Rory Sutherland

P01 Psycho-logic: solve the psychological problem before the physical one.

When the problem lives in perception (waiting, anxiety, boredom, status), changing how it feels is usually far cheaper than changing the physical reality, and the customer experiences the result as real value.

Ask: If we could not change the physical metric at all, what could we change about how the experience is perceived? / What is the “feels like” version of the target we are optimising, and is our brief a costly proxy for it? / Are we paying for an improvement customers cannot perceive, while a cheap change would be perceived as large?

Signs: The brief is written in engineering or financial units (minutes, capacity, cost per unit) while the complaint is a feeling. Large capital plans aim at gains past the point customers notice. Only one discipline has looked at the problem. People describe the experience in words (stressful, pointless, dead time) that no KPI measures.

P02 Anxiety, not time, is the cost: reduce uncertainty.

People mind not knowing far more than they mind waiting, so visible progress and reassurance can improve an experience more than making it faster.

Ask: Where do customers wait without knowing how long, or where they are in the process? / What anxiety is the customer really trying to resolve by contacting us or using us? / Where do people suffer narrow misses because of a barrier we control?

Signs: Repeat “where is it?” contacts. Complaints about waits that are not unusually long. People choosing slower but moving options. Services used mainly to check that nothing is wrong. Demand that spikes when information goes dark.

P03 Improve the quality and meaning of time, not only its length.

Time is not simply waste to be cut; making it usable, pleasant or socially positive can be worth more than shortening it, and each step of a process sends a message (progress, upgrade, rebuff) that people react to more than to the minutes.

Ask: Is the time customers spend with us wasted or usable, and what would make it valuable rather than shorter? / What social message does each step send: progress, status, rejection? / Where do people feel they lose something by complying, and is that loss ever acknowledged or refunded?

Signs: Large spending plans to cut minutes from a process. People choosing slower, calmer options when allowed. Anger at steps that send people backwards or make them wait idle. Rule-following that feels like a sacrifice nobody thanks.

P04 Perception is prediction: design for surprise.

The brain mostly perceives its own predictions and notices departures from them, so satisfaction and memory come from the gap between expectation and experience, not from absolute quality.

Ask: What do customers expect here, and where could we cheaply exceed it? / Are we spending to match expectations nobody will notice, such as a refurbishment to category standard? / Where are we over-promising, so that a good experience still registers as a let-down?

Signs: Investment that produces no change in sentiment. Experiences customers describe as “fine” and never mention. Marketing that promises more than operations deliver. Delight that comes from small unplanned acts, not from the big programmes.

P05 Cues must match the claim (everything is a bit of a placebo).

How a product is framed, priced, packaged, dosed and delivered changes how well it works for people, so the sensory and ceremonial cues around it must fit what it claims and what is at stake, even when that makes it “worse”.

Ask: Does the sensory experience of our product signal the benefit we promise? / Is the effort, packaging or ceremony around this proportional to what the customer is risking? / Which meaning-bearing cues have we stripped out for efficiency, and did they make the product work better?

Signs: A product that tests well blind but disappoints in use. Customers doubting something works because it is too easy, too cheap or too pleasant. High-stakes decisions reduced to self-service. Patients or clients who judge care by attention rather than outcome. Efforts to “command” behaviour that only respond to atmosphere.

P06 Dare to be trivial: small, oblique causes, large effects.

Human behaviour is a complex system in which the size of an intervention need not match the size of its effect, so the job is to find the smallest, often obvious-seeming lever with the biggest effect.

Ask: What tiny change in wording, labelling, mapping, default or script might shift behaviour a lot? / Would a random passer-by, or a first-time user, know what we do and how to start? / Is our response sized to look serious rather than to work?

Signs: A problem everyone treats as needing a big programme. Boards that consider details beneath them. A small perceptual bottleneck nobody owns (signage, naming, visibility, first-screen wording). Visible gestures that leave the underlying behaviour unchanged.

P07 Fix the frequent small irritations; subtract before you optimise.

Quality of life and of service is shaped mostly by small, recurring frictions that are beneath the dignity of important people to fix, and the best reform is often to stop doing something rather than do it more efficiently.

Ask: What are the ten most frequent irritations our customers or staff face, and who owns them? / What would we be better off stopping entirely? / Which “improvements” did we add without checking anyone wanted them?

Signs: Petty penalties, confirmations and forms that everyone resents and nobody defends. Tools nobody senior uses, left unimproved for years. Reform plans that only add. Staff workarounds for small daily annoyances.

P08 Blame the model, not the person.

If people persistently do something after a long evolution and the model says they should do otherwise, the model is more likely wrong than the people; “irrational” behaviour usually has a hidden logic, and old heuristics can be right for the wrong reasons.

Ask: Are we calling customer behaviour irrational when it could be adaptive for reasons our model omits? / Does this practice produce good results whatever its official reason? / What would be lost if we removed a rule or habit nobody can explain?

Signs: Plans that depend on “educating” customers out of a stubborn habit. Behaviour that persists across cultures and decades. A model whose predictions keep failing in the same direction. Proposals to remove an old custom because its rationale is outdated.

P09 Think like Darwin, not Newton: hold several explanations.

Understand human systems bottom-up and historically, through variation, selection and observation, and resist single-cause explanations; in complex systems the honest answer is often “it depends”.

Ask: What other explanations fit the same data, and would any of them lead to the opposite action? / What factors are left out of our model because they are hard to measure, and would the answer change with them in? / How did this situation come to be, and what would it look like to someone outside our group?

