step comes next.
• Relying on willpower to hold the direction. The Rider tugs the reins but tires; without a checklist or pledge, the behavior reverts. Tell: compliance decays whenever supervision lapses.

• ▸Vague aspiration: write the fuzzy goal people currently state
• ▸Scripted behavior: rewrite it as one specific, observable action
• ▸Destination: describe what success looks like in a single concrete sentence
• ▸Bright spot: name one existing case where this already works
• ▸Critical-move checklist: list the 3-5 black-and-white steps to execute
• ▸Evidence anchor: cite the fact/result that makes the move indisputable
• ▸Shrink test: state the smallest version of the ask (e.g., 'five seconds')
• ▸Commitment device: name the pledge, target, or metric that locks it in
• ▸Ambiguity check: note any part still open to more than one interpretation

1. 1Diagnose which failure you face: if people understand what to do but don't act, the Elephant is reluctant—stop adding logic and start adding feeling. Analytical appeals reach Riders, not Elephants.
2. 2Make people feel something, not just know something. Show them a visceral artifact of the problem rather than a spreadsheet—pile the gloves on the table so the gut reaction 'We're crazy!' precedes the fix.
3. 3Choose positive over negative emotion for most change. Reserve fear/anger (negative emotions) only for stone-in-the-shoe situations needing quick, specific action; if the change needs creativity and flexibility, negative emotion won't get you there.
4. 4Build a destination postcard: a vivid, near-term picture (months or years, not decades) of what's possible that also 'hits you in the gut.' Make it emotional, not just big ('You'll be third graders soon').
5. 5Shrink the change so the first step feels doable. The Elephant is easily demoralized and spooked; give reassurance even for the very first step.
6. 6Reframe people as already partway to the goal—put two stamps on their card. Make them feel closer to the finish line than they thought so they get a jolt of enthusiasm and momentum.
7. 7Grow identity: help people see themselves as the kind of person who does this ('I'm not a sloth—I'm an Exerciser'), which changes behavior from that moment forward.
8. 8Protect self-control, because it's an exhaustible resource. Don't rely on willpower to drag a reluctant Elephant; engaging emotion means the Rider doesn't have to yank the reins alone.

Jon Stegner wanted to overhaul a company's purchasing system. The easy, natural move was a 12-tabbed Excel spreadsheet of savings data and consolidation logic—an appeal aimed purely at the Rider. Instead, he dumped a pile of gloves on the boardroom table, including the same glove being bought at wildly different prices, and invited his bosses to look. Their first reaction was emotional—'We're crazy!'—followed immediately by 'We can fix this.' Had he made only the analytical case, he'd have gotten supportive nods and a follow-up meeting six weeks out that might get rescheduled indefinitely; the Riders would have understood without the Elephants being motivated. Once he broke through to feeling, the execs' Elephants got fired up to move, and everyone could think of things to try. A single employee with a summer intern, focused on one product, produced a change with billion-dollar reach—the power of speaking to the Elephant, not just the Rider.

• If people have understanding but no action, then the Elephant is unmotivated—switch from logic to feeling, not from feeling to more logic.
• If the change needs quick, specific action (stone-in-the-shoe), then negative emotion may help; if it needs creativity and flexibility, then use positive emotion instead.
• If the goal is far off or abstract, then reframe it as a near-term, gut-smacking destination postcard people can picture.
• If people feel they're starting from scratch, then give them a head start—show them the two stamps already on the card so they feel partway there.
• If you're relying on willpower to sustain the change, then expect it to fail—self-control exhausts like a muscle, so engage emotion so the Rider doesn't have to force the Elephant.

• Speaking only to the Rider: the 12-tab spreadsheet that makes an accountant weep but leaves Elephants cold. Tell: supportive nods, requests for a follow-up meeting 'next year,' then rescheduling.
• Reaching for the burning platform by default: assuming fear always motivates. Tell: you get panic or avoidance but no sustained creative effort, because negative emotions only target narrow, specific actions.
• Setting a big goal with no emotional charge—a BHAG that's big but doesn't 'hit you in the gut.' Tell: people can recite the goal but feel nothing and don't move.
• Leaving people at the starting gate of a shorter journey instead of partway through a longer one. Tell: low follow-through—only 19 percent finish (eight-stamp group) versus 34 percent when given a head start.
• Depending on willpower to drag a reluctant Elephant. Tell: early compliance that collapses (radish-eaters quit in eight minutes vs. nineteen)—self-control ran out because it's an exhaustible resource.

