People judge experiences not by the total or average but by the most intense moment and how it ends — design the peak and the ending, and the memory takes care of itself.
Model #0146Category: Psychology & BehaviorSource: Kahneman / Redelmeier / KatzDepth to apply:
In the early 1990s, Daniel Kahneman and his collaborators conducted a series of experiments that overturned a fundamental assumption about human memory and judgment. The assumption — held by economists, psychologists, and anyone who has ever designed a customer satisfaction survey — was that people evaluate experiences by integrating all the moments they contain. A longer painful experience should be judged as worse than a shorter one. A dinner with eight excellent courses and one mediocre dessert should be rated nearly as well as one with nine excellent courses. The brain, it was assumed, functions like an accountant: it tallies the pleasure and pain at each moment, sums the total, and produces a global evaluation proportional to the sum. Kahneman demonstrated that this assumption is comprehensively wrong. People do not evaluate experiences by their total or their average. They evaluate them by two moments: the most intense point — the peak — and the final moments — the end. Everything else is largely discarded. The duration of the experience, the cumulative total of pleasure or pain, the average intensity across all moments — none of these factors reliably predicts how people remember and judge what happened to them. The peak and the end do.
The landmark demonstration came from a colonoscopy study conducted by Kahneman, Donald Redelmeier, and Joel Katz, published in 1996. Patients undergoing the procedure — which was, at the time, performed without sedation and was genuinely painful — were asked to report their pain intensity in real time, moment by moment, throughout the examination. The researchers then compared these real-time reports with the patients' retrospective evaluations of the total experience. The results violated every intuition about rational assessment. Patients whose procedures lasted longer — sometimes significantly longer — rated the experience as less painful overall, provided the final minutes involved reduced intensity. A patient who endured twenty minutes of moderate pain with a gentle tapering at the end remembered the experience as less terrible than a patient who endured eight minutes of the same moderate pain but whose procedure ended abruptly at a moment of high intensity. The longer procedure contained more total pain. It was remembered as less painful. The critical variable was not how much pain was experienced but how the experience ended. The ending overwrote the record.
A subsequent experiment — the cold-water study — distilled the finding to its purest form. Participants immersed a hand in 14°C water for sixty seconds. In a separate trial, they immersed the same hand for sixty seconds at 14°C followed by thirty additional seconds at 15°C — still painful, but slightly less so. When asked which trial they would repeat, participants chose the longer trial. They preferred ninety seconds of pain to sixty seconds of the same pain, because the ending was marginally better. The result is as counterintuitive as any finding in behavioural science: people will choose more suffering if the ending is gentler, because the ending — not the total — is what the remembering self encodes.
Kahneman called this phenomenon the Peak-End Rule, and its implications extend far beyond medical procedures. The rule reveals that human memory is not a recording device — it is an editing system. The remembering self, as Kahneman would later call it, does not replay the experience in full. It constructs a summary from two data points: the moment of greatest intensity and the final moment. This summary then becomes the memory, and the memory — not the experience itself — determines all subsequent judgments: whether the person would repeat the experience, whether they would recommend it to others, whether they feel satisfied or cheated, whether they return as a customer or never come back. The experiencing self lives through every moment. The remembering self keeps only the peak and the end. And it is the remembering self that makes decisions.
The rule's power becomes visible when you examine the domains where experience quality determines business outcomes. In product design, the Peak-End Rule explains why Apple invests obsessively in the unboxing experience — the final moment before the customer begins using the product. The unboxing is the end of the purchase journey, and it creates a peak of anticipation and delight that disproportionately shapes how the customer remembers the entire buying decision. In hospitality, the rule explains why the Ritz-Carlton trains staff to deliver a warm, personalised farewell — because the departure moment is the end that will be remembered when the guest decides whether to return. In presentations, the rule explains why the most memorable TED talks build to an emotional climax and close with a resonant final line — because audiences remember the peak moment and the closing, not the fifteen minutes of exposition that connected them.
The corollary that makes the Peak-End Rule operationally dangerous is what Kahneman called duration neglect — the near-complete insensitivity of retrospective evaluations to the duration of the experience. A two-hour flight delay that ends with a sincere apology and a voucher is remembered more favourably than a thirty-minute delay that ends with an automated announcement and no acknowledgment. A six-month onboarding process that culminates in a celebrated first-win milestone is rated as more effective than a three-week onboarding that ends with an abrupt transition to business-as-usual. The durations are vastly different. The retrospective judgments do not track them. They track the peak and the end — which means that any experience designer who optimises for average quality across all moments is solving the wrong problem. The leverage is concentrated in two moments, and everything between them matters far less than intuition suggests.