Signs: One theory treated as “the only game in town”. Segments described as homogeneous. A team that shares one background and one data source. Confident explanations that arrived as soon as the numbers fitted economic theory.

P10 The press office brain: reasons come after the decision.

People decide largely on instinct and emotion and then produce plausible reasons, so stated reasons and objections are often post-hoc cover for an unspoken driver or fear.

Ask: Is the explanation we accept merely the first plausible one? / What might people be afraid or embarrassed to admit about this choice? / Are we designing for the reasons people give, or for the emotional drivers behind them?

Signs: Objections that shift when answered. Rational-sounding resistance that no information changes. Surveys with tidy reasons and behaviour that contradicts them. Hesitation clustered around a moment that is awkward, intimate or socially exposed.

P11 Stated versus revealed preference: unmet needs and met un-needs.

What people say they want differs from what changes their behaviour, so research-led firms invest in things customers don’t value while missing needs nobody can articulate, and keep doing things nobody wants.

Ask: Which of our investments are justified only by what customers say, and which by what they do? / What do we do that customers tolerate but don’t want? / What nearly stopped people buying, and what do successful oddities in our category tell us?

Signs: High stated intent with low uptake (people say they’d pay more for green). Features requested in research and then unused. “Services” customers decline or work around. A product succeeding for a reason its makers never intended.

P12 Detective before courtroom: investigate before you prove.

Problem-solving needs an investigative phase that uses anecdotes, anomalies and hunches to decide where to look, before an evidential phase that demands proof; anything genuinely new has no data until you act.

Ask: Are we rejecting this option because it lacks data that could only exist after we try it? / What odd thing has someone noticed recently, and did we throw it away as anecdote? / What would have to be true for the opposite of our core assumption to hold?

Signs: Proposals vetoed for “lack of evidence” in a domain where things change. Frontline observations that never reach decision-makers. Research read for what people say rather than why. Compliance or finance setting the proof bar for exploration.

P13 All data comes from the past.

Data and the tools trained on it inherit status-quo bias, the bias toward what is countable, and the gaps of what was recorded; behavioural data shows what happened without why, and one bad probabilistic assumption can make a conclusion wildly wrong.

Ask: What does our data systematically not contain? / What other psychological explanations fit this pattern, and would they lead to the opposite action? / Are we treating an output as fact because a computer produced it, and which probabilistic assumption is it resting on?

Signs: Dashboards with no emotional or attitudinal signal. Confident conclusions drawn from correlations. Models trained on a world that has since changed. Chains of probability multiplied as if independent.

P14 Averages lie: 10 × 1 is not 1 × 10.

Designing for the average designs for nobody, and the same total spread differently (across people or across time) is a different human reality, so look at distributions, frequencies and individual trajectories.

Ask: Whose experience is hidden in our average, and who bears the burden most often? / How does this look to one customer experiencing it repeatedly over years? / Which weak link in our offer is multiplying everything else toward zero?

Signs: Policies that are fine on average but crippling for frequent users. Products designed to a composite persona. Pricing built for the average buyer. Satisfaction scores that look stable while a minority churns. A single catastrophic failure mode offset in the plan by many small strengths.

P15 Pair every cost with a true upside (sweet lemons).

People can live with an unavoidable cost if they are given a genuine upside or a story that it was their choice; what enrages them are burdens with no redeeming feature.

Ask: Which of our unavoidable negatives has no upside story attached? / What is our most obvious weakness, and what true benefit does it carry? / When we force a compromise, is there a plausible reason the customer could tell themselves it was what they wanted?

Signs: Recurring complaints about a constraint you cannot remove (a slow step, a distant location, a mandatory charge). Anger at burdens that seem to buy nothing. A compromise option that people take resentfully.

P16 Naming and story create value.

What you call something, the story you tell about it and the frame you place it in change how people value it, use it and whether they will identify with it, with no change to the thing itself.

Ask: How are we naming this, and in whose units? / What word or frame makes people reluctant to identify with our product or behaviour? / Is there a flaw or by-product we could rename as a feature, or existing inventory we could describe so that it commands a premium?

Signs: A good product with a clinical, embarrassing or technical name. Specs expressed in the maker’s units. Existing assets invisible because unnamed or unmapped. A product people use but won’t admit to.

P17 Change the frame of comparison.

Value is judged relative to whatever is at hand, so changing what something is compared with can turn a weakness into a strength without changing the product.

Ask: What is this being compared with at the moment of choice, and what would we like it compared with? / Is there a dimension on which we win that customers are not attending to? / Is ours a category where customers never ask what else the money could buy?

Signs: Losing on the one dimension the category always compares. A premium product judged against a cheap substitute rather than the expensive alternative it replaces. Customers in a category who borrow to the maximum without considering alternatives.

P18 Price is a feeling: change how you charge, not only what.

To economists price is a number; to customers it is a feeling shaped by timing, frequency, method, bundling, framing and who else benefits, so the mechanics of charging (and of selling and packaging) are a field for innovation without changing the product.

Ask: Which part of our price hurts most emotionally, and can we change the mechanic rather than the amount? / Have we experimented with how people pay, not only how much? / Keeping the product fixed, what new way to charge, package or present it could we try?

Signs: Price objections that persist after discounts. A frequent-use product charged per use, or a rare-use product on subscription. Discounts dropped without being marketed. Allowances so generous nobody hurries. Buyers who must justify each line item to someone else.

P19 Price is a signal: the dead middle and the suspicious bargain.

Price tells buyers about quality and occasion, so the middle is often a dead zone, underpricing is the costlier error, and a low price without a believable reason reads as a warning.