• ▸Change being asked for: state the specific behavior in one line.
• ▸Current diagnosis: is this understanding-without-motivation, or genuine confusion? (If confusion, direct the Rider instead.)
• ▸Feeling to evoke: what should people feel, and is positive emotion the fit (vs. a rare stone-in-the-shoe case)?
• ▸Visceral artifact: what 'pile of gloves' can you show that triggers gut reaction before analysis?
• ▸Destination postcard: one vivid near-term picture (months/years) that hits the gut.
• ▸Head start / two stamps: how will you show people they're already partway, not at zero?
• ▸Identity to grow: which new self-label ('I'm an Exerciser') can people adopt from this moment?
• ▸First tiny step: the single smallest move that reassures the easily-spooked Elephant.
• ▸Willpower check: where are you relying on self-control that will exhaust—and what emotion replaces it?

1. 1Diagnose the situation, not the person: when a behavior keeps failing, ask 'How can I change the situation so the right behavior is more likely?' rather than concluding people are lazy or careless—this is Pronovost's move with line infections.
2. 2Tweak the environment to make the desired behavior the path of least resistance and the undesired behavior harder: remove the trigger (delete the Outlook alert sound), block the temptation (cover the taskbar icon with a Post-it), or resize the container (swap dinner plates for salad plates).
3. 3Build habits by pairing an action with a concrete situational trigger: specify exactly when and where the behavior fires ('right after I drop Anna at school, I head to the gym') so it runs on autopilot without spending the Rider's exhaustible self-control.
4. 4Where possible, engineer the environment so the habit is forced, not chosen—like Rackspace throwing out the call-queuing system so staff had no option but to answer the phone.
5. 5Rally the herd: make the desired behavior visibly the norm and recruit peers, because behavior is contagious. Get a critical mass onboard privately, then point holdouts to the majority ('almost everyone but you is participating').
6. 6Engineer early, visible wins to generate hope (Elephant fuel), and find a yardstick to make progress measurable when no off-the-shelf scale exists.
7. 7Preload difficult decisions and stack supportive habits—checklists, stand-up meetings, laying out jogging clothes the night before, setting coffee to auto-brew—so the Rider isn't fighting the Elephant in the moment.

Steve Kelman wanted federal agencies to break tradition and consider past performance in contracts, but bureaucratic inertia was enormous. Rather than argue people into it, he shaped the Path socially. He worked the phones and had his employees and advisers rally their networks, and he decided to push the pledge publicly only once he was sure he had at least eight agencies signed up. After reaching nine, participation snowballed—he could now tell holdouts, 'Almost everyone but you is participating.' Twenty agencies eventually took the Past Performance Pledge, and to keep it real he prodded them to name fifty-eight specific upcoming contracts where past performance would be considered. The result: five years later 70 percent of frontline employees called themselves proponents of reform, and Brookings graded it the only 'reinventing government' initiative to earn an A.

• If a behavior fails repeatedly across many people, then treat it as a situation problem and redesign the environment—don't diagnose character (avoid the Fundamental Attribution Error).
• If self-control is the only thing standing between someone and the wrong behavior, then remove the temptation entirely (toss the Cheetos, shrink the plates) rather than relying on willpower, which is exhaustible.
• If you want a behavior to persist, then attach it to a specific situational trigger and, when feasible, make the environment force it rather than merely invite it.
• If you're recruiting a herd, then secure a private critical mass first, make participation visible, then use the majority to pull in holdouts.
• If early progress isn't naturally measurable, then build a yardstick and make advances visible to keep the Elephant motivated.

• Blaming the person instead of the situation—the tell is language like 'he's hopelessly absentminded' or 'they're negligent doctors,' which points at character and skips the environmental fix (a checklist taped to the steering wheel).
• Relying on willpower to override temptation—the tell is repeated resolutions that collapse, like checking Outlook every time the alert sounds; the fix is deleting the trigger, not resisting it.
• Making only mental resolutions with no situational trigger—the tell is vague intent ('I'll go to the gym more') with no specified when/where, so the behavior never fires.
• Assuming environmental tweaks alone will produce every habit—the tell is expecting a rearranged room to make you learn piano; some habits also need mental groundwork like action triggers.
• Announcing a herd change before you have critical mass—the tell is pushing a pledge publicly with too few onboard, so it reads as fringe instead of inevitable and never snowballs.