Why does the ending carry such outsize weight? Partly because the final moment is the most recently experienced — recency ensures it is the most cognitively accessible when the remembering self constructs its summary. Partly because the ending carries narrative significance: the human mind is a story-processing engine, and stories are defined by their conclusions. A story that ends well is a good story regardless of the struggles in the middle. A story that ends badly is a tragedy regardless of the joys that preceded it. The Peak-End Rule is not merely a memory bias — it is a reflection of the narrative architecture through which humans interpret all temporal experience. We do not remember sequences of moments. We remember stories. And stories are defined by their climax and their resolution.
The Peak-End Rule belongs in Tier 1 because it governs how every experience your company creates — every customer interaction, every employee milestone, every investor meeting, every product touchpoint — will be remembered and judged. It is the mechanism that converts momentary interactions into lasting impressions, and lasting impressions into repeat behaviour, referrals, retention, and lifetime value. The companies that understand the rule design their experiences backward from the ending, ensuring that the final moment is intentionally crafted rather than accidentally defaulted. The companies that do not understand it distribute their effort evenly across the experience — investing equal resources in moments that will be forgotten and moments that will determine whether the customer ever comes back.
Section 2
How to See It
The Peak-End Rule is operating whenever the remembered quality of an experience diverges from its objective average — when a mostly excellent experience is remembered as mediocre because it ended poorly, or when a mostly painful experience is remembered as tolerable because it ended gently. The diagnostic signature is a mismatch between what happened across the full duration and how it is later described. When post-experience ratings seem disproportionately influenced by the final moments rather than the totality, the Peak-End Rule is shaping the judgment.
Duration neglect is the companion signal. When a customer rates a fifteen-minute support call as "quick and helpful" and rates a six-minute call on a different issue as "painfully slow" — despite the first call being more than twice as long — the evaluations are tracking peak and end quality, not elapsed time. The most common manifestation in business: teams that worked on a six-month project remember it as "that amazing sprint" because the launch was celebrated, while teams that worked on a three-month project remember it as "that endless slog" because it ended with a whimper.
You're seeing Peak-End Rule when a customer who experienced thirty minutes of efficient, pleasant service remembers the entire interaction negatively because the final sixty seconds involved a billing error — or when a customer who waited too long, encountered friction, and dealt with a system crash remembers the experience positively because the support agent who resolved the issue was exceptionally warm and the resolution arrived at the very end. The length and average quality of the experience have been overwritten by its peak and its conclusion.
The most reliable early warning sign is a disconnect between process metrics and satisfaction metrics. When your operational data shows consistently high throughput, low error rates, and fast resolution times — but your NPS or CSAT scores tell a different story — the Peak-End Rule is the most likely explanation. The process metrics measure the experiencing self's journey. The satisfaction metrics measure the remembering self's summary. If the process is good but the peak was absent and the ending was accidental, the remembering self has nothing favourable to report — and the satisfaction score reflects the memory, not the process.
A related signal appears in competitive contexts. When a competitor with an objectively inferior product or service consistently outperforms you on satisfaction and loyalty metrics, examine their peaks and endings before auditing your average quality. The competitor may be delivering a worse average experience but a better-remembered experience — and the remembered experience is the one that drives repeat behaviour and referral.
Product & Customer Experience
You're seeing Peak-End Rule when a SaaS product receives consistently high NPS scores despite a clunky middle-of-workflow experience — because the onboarding flow creates a moment of delight when the user achieves their first result (peak) and the end-of-session summary delivers a satisfying sense of progress (end). Conversely, a competitor with a smoother average workflow receives lower satisfaction scores because its onboarding ends with a generic dashboard that creates no emotional punctuation. The product with the rougher average experience is remembered as better because its peak and end were designed, while the smoother product's were not. Satisfaction surveys are not measuring the experience. They are measuring the memory of the experience — and that memory is authored by two moments, not by the full timeline.
Presentations & Communication
You're seeing Peak-End Rule when an audience rates a forty-five-minute keynote as "incredible" despite twenty minutes of meandering content in the middle — because the speaker opened with a provocative story, built to an emotional climax at minute thirty-two, and closed with a memorable call to action. A second speaker delivers uniformly solid content for forty-five minutes but ends with "and that's all I have — any questions?" The first speaker is remembered as transformational. The second is remembered as competent but forgettable. The audience is not averaging the quality across all moments. They are retrieving the peak — the emotional high — and the end — the closing line — and constructing their evaluation from those two data points. The twenty minutes of mediocre content in the first talk are not in the memory that generates the rating.