Ask: Is our price in the dead zone between bargain and treat? / Who would buy more of this if it cost more? / If we are cheaper, do customers understand why, or will they imagine a hidden defect?

Signs: A mid-priced range squeezed from both sides. Discounts that reduce demand. Customers trading up for special occasions to something you don’t offer. A cheap offer that converts badly despite good value.

P20 Rigged is worse than unequal: make fairness visible.

People tolerate unequal outcomes and premium tiers but withdraw when the game feels rigged, when they suspect a price is personalised against them, or when they feel they are subsidising others; fixed, public, visible processes keep them in.

Ask: Could customers suspect we are personalising a price or outcome against them? / Does our pricing make some users feel they pay for other people’s value? / Where we ration by queue or charge a premium, would people resent who profits?

Signs: Backlash at surge or dynamic pricing. Light users cancelling bundles. Complaints that a premium lane or tier is “unfair”. Low trust in an opaque process even when outcomes are good.

P21 Change the context, not the mind.

It is easier to change the setting, interface, ordering or salience of a decision once than to persuade people one by one, and people rarely notice how much the choice architecture is steering them.

Ask: Where, when and through what interface does the customer decide, and could we move that? / What does the menu, form, default or first screen make easy, and what option is invisibly excluded? / Is the problem really lack of awareness, or is the thing we want chosen simply not salient at the moment of choice?

Signs: Persuasion campaigns that move attitudes but not behaviour. Supply that exists but is hard to see. Social awkwardness suppressing a choice (ordering something unusual, asking for help). One attribute (usually price) dominating because it is the only one shown.

P22 Create the option people didn’t know they had (and keep the plan B).

Much value comes from making a new option visible or available, including options people rarely use, because the existence of a fallback or of spontaneous access changes behaviour far beyond its direct use.

Ask: What option has no one ever been offered, or exists but is invisible? / What safety net would make customers willing to try the riskier or newer option? / Where do we require customers to plan ahead, and whom does that exclude?

Signs: Either/or decisions with no middle path. Low uptake of something people fear getting stuck with. Appointment-only access, short hours, or waiting lists that leak. Fallback channels being cut because few use them.

P23 Choice needs comparison and constraint.

People cannot choose well with one option or with too many; they need a small, comparable set they can place on a mental ladder, with guard rails, and sometimes they will pay to be relieved of choosing.

Ask: Are we offering one “optimal” answer, an overwhelming configurator, or a manageable comparative set? / Can a normal person build a mental map of our range? / Would a membership or default simplify customers’ lives?

Signs: Long ranges with overlapping products at similar prices. Abandonment at configuration or comparison pages. Customers asking “which one should I get?”. Recommendation tools that return one result.

P24 Control: people prefer what they chose.

Giving people meaningful control over how, when and where things happen raises satisfaction and motivation and moves blame, and control can be worth more than money.

Ask: Where could we let customers or staff choose (supplier, timing, place, method)? / Which non-cash currencies could we offer, and whom would they bring in? / Are we scrutinising this flexibility because it reduces output, or because people like it?

Signs: Imposed processes that experienced people quietly rework. Talent excluded by rigid hours or locations. Complaints directed at you for failures of a supplier you chose for the customer.

P25 Convenience beats conviction: friction carries meaning.

A great deal of behaviour happens because it is easy, not because it is believed in, and difficulty also signals that an action is unusual or unwise, so removing friction changes behaviour more than argument does.

Ask: What does the difficulty of this action tell the customer about whether they should do it? / Is our rule easy to cheat, hard to remember or socially isolating? / Could we do the form-filling or setup for people?

Signs: Good intentions that never convert. Long forms and multi-step setups. Low take-up of something people agree is sensible. Behaviour-change rules expressed as thresholds people must calculate.

P26 Effort creates value.

Effort invested in acquiring, making or committing to something raises its perceived value, memory and follow-through, and visible effort by the seller makes a result trusted, so removing all friction can make things forgettable or suspect.

Ask: Where has our convenience made something forgettable, trivial or untrusted? / Would a little deliberate participation raise perceived value or commitment here? / Are we judging our own staff by visible effort rather than results?

Signs: Frictionless products with low engagement or poor adherence. Results delivered so fast they are doubted. Low-price offers that look shoddy. Managers suspicious of work they cannot see.

P27 Preferences are found, not fixed: filters and sequences hide options.

People discover what they want while searching, so filter-and-match systems, sequential vetoes and one-at-a-time selection hide good options, push everyone onto the same few, and breed conformity.

Ask: What good options can our customers not find through our search, and is low demand really low visibility? / Is our first filter a proxy that is necessary but not sufficient, and is everyone else using the same one? / Are we choosing in isolation, or in sequence with vetoes, things that should be chosen together?

Signs: Search tools built on fixed criteria and price bands. Approval chains where each stage can only say no. Hiring one person at a time and getting clones. Everyone screening on the same credential. Options just past a threshold that never sell.

P28 Costly signalling: the cost of the message is part of the message.

People weigh a communication, a promise or a product by what it visibly cost the sender in money, effort, talent, restraint or risk, because cost is hard to fake; cheap messages convey little confidence and cheap complaints little distress.

Ask: What does the way we communicate (cost, medium, effort) say, independent of the words? / Having identified our best prospects precisely, are we then reaching them in the cheapest, most offhand way? / What could we stop offering that would make us look more committed or clearer?