• ▸Target behavior: the specific desired behavior stated in observable terms
• ▸Situation diagnosis: what in the current environment makes the wrong behavior easy (the trigger/temptation to remove)
• ▸Environment tweak: the concrete change that makes the right behavior easier (e.g., resize, relocate, remove alert, force via system change)
• ▸Action trigger: 'When [situational cue], I/we will [behavior]'—name the exact time and place
• ▸Habit support: the recurring ritual or checklist that runs the behavior on autopilot
• ▸Herd critical mass: the number/list of people you'll privately sign up before going public
• ▸Norm-making move: how you'll make the behavior visibly the majority and address holdouts
• ▸Progress yardstick: how you'll measure and display early wins to fuel the Elephant

1. 1Start with the noble but ambiguous goal ('eat healthier,' 'reform procurement,' 'change direction') and admit it can't be executed as stated — a Rider facing it spins his wheels.
2. 2List every plausible way someone could interpret the goal (eat more grains? take vitamins? switch to Diet Coke?). The length of that list is your evidence that ambiguity, not resistance, is the obstacle.
3. 3Hunt for the single high-leverage behavior that moves the goal most — the one concrete swap that, if adopted, delivers a disproportionate share of the outcome (Booth-Butterfield and Reger found whole-to-1% milk hit the USDA saturated-fat target by itself).
4. 4Translate that behavior into a specific script naming the exact action, place, and time: not 'act healthier' but 'next time you're in the dairy aisle, reach for a jug of 1% milk instead of whole milk.'
5. 5Write decision rules that pre-answer the hard trade-offs for people, so every choice point resolves itself (Behring's four rules: faster is better than best, minimize up-front cash, unblock revenue, use what you've got).
6. 6Refuse to hide behind big-picture vision. Get into the details where the paralysis actually lives; 'set the vision and stay out of the details' fails in change because the paralyzing part is the details.
7. 7Measure the concrete behavior, not the abstraction — track the specific move (share of low-fat milk sold) so you can see whether clarity is producing movement.

Two West Virginia professors, Booth-Butterfield and Reger, wanted people to eat healthier — a goal so open-ended the Rider drowns in options (breakfast or dinner? home or restaurant? grains or meat?). Instead of preaching, they narrowed the whole ambiguous goal to one behavior: switch from whole milk to 1% or skim, because milk is the single largest source of saturated fat in the typical American diet and the swap alone brings the average diet to USDA saturated-fat levels. Their script was concrete about action, place, and time: reach for a jug of 1% milk in the dairy aisle instead of whole milk. Across eight intervention-area stores, low-fat milk's market share rose from 18 percent before the campaign to 41 percent after, holding at 35 percent six months later. The lesson: the West Virginians weren't people who 'didn't care about their health' — if they were, behavior wouldn't have shifted so easily. What looked like resistance was a lack of clarity, and one crystal-clear move dissolved it.

• If the goal can be interpreted in many valid ways, then it's ambiguous — script one concrete behavior before asking anyone to move.
• If people share the goal but aren't acting, then treat it as a clarity problem, not a motivation or character problem (child abusers in PCIT shared normal parenting goals; their methods and instructions were wrong).
• If the road ahead is uncertain, then expect the Elephant to default to the status quo, because uncertainty breeds anxiety and the familiar path is always the status quo.
• If a decision point recurs, then pre-write a rule that dictates the correct choice (e.g., 'use invasive options only as a last resort'), so the Rider spends no energy re-deciding.
• If you're tempted to lead hands-off with vision alone, then stop — the paralyzing part of change is in the details, so script the critical moves.

• Diagnosing a clarity gap as resistance or apathy. The tell: you conclude these are 'the kind of people who don't care,' yet their behavior shifts easily the moment you give a concrete instruction.
• Setting vision and staying out of the details. The tell: leaders proud of high-level direction watch people revert to the default path because no one told them the exact next move.
• Offering the abstract goal ('act healthier,' 'be more innovative') and expecting execution. The tell: everyone agrees with the goal but no one starts, because their Riders are still contemplating limitless interpretations.
• Leaving recurring trade-offs unscripted. The tell: people re-agonize over the same investment or care decision each time, and default to the safest familiar option (invasive hip surgery, buying new locomotives).
• Mistaking option overload for the only source of paralysis. The tell: you strip choices but ignore ambiguity — not knowing which options even exist paralyzes just as surely as 24 jams on a table.

• ▸Ambiguous goal: state the noble goal in the vague words people currently use.
• ▸Interpretation list: write out 5+ ways someone could reasonably interpret that goal.
• ▸High-leverage behavior: name the one concrete action that moves the goal most, and why.
• ▸Exact script: specify the action + place + time in one sentence a stranger could follow.
• ▸Pre-decided trade-offs: list 2-4 if/then rules that answer the hardest recurring choices.
• ▸Default-path risk: name the status quo behavior people revert to under uncertainty.
• ▸Behavior metric: state the specific, measurable action you'll track (not the abstraction).
• ▸Baseline & target: record the starting number and the movement that proves clarity worked.