Employee Experience
You're seeing Peak-End Rule when departing employees who had mixed tenures — some great years, some difficult ones — rate their overall experience based almost entirely on their final months. An employee whose last project was a visible success, whose manager conducted a thoughtful exit conversation, and whose departure was marked with a genuine team farewell remembers a positive tenure. An employee with an identical mixed history whose final months involved a reorganisation, an impersonal exit process, and a last day that felt administrative rather than personal remembers a negative tenure. The Glassdoor review, the referral willingness, the alumni network engagement — all are shaped by the ending, not the average. Companies that invest in offboarding as carefully as onboarding understand the Peak-End Rule. Companies that treat departures as administrative functions are writing their Glassdoor reviews by neglect.
Sales & Negotiation
You're seeing Peak-End Rule when a prospect who endured a long, complex sales cycle — multiple demos, procurement reviews, security questionnaires — remembers the experience favourably because the deal closed with a celebratory kickoff call, a personalised welcome package, and a named customer success manager who made first contact within twenty-four hours. A second prospect with an identical sales cycle but whose deal closed with a form email and a generic onboarding link remembers the process as "exhausting." Both cycles contained the same friction. The first ended with intentional design. The second ended with operational default. The memory that determines whether the customer becomes a reference, expands their contract, or tells peers about the experience is shaped by the ending — not by the months of demos that preceded it.
Section 3
How to Use It
Decision filter
"For every experience I design — product, presentation, meeting, customer journey — I ask: what is the peak moment and what is the ending? If I cannot identify an intentionally designed peak and a deliberately crafted conclusion, the experience is leaving its most memorable moments to chance. I design backward from the end."
As a founder
The Peak-End Rule is the highest-leverage framework for customer experience design because it concentrates your investment where memory formation actually occurs. Most founders distribute experience quality evenly — investing proportionally in every touchpoint across the customer journey. The Peak-End Rule says this is wasteful. The moments that determine retention, referral, and lifetime value are the peak and the end. Everything else is filler that the remembering self will largely discard.
Design your product's first-use experience to contain a deliberate peak — a moment where the user achieves something meaningful and the interface celebrates it. Slack's first-message-sent moment, Stripe's first-successful-payment moment, Notion's first-published-page moment — each is an engineered peak that creates the emotional anchor the user will associate with the product. Then design the session-end experience with equal care: a progress summary, a motivational prompt, a preview of what comes next. The default end of most products is the user closing the browser tab. That is an accidental ending that leaves the memory to chance. A designed ending — even a simple one — converts a forgettable session into a memorable experience.
Apply the same logic to your fundraising process. The pitch meeting's peak should be a single, vivid demonstration of the product or market insight that creates genuine surprise. The meeting's end should be a clear, confident close — not a trailing "so, yeah, any questions?" Every investor meeting is an experience that will be remembered by its peak and its end. Design both deliberately.
As a leader
Inside organisations, the Peak-End Rule governs how employees remember meetings, projects, quarters, and tenures. A team that ships a product after months of grueling work will remember the experience based on the most intense moment — the launch celebration, the first customer win, the CEO's public recognition — and the ending — the retrospective, the transition to the next project, the acknowledgment (or absence) of the effort. If the peak was a celebrated launch and the end was a thoughtful retrospective that named specific contributions, the team will remember the sprint positively and volunteer for the next one. If the peak was the same celebrated launch but the end was an immediate reassignment to the next project with no acknowledgment, the team will remember exhaustion — because the ending overwrote the peak.
The rule applies with particular force to how you close meetings. The final two minutes of any meeting determine how participants remember the entire hour. End with clarity — decisions made, owners named, next actions specified — and the meeting is remembered as productive. End with ambiguity — trailing conversation, no summary, uncertain next steps — and the same meeting is remembered as a waste of time, regardless of the quality of discussion that preceded it.
The same principle governs how you close quarters, close fiscal years, and close eras of the company's history. A quarter that ends with a team celebration of specific wins — even modest ones — is remembered as a successful quarter. A quarter with identical results that ends with an immediate pivot to next quarter's targets is remembered as a grind. The performance was the same. The ending was different. And the ending authored the memory that shapes the team's energy, motivation, and willingness to push through the next cycle.