Signs: Trust-dependent decisions sold through the cheapest channel. Targeting so precise that the medium carries no confidence. Broad menus that make nothing look special. Organisations reacting to cheap, high-volume complaints as if they were costly ones.

P29 Reputation is collateral (skin in the game).

People trust sellers, brands, intermediaries and recommenders who have something to lose and who make claims in front of an audience able to call them out; they pay less for the information than for the fact that someone can be punished.

Ask: What does the buyer think we would lose if we let them down, and can they see it? / Who personally loses face if this goes wrong? / Who else sees this claim, and could they refute it?

Signs: Buyers preferring a physical, reachable or named seller despite a cheaper anonymous option. Anonymous or one-to-one channels where nobody else sees the promise. Low trust in automated decisions with no accountable person.

P30 Satisficing: people buy to avoid disaster.

Under uncertainty people optimise for “never terrible” rather than “best on average”, so reliability, low variance and the absence of a catastrophic tail often beat peak quality.

Ask: What is the worst that can happen with our offer, and how visible is that risk to the customer? / Is our appeal our peak quality or our guaranteed floor, and which are we communicating? / Where are choices rushed, one-off or high-stakes, so that reliability matters more than excellence?

Signs: Customers choosing a familiar, unexciting option over a better-rated unknown. One small cue (a vague photo, a low rating) killing conversion. A product that is often brilliant and occasionally awful.

P31 Trust is a threshold, shared across the category.

People do not trade a small risk of a bad experience against a lower price; below a trust threshold they refuse to deal at all, and a few bad actors or experiences can taint a whole channel or category.

Ask: Is the real barrier to adoption the product, or fear of a service nightmare? / If every competitor did what we are doing, what would happen to trust in the category? / What share of bad experiences would make people abandon the channel?

Signs: Adoption stalled despite a better price. Customers citing one bad experience with someone else as their reason. Category-wide friction that every player adds and every customer hates.

P32 Status signalling, countersignalling and moving currencies.

Much consumption is a message to an audience and to oneself, so what people buy depends on who is watching, who else uses it and what currently confers status; the urge is constant but its currencies keep moving.

Ask: Who sees this purchase, and what does the buyer want them to conclude? / Is there a segment that wants the benefit without the signal, or the signal without the flaw? / Who do non-customers imagine our customers are, and does that put them off?

Signs: A brand losing appeal as it becomes ubiquitous or overtly branded. Buyers who want a product’s benefit but not its identity. Products that sell differently in anonymous cities and small towns. A visible customer group that deters others.

P33 Credibility comes from others.

Some qualities (classic, luxurious, funny, reliable) persuade only when other people say them and backfire when self-claimed, and success achieved despite unfashionable appearances is a stronger quality signal than success explained by image.

Ask: Are we asserting a quality that only has force coming from third parties? / Is this choice popular because it is good, or because it is fashionable or looks the part? / Who succeeds in our market against the grain, and what does that tell us?

Signs: Copy full of self-awarded superlatives. A best-seller nobody talks about. Spokespeople chosen for looking the part.

P34 Fame and brand earn permission (easy mode).

A famous, consistent brand earns forgiveness, lower price sensitivity, more effective direct response and licence to launch unfamiliar things, so brand building and performance marketing multiply each other, and abandoning a working long-run asset is vandalism.

Ask: Is our split between brand and performance based on effectiveness or on ease of proof? / Are we using our brand’s permission to do something new, or only improving specs? / Are we changing this asset because it is worn out, or because someone new wants their own mark?

Signs: Rising acquisition costs as brand spend falls. Customers unforgiving of small failures. Long-running assets dropped with each new leadership team. A strong brand used only for incremental launches.

P35 Relational beats transactional (homo economicus is your worst customer).

Regular customers believe they are in a relationship that earns goodwill over time, while finance treats every transaction as standalone; firms that recognise tenure and make small unconditional gestures win, and designing for the pure price-maximiser optimises for the worst customer.

Ask: Does our policy treat a ten-year customer like a stranger? / What do we do after the purchase that we have no short-term reason to do? / Which customer is our pricing and process built for, and can anyone pay a little more for better?

Signs: Complaints that open “I have been a customer for X years”. Rigid rules with no discretion for regulars. Post-purchase experience cut to the bone. Policies designed around preventing abuse by the few.

P36 People judge the person: human touchpoints multiply (brand quake).

When the product is too complex to judge, people judge the person or firm offering it, so a handful of human interactions shape the view of a whole organisation more than thousands of automated ones or any campaign.

Ask: Which rare human contacts carry disproportionate weight in how customers judge us, and what are those people empowered to do? / How are service staff measured: call length, or retention and the rescue of unhappy customers? / What cue are customers actually using to judge us, and are we spending on it?

Signs: Satisfaction tracking staff behaviour more than operational metrics. Service measured as a cost centre. Hidden phone numbers. Cuts to frontline pay while acquisition budgets grow.

P37 When it goes wrong: explain, own it, and fix life after the sale.

A failure handled impressively can leave a customer more loyal than no failure, a reason (even a thin one) turns anger into acceptance, and the experience after the purchase is where most firms quietly lose people.

Ask: When we fail customers, do we tell them why? / How do we behave when a problem is not our fault? / What happens the moment a customer leaves the happy path, and can they reach a human?

Signs: Delays with no explanation. Post-purchase contact handed to the cheapest channel. Policies that refuse responsibility for third-party failures. High acquisition spend alongside poor after-sales reviews.

P38 Imply, ask, don’t order.

Independent-minded people resist being told; asking for help, implying a norm, personalising obliquely and admitting uncertainty bring them with you better than instructions, mandates or displays of what you know.