1. 1Separate understanding from motivation: verify your audience already grasps the need to change intellectually, then diagnose whether their Elephant feels anything about it—if you get supportive nods but no movement, feeling is missing.
2. 2Make the problem viscerally concrete rather than analytical. Instead of a spreadsheet or a report, stage a physical, undeniable demonstration—Stegner dumped every glove the company bought onto a boardroom table so his bosses could see and touch the waste.
3. 3Aim for a specific emotion, not 'motivation' in general. Choose the one that fits the moment: hope when people feel stuck, urgency when they're complacent, pride when they've succeeded, empathy when the change serves others.
4. 4Deliver a jolt that produces a two-step reaction: first shock ('We're crazy!'), then agency ('We can fix this'). Shock alone paralyzes; pair it immediately with a believable, do-able next action so the fired-up Elephant has somewhere to go.
5. 5Set a destination postcard—a vivid, near-term, gut-hitting picture of success ('You'll be third graders soon!')—so the felt emotion points at a concrete future, not an abstraction.
6. 6Mine bright spots for hope: find people succeeding against the odds (the healthy kids, the sober moments, the two transformed managers) and show your audience 'there really is a way, and it's something I can do.'
7. 7Let people act their way into feeling. Design hands-on participation—the cooking groups where mothers prepared food together—so the emotion arises from doing, not from being told.

Jon Stegner wanted to overhaul his company's wasteful purchasing system, a potential billion-dollar change. He could have built a 12-tabbed Excel spreadsheet that would make a tax accountant weep, and the analytical case was genuinely compelling. But an analytical appeal would have earned supportive nods and a follow-up meeting six weeks out—reaching Riders, not Elephants, yielding direction without motivation. Instead he had a summer intern gather every different glove the company purchased and dumped the whole pile on the boardroom table, inviting the executives to see them. First they thought 'We're crazy!'—then 'We can fix this,' and everyone could name a few things to try. That jolt got their Elephants fired up to move on a single product that stood in for the entire ordering process. The scope of the presentation didn't match the scope of the proposal, yet it worked, because he broke through to feeling.

• If you get supportive nods but no action, then you've reached the Rider and missed the Elephant—stop adding analysis and add feeling.
• If people feel stuck or fatalistic, then supply hope via bright spots ('a poor family CAN have a well-nourished child'); if they feel complacent, then supply urgency or shock.
• If a goal is abstract ('eat healthier,' 'act healthier'), then convert it to a gut-hitting destination postcard that also gives direction ('third graders soon,' 'reach for 1% milk').
• If the emotion has no immediate do-able action attached, then it will curdle into paralysis—always pair the jolt with 'here's what we can try.'
• If the change is an imported outside solution, then people's Elephants will bristle ('those people aren't like us')—source the emotion and solution from within the group instead.

• Substituting intellectual proof for feeling: you deliver a rigorous spreadsheet and get agreement without movement. Tell: everyone agrees it's important 'next year' and reschedules the meeting.
• Willpower-only mobilization: you rely on the Rider tugging the reins to force the Elephant. Tell: early compliance that fades fast as people 'get too busy'—the Rider gets exhausted.
• Shock without agency: you horrify people about how bad things are but give no path forward. Tell: defensiveness, denial, or paralysis instead of 'we can fix this.'
• Defaulting to the negative/problem frame: you ask 'What's broken?' and lead with fear only, missing hope and pride. Tell: your language is all snow (bad emotion words dominate) and the Elephant digs in.
• Making the emotional appeal abstract: you exhort people to 'care more' or 'be motivated.' Tell: no specific feeling, no near-term picture, no concrete first sign of success—nothing for the Elephant to move toward.

• ▸Change I'm driving: state the specific behavior change in one sentence.
• ▸Do they already understand? note evidence the Rider gets the need (nods, agreement, prior reports).
• ▸Target emotion: pick ONE—hope, pride, enthusiasm, empathy, or urgency—and why it fits.
• ▸The jolt: describe the concrete, physical, or undeniable demonstration (my 'pile of gloves').
• ▸Shock-then-agency line: write the 'We're crazy!' realization AND the 'We can fix this' first action.
• ▸Destination postcard: one vivid, near-term, gut-hitting picture of success.
• ▸Bright spot: name a real person/case already succeeding against the odds to prove it's possible.
• ▸Act-into-feeling step: the hands-on activity that lets people experience the change, not just hear it.
• ▸Source check: confirm the solution comes from within the group, not imported (avoids 'not like us').