As an investor
The Peak-End Rule governs how founders remember your firm after the fundraise — and that memory determines whether they choose you for the next round, refer other founders, or warn them away. The peak of the fundraising experience from the founder's perspective is the moment of commitment — the partner's handshake, the term sheet delivery, the call that says "we're in." Most firms handle this moment well because it aligns with their own interests. The ending, however, is where most firms fail. The post-close experience — the legal process, the wire transfer, the first board meeting — is often handled by operations staff with no awareness that this is the ending the founder's remembering self will encode. A wire transfer that arrives three weeks late, a legal review that feels adversarial, or a first board meeting that feels perfunctory becomes the ending that overwrites the partner's enthusiasm from the pitch process.
The best firms design the post-close ending deliberately: a same-day congratulatory note, a rapid and founder-friendly legal process, a first board meeting structured to add immediate strategic value. These endings are not expensive to deliver. But they are the moments that determine whether the founder's memory of the fundraise — and by extension, of the firm — is positive or negative. The firms with the strongest founder networks are not necessarily the ones that offer the best terms. They are the ones that engineer the best endings.
Common misapplication: Concluding that only the peak and end matter and neglecting everything else. The Peak-End Rule describes how experiences are remembered, not what constitutes a good experience. A customer who encounters persistent friction throughout the journey may abandon before reaching the end — in which case there is no end to design. The rule operates on completed experiences. The middle of the experience must be good enough to retain the person through to the end, where the rule's leverage applies.
Second misapplication: Treating the rule as permission to end every interaction with an over-the-top flourish. The Peak-End Rule describes how memory selects moments, not a mandate to manufacture theatrical conclusions. A team meeting that ends with an awkward forced cheer or a customer email that closes with three paragraphs of effusive gratitude will register as performative — and inauthenticity can itself become a negative peak.
Third misapplication: Manufacturing artificial peaks that feel manipulative rather than genuine. A contrived celebration, an exaggerated thank-you, or a performative gesture registers as inauthenticity — which can become a negative peak that the rule then amplifies. The most effective peaks are genuine moments of value delivery: the user achieving a meaningful result, the customer receiving unexpected care, the employee being recognised for a specific contribution. Design the conditions for the peak; do not fabricate the emotion.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The founders below demonstrate that the Peak-End Rule is not an abstract cognitive principle but an operational design framework. In each case, the leader — whether consciously or intuitively — engineered the peak moment and the ending of a critical experience to disproportionately shape how customers, audiences, or stakeholders remembered the entire interaction. The pattern is consistent: the leaders who build the strongest emotional loyalty are not those who deliver the highest average quality across every moment but those who concentrate extraordinary quality at the moments memory actually encodes.
The five cases span consumer technology, e-commerce, hospitality, retail, and entertainment — demonstrating that the Peak-End Rule operates wherever an experience unfolds over time and a memory of that experience determines future behaviour. In every case, the critical design decision was not what to include in the middle of the experience but how to engineer the peak and what to leave the audience with at the end. What unites these leaders is not industry or era but a shared instinct — or in some cases, a deliberate methodology — for identifying the two moments that would author the customer's memory and then investing disproportionate creative energy at precisely those points.
Jobs understood the Peak-End Rule at an intuitive level that preceded the academic research. Every Apple product launch was structured as a peak-end experience. The keynote built through a deliberate arc — problem framing, product revelation, feature demonstrations — toward a single peak moment designed to produce involuntary audience reaction. The iPhone launch in 2007 had its peak at the moment Jobs first scrolled through the contact list with his finger on the touchscreen — the gesture that made the audience gasp because it violated every expectation about how phones worked. The end was equally deliberate: Jobs's "one more thing" closing became a signature because it delivered a second emotional peak at the precise moment the audience expected the experience to conclude, ensuring the final memory was one of surprise and delight. Beyond keynotes, Jobs applied the rule to the product itself. The unboxing experience — the slow reveal, the precision fit of the packaging, the device resting in its tray like a jewel — was the engineered end of the purchase journey. Customers who had navigated the Apple Store, waited in line, and completed the transaction remembered the experience through the unboxing peak and the moment they powered on the device for the first time. The middle of the purchase process — the wait, the payment, the receipt — was forgettable by design. The peak and the end were unforgettable by design.