Ask: Where are we compelling customers or staff when we could be asking? / Does our message sound like an order to an audience that prizes independence? / How would a less open or more privacy-sensitive person read this personalised message?

Signs: Mandatory changes met with resentment. Campaigns that harden opposition. Personalised messages that prompt complaints. A rollout where leaders project certainty nobody has.

P39 Let people pause, not cancel.

Offering a pause or pay-as-you-go option instead of a binary continue-or-cancel keeps customers who would otherwise never come back.

Ask: When customers want to stop, do we give them a reversible option that keeps the relationship? / What share of cancellers would have paused if asked?

Signs: Seasonal or life-event churn. Cancellation flows with only two buttons. Lapsed customers who never return although they were satisfied.

P40 Intermediaries carry hidden value.

Middlemen and human advisers survive disintermediation because they carry aggregated clout, pattern knowledge, reassurance and accountability, and can match odd supply with odd demand; services with people doing useful things are undervalued by tech-first investors.

Ask: What does the intermediary do beyond the visible transaction: clout, matching, reassurance, accountability? / Where do customers need someone they trust to say “don’t do that, do this instead”? / What is the worst part of this journey, and would people pay for a person to remove it?

Signs: Disintermediated markets where outcomes worsen for complex or high-stakes cases. Customers recreating advisers informally. A wealthy, less digital segment poorly served.

P41 Experience is the argument: the ratchet and frozen reference points.

Many of the best innovations are things people don’t want until they have tried them and then never give up, while people who haven’t tried judge a category by an experience that may be decades old; so measure repeat use and reversion, not first adoption, and invest in trial.

Ask: Of the people who try it, how many use it again, and how many go back to the old way? / Is resistance about the product, or about never having used it? / When did a typical customer last use this category, and what do they think it is like now?

Signs: Low stated interest but near-zero reversion among users. Slow early uptake read as rejection. Infrequent purchases replaced only on failure. Fast sign-ups with little repeat use.

P42 Habit and copying: new things need more selling, not less.

Because our defaults are “do what I did before” and “do what others do”, a genuinely new product or behaviour faces extra anxiety, needs visible peers and reassurance, and requires as much effort in selling as in inventing; norms then flip suddenly once enough people are seen to adopt.

Ask: Do adopters have to justify or explain this choice to others? / Can people see what their peers do in this category, and if not, how could we show them? / Is the barrier a stigma that would vanish once a visible minority adopted, and who holds the veto?

Signs: Products assumed to “sell themselves”. Behaviour that is rational but looks odd or rude. Low uptake in categories where nobody sees what others do. A minority with strong preferences quietly deciding the default.

P43 Who adopts first, and what surrounds it, sets the path.

Inventions spread only when complementary conditions line up and when work is reorganised around them, and the first users decide whether a product carries status or stigma.

Ask: Who would buy this now to be seen with it, and what would their being first signal to others? / Are we bolting a new technology onto an old process, and what would a firm starting today do? / Is this failing because the network or the surroundings aren’t ready, rather than because the idea is bad?

Signs: A good technology with stigmatised early users. New tools with no productivity gain. Networks judged on early losses. Infrastructure rejected because benefits can’t be proven in advance.

P44 Make the first step small, obvious and reversible.

A small early step changes the default for the big later decision more cheaply than subsidising the big decision, people accept far more when something is temporary or removable, and products that do one thing make the new behaviour obvious.

Ask: What small prior commitment would make our desired choice the obvious next step, and are our rules blocking it? / Could we make this reversible, mobile or time-limited to reduce opposition? / Does our new product do so much that nobody knows what to use it for first?

Signs: Subsidies aimed at the final purchase while the enabler is neglected. Opposition to permanent change. Feature-rich launches with confused users.

P45 Remove the hidden obstacle.

Many products succeed by removing an emotional, social or image barrier rather than by improving function, because the obstacle is often not the product but what using it says or where it can be used.

Ask: Who would buy this but for how it looks, feels or what it says about them? / Where can’t the product currently be used or shown?

Signs: Strong functional ratings with weak sales in a segment. Products used privately but never publicly. A category image that embarrasses part of the market.

P46 Make it attractive before making it compulsory.

Persuade first, then incentivise, and legislate last; give people a selfish reason to do the collectively good thing, and design messages for the undecided observer, not your own side.

Ask: Have we tried to make people want this before mandating it? / What does the individual personally gain from doing the good thing? / Is this message designed to impress our own side or to persuade the people we need, and is there an adjacent problem both sides hate?

Signs: Mandates and penalties proposed as the first move. Pro-social campaigns that appeal only to sacrifice. Messaging that insults non-adopters. Debate stuck on the most divisive aspect while agreed problems go unsolved.

P47 A surprising signal resets expectations.

Behaviour runs on assumptions about the future, so a cheap, surprising signal that the rules have changed, or a disruption that gives people a story for behaving differently, can shift behaviour more than a large incentive.

Ask: What unthinking assumption about the future is holding current behaviour in place? / What habit-bound change could we introduce now that people have a reason to behave differently?

Signs: Large incentives with little effect. A disruption (a move, a price shock, an outage, a reorganisation) that has already broken habits.

P48 Reverse benchmarking.

Instead of copying what the best competitor does well, find what the whole category does badly, ignores or treats as an afterthought, and be disproportionately good at that while staying competent elsewhere.

Ask: What does everyone in this category neglect, not measure, or treat as second-class (a customer group, a moment before or after the core service, aesthetics)? / What does our best-in-class non-competitor do that we don’t? / What did the best competitor fail to do?

Signs: A category where all players publish the same metrics. A customer minority universally underserved. A moment in the journey everyone treats as overhead. Customers praising a rival for something unrelated to the core product.

P49 Logic converges: conformity leaves a vacancy.

When every firm optimises the same metrics with the same logic and the same map of the market, they become alike and compete only on price, which leaves a cheap opening for whoever does the opposite, uses the abandoned channel, or notices what the shared model gets wrong.

Ask: Which metrics did we inherit from consultants or analysts that every competitor also uses? / What has the whole industry recently abandoned as old-fashioned or unscalable that customers might now value because it is rare? / What model of the market does everyone rely on, and what does it leave out?

Signs: Competitors indistinguishable except on price. A channel every rival left. One dimension (usually price or a headline spec) dominating comparison. Plans that can be predicted from the industry playbook.

P50 The opposite of a good idea can be another good idea.

Unlike in physics or school maths, in psychology two contradictory strategies can both win, every trend creates a counter-trend, and the dangerous place is usually the compromise in the middle.

Ask: What is the exact opposite of our category’s best practice, and which segment would love it? / What dominant trend is our category riding, and what human need has it vacated? / Are we stuck in a compromise between two coherent opposite strategies?

Signs: A category converging on one model (fully automated, or fully premium). Nostalgia or backlash among a segment. A middle-of-the-road offer losing to both extremes.

P51 Want what others don’t.

Markets price in widely shared preferences and dislikes, so the advantage lies in attributes most people avoid but some segment doesn’t mind, and in the slightly less famous option in winner-take-all markets.

Ask: What do most customers mark down that some segment would value or not mind? / Could negatives be made searchable, so people can find what they don’t mind? / Does our channel concentrate rewards on the most famous option, and where do we sit?

Signs: Inventory that is discounted for a feature only some people dislike. Off-peak capacity priced as if worthless. Divisive reviews. A crowded leader with a nearly-as-good, overlooked alternative.

P52 Delight attributes: the gratuitous touch (Kano).

Features don’t add satisfaction in proportion; small, tangential, apparently unnecessary touches often create more preference than improving the core, and they are the first thing efficiency drives remove.

Ask: Which of our features are table stakes, which are ignored, which do customers dislike, and what small odd thing could delight? / What tiny feature would people talk about, and who owns the “trivia” that falls between departments? / What gratuitous thing could we do that no competitor does and no spreadsheet would justify?

Signs: Products competing on specs nobody notices. Cost reviews targeting small “unnecessary” touches. Customers talking about a minor detail more than the core offer.

P53 Design for the edge.

Extreme, impaired and neglected users show where a product needs to go, and solutions designed for them often serve a much larger mainstream market.

Ask: Who struggles most with this, and would fixing it for them help everyone? / What everyday situations put all users temporarily in the struggling user’s position? / Who uses this in the most extreme way, and what are they improvising?

Signs: Small type, small buttons, few seats or toilets. Workarounds built by heavy users. A large, predictable segment ignored as unfashionable.

P54 Cooperate on what no brand can fix alone.

Some barriers belong to the whole category and brand-by-brand marketing cannot fix them (and can deepen them), while partnerships that pair one firm’s innovation with another’s trust are under-used because they are cheap.

Ask: Is the barrier our brand or the category’s reputation, and is our competitive message reinforcing the category’s biggest objection? / Who already has the trust we lack, and what problem of theirs could we solve? / Is this move adjacent, or two steps removed from what we’re trusted for?

Signs: Every competitor advertising the same reassurance. Category-wide objections in research. A trusted partner who could vouch for you. Partnerships with no internal owner.

P55 Fat tails: a few ideas create most of the value.

Most of the value in marketing, innovation and R&D comes from rare, unpredictable breakthroughs that pay for decades, so judging every spend on short-term, granular returns or paying by the hour underfunds the ideas that matter.

Ask: Is anyone still looking for treasure, or is everyone mining measurable 3 to 7% gains? / How long do we credit an idea’s value, and would our metrics have rewarded the idea still paying off ten years later? / How does the person with a large recurring idea get rewarded?

Signs: All activity justified by short-term attribution. Hourly or cost-plus pay for creative and strategic work. No budget line that is allowed to fail.

P56 Explore as well as exploit (optionality over optimality).

Systems optimised purely for efficiency get stuck at a local maximum and break when conditions change, so a share of apparently wasteful exploration and redundancy is essential, and paths that keep options open are worth a small efficiency loss.

Ask: Who are our scout bees, and would a quarterly-reporting mindset cut them? / Are we choosing the locally best option at the cost of future possibilities? / Do we depend on a single input, channel, supplier or process?

Signs: Every person and pound allocated to known-good activity. One channel or supplier carrying the business. Efficiency programmes that remove slack and redundancy.

P57 Increase your surface area for luck.

Success is probabilistic, so fame, slack, odd meetings, openness to ideas from unexpected sources and protected rule-breaking increase the chance of lucky breaks, and the people closest to a problem will find uses for your idea you never imagined.

Ask: Are we rejecting an idea because of where it came from? / Which part of our activity is allowed to break the usual rules, and what did our best innovations have to hide to survive? / Are we cutting the fun, social, publicity and random activities that create lucky breaks?

Signs: Tightly planned programmes with no slack. Publicity and events cut in bad months. Innovations that only exist because someone hid them from process.

P58 Test, don’t argue (two-way doors).