Hsieh built Zappos into a billion-dollar company by applying the Peak-End Rule to customer service with a rigour that most companies reserve for product engineering. Zappos's legendary service — free shipping both ways, 365-day return policy, customer support representatives empowered to spend unlimited time on calls — was not designed to optimise average satisfaction across the journey. It was designed to create peaks and endings that customers would remember and share. The peak was the moment of unexpected generosity: the overnight shipping upgrade the customer didn't request, the handwritten thank-you note in the box, the support agent who stayed on the call for four hours because the customer needed someone to talk to. The end was the effortless return — a process so painless that it transformed a potentially negative moment (returning a product) into a positive one (the company trusts me). Hsieh understood that the return experience was the end of the customer journey for a significant percentage of first-time buyers — and that engineering that ending would determine whether they became repeat customers. The strategy was counterintuitive by conventional efficiency metrics: spending more on shipping, accepting more returns, extending call times. But measured by the Peak-End Rule, every dollar spent on peaks and endings generated disproportionate returns in loyalty, word-of-mouth, and lifetime value.
Brian CheskyCo-founder & CEO, Airbnb, 2008–present
Chesky's "11-star experience" framework is the Peak-End Rule translated into product design methodology. In 2012, Chesky challenged his team to imagine what a five-star Airbnb stay would look like, then a six-star, a seven-star — all the way to eleven stars. The exercise was not about delivering an eleven-star experience (at eleven stars, Elon Musk would greet you in space). It was about identifying the peak moment that would make the experience unforgettable and then engineering a realistic version of it. The insight was that a stay with one extraordinary moment — a handwritten welcome note with local restaurant recommendations, a surprise bottle of wine matched to the guest's preferences, a host who arranged a private neighbourhood tour — would be remembered as superior to a stay with uniformly good but unremarkable amenities. Chesky extended this to the checkout experience — the end of the stay. Airbnb's review prompt, sent immediately after checkout, captures the guest at the moment when the Peak-End Rule is forming their summary judgment. The timing is deliberate: request the evaluation while the ending is fresh and the peak is accessible, before the memory fades into a generalised impression. Chesky's framework demonstrated that designing for peaks and endings is not a luxury reserved for premium products — it is a systematic methodology that any experience can adopt.
Howard SchultzCEO, Starbucks, 1986–2000, 2008–2017, 2022–2023
Schultz designed the Starbucks experience around the Peak-End Rule before the term existed in business vocabulary. The peak of the Starbucks visit is the moment the barista calls your name and hands you a personalised drink — a small ritual of individual recognition in an otherwise anonymous urban routine. The end is the first sip taken while settling into the "third place" environment Schultz designed: the lighting, the music, the seating arranged for both solitude and community. Schultz's genius was recognising that the coffee itself — the product — was the middle of the experience, not the peak or the end. The peak was the human moment of name recognition and personal connection. The end was the environmental feeling of arrival and comfort. When Starbucks's growth-era efficiency push in the mid-2000s eliminated some of these experiential elements — automated espresso machines that blocked eye contact, drive-through windows that removed the in-store ritual — customer satisfaction declined despite the coffee quality remaining constant. Schultz's return as CEO in 2008 was explicitly framed as a restoration of the experience, not the product. He understood that customers were not evaluating the coffee. They were evaluating the memory of the visit — and that memory was authored by the peak and the end, both of which the efficiency push had degraded.
Bob IgerCEO, The Walt Disney Company, 2005–2020, 2022–present
Disney's theme parks are the most sophisticated commercial application of the Peak-End Rule at scale. Iger inherited and extended Walt Disney's original insight: that guests would tolerate long waits, crowded walkways, and expensive food if the peak moments — the ride climax, the character encounter, the fireworks crescendo — were extraordinary and the ending was magical. The fireworks show is not scheduled at the end of the park's operating hours by coincidence. It is the designed ending of the visit — a spectacular peak-end combination that ensures the final memory is wonder, not the aching feet from eight hours of walking. Under Iger's leadership, Disney invested in MagicBand technology and the Genie+ system not primarily to reduce average wait times but to increase the predictability and quality of peak moments. The ability to schedule specific ride times and character meet-and-greets meant guests could engineer their own peaks, increasing the probability that the remembered highlights would be the highest-intensity experiences rather than the random encounters dictated by queue length. Iger understood that a guest who experienced two extraordinary peaks and a magical ending would rate the visit higher — and return sooner — than a guest who experienced uniformly moderate enjoyment across more attractions.
Section 6
Visual Explanation
Peak-End Rule — Memory does not average the experience. It encodes the peak (most intense moment) and the end (final moment), discarding duration and everything in between. The remembered evaluation — which drives all future decisions — is authored by these two moments alone.