Most business decisions are reversible and should be tested cheaply rather than debated, including deliberately counter-intuitive options, because a surprising result is worth far more than confirming the obvious; heavy rigour belongs to one-way doors.

Ask: Can we undo this cheaply, and if so why are we still debating it? / What is the most counter-intuitive thing we could test cheaply here, such as raising the price instead of cutting it? / What is the cheapest test of whether people will act, not just say they like it?

Signs: Long debates over reversible choices. Business cases demanded for small pilots. Tests that only compare minor variants of the consensus.

P59 The answer to a different question (anomaly, analogy, trade-off).

Many breakthroughs start as a fluke, a side effect, a “failed” result or a practice from another field that turns out to answer a different question, so look for anomalies, borrow from other categories and refuse assumed trade-offs.

Ask: What have we shelved as a failure, side effect or bug that would be a feature for another customer or context? / Which other industry has already solved this? / Which either/or is everyone assuming, and does someone point out an anomaly here and get ignored?

Signs: Rejects and side effects nobody revisits. A problem only ever discussed by insiders. A trade-off treated as a law.

P60 An element of stupidity.

Problems that yield to logic have mostly been solved, so an idea with a slightly ludicrous element is weak evidence it is untried, while a logically obvious gap often has no market in it.

Ask: If this idea makes perfect sense, why hasn’t someone already done it? / Is there something about our idea that is a bit stupid, comical or weird, and is that why nobody else would try it? / What would our research process have said about our own best past idea?

Signs: Shortlists containing only defensible ideas. Innovation pipelines filtered by concept tests. Successful category oddities that “make no sense”.

P61 Markets discover; reasons come later.

The main value of markets and businesses is discovering things nobody could have justified in advance, which requires loose definitions of success and permission to try ideas before their rationale exists.

Ask: Does this organisation require an idea to win an argument before it can be tried? / Is our definition of success so narrow that it would reject a valuable opportunity that wasn’t in the plan? / Are we refusing ideas because we can’t yet explain why they would work?

Signs: Business-case templates applied to every experiment. Opportunities outside the plan declined. Things that work without an explanation being stopped.

P62 The doorman fallacy.

Defining a role, channel or process by its most visible function and then automating or cutting it destroys the unmeasured functions it also performed.

Ask: If this role, step or channel vanished, what would we lose that is not in its job description or KPI? / Who else benefited from it, and how? / If we automated this output, what understanding would we stop building?

Signs: Automation business cases built on headcount alone. Channels cut because cheaper ones exist despite much higher conversion in the costly one. Roles described in one line. Post-cut declines in loyalty, conversion or service that nobody links back.

P63 Options become obligations (Joni Mitchell syndrome).

New technologies arrive as welcome options; once the savings are noticed the old alternative is removed and the option becomes compulsory, and because we compare only on the new thing’s distinctive features, what the old way did well is forgotten.

Ask: Is this new channel additive, or are we about to remove the old one, and who can’t use the new one? / What did the channel or process we are replacing quietly do well? / Are we comparing only on the features where the new thing is better?

Signs: A digital channel launched “as a choice” with a plan to retire the alternative. Older or less digital customers stranded. Effectiveness (conversion, response, returns) never compared between old and new.

P64 Costs bark, opportunity costs stay silent.

Organisations over-manage visible costs and losses and ignore larger invisible ones (revenue never earned, failures prevented, work pushed onto customers), so the people who cut costs take credit for savings while nobody is charged for the value destroyed.

Ask: What are we failing to do that costs us revenue nobody tracks? / Who gets credit for this saving, and who would ever be blamed for the sales, complaints or innovation it costs? / How many hours of customer labour does buying from us require, and who used to do that work?

Signs: Discounts as the first response to slow sales. Cost programmes with no measure of value lost. Unlit, unanswered, unstaffed moments of demand. Systems judged on their rare failures while the failures they prevent go uncounted. More forms and self-service for customers presented as efficiency.

P65 Quantification bias (the McNamara fallacy).

Organisations favour what can be counted and what shows up quickly, so they over-invest in acquisition, cost-cutting and technology and under-invest in retention, service, brand and the human feelings that drive behaviour.

Ask: Are we spending where the money is effective, or only where we can prove it is effective? / Which of our KPIs could improve while customers feel no better, and what do they value that has no metric? / Who benefits from the metrics we use, and did we ask what else this technology budget could buy?

Signs: Acquisition teams better paid than retention or service. Satisfaction data aggregated so extremes vanish. Feedback loops of a quarter in a business where effects take years. Large technology budgets approved easily while small human spend is policed.

P66 Efficiency and opportunity are different mindsets.

Finance’s efficiency mindset and marketing’s opportunity mindset both matter, but they suit different functions; when procurement logic runs marketing, HR, R&D or strategy, or when new technology is framed only as a way to cut cost, firms lose the opportunities that never show on the balance sheet.

Ask: Which of our functions is being judged with efficiency metrics when its job is to find what competitors aren’t finding? / Are we deploying this technology to remove people and costs, or to make people more valuable, and who is framing the decision? / Is a customer-lens mind in the room when this decision is made, and does it have power?

Signs: New technology sold to the board on headcount. Premium brands skimping on the touchpoints customers pay for. A lone marketer who goes quiet in board meetings. Strategy documents that are really cost plans.

P67 Winning an argument is not solving a problem.

Organisations reward people for building logically defensible cases, which is a different and often opposite skill from solving problems, and they judge decisions by the rigour of the reasoning rather than by the outcome.

Ask: Are we choosing this because it will work or because it is defensible? / Are we demanding advance proof for ideas that can only be proven by trying? / Would we fire someone for being unimaginative as readily as for being irrational?

Signs: The person with the spreadsheet wins every meeting. Options pre-narrowed to three, with one straw man. Theory-led policy with no feedback from practice. Good ideas dropped because they couldn’t be justified in advance.

P68 Defensive decision-making: blame without credit.

Inside institutions, decisions are shaped less by expected value than by whether the decision-maker can be blamed, and people who bear the blame for failures but capture little of a breakthrough’s upside rationally stop trying for breakthroughs.

Ask: Is this decision being made because it is best, or because nobody could be blamed for it? / Is this framework helping the decision, or protecting the decision-maker? / How does the person who has a large recurring idea get rewarded?

Signs: Default choice of the biggest, safest supplier. Elaborate processes for creative work that pretend to be linear. Big wins with no reward and small failures punished. Time-based fees for value-based work.

P69 The tantric sector: process replaces judgement.

Compliance, legal, HR and procurement functions grow by slowing things down; their real cost is the useful things that never happen, because demands for comparability, pre-justification and legal safety screen out new, one-of-a-kind ideas and replace discretion with procedure.

Ask: Would our procurement or approval process even let the best idea bid, and would a better method be legal under our current procedures? / Is this rule producing a good outcome here, or protecting someone from blame? / Is our only response to a problem either nothing or a formal process?

Signs: Long sign-off chains for small decisions. Spend mandated by compliance that nobody scrutinises. Objectors heard, silent supporters uncounted. Unbudgeted ideas with no route to funding. Lack of progress in a field that is regulatory rather than technical.

P70 Long horizons: profit is a constraint, not the goal.

A business must make a profit to survive, but its purpose is to discover new value for customers, staff and others; firms that can think across several time horizons at once, rather than to the next quarter, serve customers better and resolve apparent conflicts between stakeholders.

Ask: What is this organisation actually optimising for, and over what time horizon? / Which stakeholder is being quietly sacrificed for a spreadsheet gain, and is there a move that serves all of them? / Could this decision be approved in a quarterly-reporting company, and if not, and it’s good, is that our opportunity?

Signs: Decisions justified by the next reporting period. Cost cuts that move pain from investors to customers or staff. Benchmarking only against firms whose data is easy to get.

P71 Fix the system, not the part or the person.

Most failures come from system design in which every individual acts rationally, and many problems disappear when you look at the whole flow of use rather than one component; a rule that looks slower for each person can make the whole system faster.

Ask: What does our payment or measurement system reward that we would never ask for directly? / What does the customer do before and after using our product, and which step could disappear? / Would a rule that slows each individual raise the throughput of the whole?

Signs: Repeated failures blamed on individuals. Each department optimising its own step. Congestion at handovers. Incentives that pay for activity rather than outcomes.

P72 People respond to the rule, not its intent.

Rules, prices and interventions must be designed for how people will actually behave next, because behaviour drifts, people game systems, effects decay as they adapt, and those who learned the old rules feel a change as theft.

Ask: If we charge for, ban or remove this, what will people do instead, and what will that cost? / Which natural social cost has our product or policy removed, and what behaviour did that release? / Who learned to play by the old rules, and how will they experience the change?

Signs: A saving that creates a new cost downstream. Interventions that worked at launch and faded. Gaming of targets. Fury from loyal customers at a reform that is economically sensible.

P73 Follow the gains to the bottleneck.

Whenever useful activity depends on a location, a platform or another chokepoint, the gains from everyone else’s improvements flow to whoever owns the bottleneck; meanwhile some spending sustains goods others enjoy and some only competes for scarce ones.

Ask: Who owns the bottleneck in our value chain, and who captures the gains from improvements we make? / Of every pound a customer pays, how much reaches the person creating value? / Does our product or space get better when more people use it, and is that value priced?

Signs: Margins steadily absorbed by a platform or landlord. Fees that look like tolls rather than prices for a service. Channels where all traffic passes through one gatekeeper. Social or shared spaces closing despite heavy use.

P74 Deploy slow deliberately: some value lives in the duration.

Compressing a process destroys value whenever the worth of the activity is in the doing rather than in the output, and because every optimisation model assumes faster is better, nobody ever asks which processes should be deliberately slowed.

Ask: Is the value of this activity in its output, or in what doing it forces us to notice? / If we halved the time this takes, what would stop happening that nobody has written down? / Which of our processes would customers pay more for if they were visibly slower? / What would the slow version of this new technology look like, and would anyone want it?

Signs: An automation or AI business case that measures the artefact and not the thinking the work produced. Premium products sold only through the cheapest available channel. Batching or parallel processing introduced into something customers bought as handmade. Target audiences defined as the people cheap enough to reach rather than the people likely to buy. Teams whose calendars are cut into intervals too short to think in.

P75 Design the protocol, not just the tool.

A technology that each person adopts optimally for themselves can leave everyone worse off, because the costs of one person’s speed fall on everybody else; the fix is a convention governing how the thing is used, and almost nobody designs one.

Ask: Who absorbs the cost when one of our users goes faster, and do they get any of the saving? / Would this be better for everyone if we all agreed when to use it, rather than each deciding? / Is there a default delay, batching window or agreed day that would make this tool collectively better? / Who would have to set that convention, and does anyone own it?

Signs: Universal adoption of a faster channel that leaves everybody busier. Complaints about a tool that nobody is willing to stop using. Costs that appear on the recipient’s side of a transaction and never on the sender’s. A behaviour defended as individual choice that visibly changes the experience of non-adopters.