Section 7
Connected Models
The Peak-End Rule does not operate in isolation — it interacts with a network of cognitive biases and design principles that either amplify its effects or provide correctives against its blind spots. The most consequential experience-design errors occur not from misunderstanding the Peak-End Rule alone but from failing to recognise how it compounds with related biases to produce memories that diverge dramatically from the objective quality of the experience. Understanding the connections below transforms the Peak-End Rule from a useful heuristic about memory into a comprehensive framework for designing experiences that are remembered, shared, and repeated.
The six connections map how the Peak-End Rule reinforces biases that amplify the dominance of recent and intense moments, creates productive tension with frameworks that demand holistic assessment, and leads to broader patterns of behaviour that determine customer loyalty, brand perception, and organisational culture at scale.
The reinforcing connections (Recency Bias and Loss Aversion) describe how related biases compound the Peak-End Rule's power — making endings even more dominant and negative peaks even more damaging. The tension connections (Systems Thinking and Goodhart's Law) provide essential correctives that prevent the rule from being applied reductively. The leads-to connections (Halo Effect and Framing Effect) describe the downstream consequences when peak-and-end memories aggregate into brand perceptions and narrative constructions that shape market behaviour at scale.
Reinforces
Recency Bias
The Peak-End Rule and recency bias share a common mechanism — the disproportionate weight that memory assigns to the most recently encountered information — but they operate at different scales. Recency bias describes the tendency to overweight the latest data point in sequential judgments: the most recent quarter, the last interview candidate, the final item in a presentation. The Peak-End Rule extends this mechanism to continuous experiences: the final moments of a journey dominate the remembered evaluation regardless of what preceded them. The reinforcement is direct — recency bias ensures the ending is the most cognitively available moment, and the Peak-End Rule ensures that cognitive availability translates into evaluative dominance. Together, they create a compounding effect: the end of an experience is not just remembered better (recency) but also weighted more heavily in the global judgment (Peak-End). For experience designers, this means the ending is doubly leveraged — it benefits from both the accessibility advantage of recency and the evaluative privilege of the Peak-End Rule. Investing disproportionately in the final moments of any experience is not a bet on one bias but on the convergence of two.
Reinforces
Loss Aversion
The Peak-End Rule interacts with loss aversion to make negative endings catastrophically more damaging than positive endings are beneficial. Loss aversion — the finding that losses are psychologically approximately twice as painful as equivalent gains are pleasurable — means that a negative peak or a negative ending is encoded with roughly double the emotional intensity of a positive one. A customer journey that ends with a billing dispute, a support failure, or an unresolved complaint does not merely end neutrally — it ends with a loss-coded memory that the Peak-End Rule then promotes to the dominant position in the remembered evaluation. The reinforcement amplifies the asymmetry: a positive ending improves the remembered experience moderately; a negative ending damages it disproportionately. This explains why companies that excel at damage recovery — resolving complaints quickly, generously, and personally — generate higher loyalty than companies that never make mistakes. The recovery moment becomes a positive peak, and the resolution becomes a positive ending — converting what loss aversion would have made a devastating memory into a loyalty-building one. The interaction between Peak-End and loss aversion means that how you handle failures matters more than how you deliver successes.
Section 8
One Key Quote
"Odd as it may seem, I am my remembering self, and the experiencing self, who does my living, is like a stranger to me."
— Daniel Kahneman, Thinking, Fast and Slow (2011)
Kahneman's observation identifies the central paradox that makes the Peak-End Rule so consequential for anyone designing experiences: the self that lives through the experience and the self that evaluates it afterward are not the same entity, and they operate by different rules. The experiencing self processes every moment with equal weight — a minute of pain is a minute of pain whether it occurs at the beginning, middle, or end. The remembering self discards most of what the experiencing self endured and constructs a summary from two data points. Every subsequent decision — whether to return, whether to recommend, whether to buy again — is made by the remembering self, using the edited summary, not by the experiencing self, using the full record.
The implication for founders and leaders is that optimising for the experiencing self and optimising for the remembering self require different strategies. Optimising for the experiencing self means making every moment as good as possible — reducing wait times, eliminating friction, maintaining consistent quality. Optimising for the remembering self means concentrating resources at the peak and the end — creating one extraordinary moment and ensuring the final impression is carefully designed. The companies that achieve the highest customer loyalty and the strongest word-of-mouth typically do both, but when resources are constrained — and they always are — the Peak-End Rule tells you where the marginal dollar generates the most remembered value. It is at the peak and the end, because those are the moments the remembering self will consult when the next decision arrives.
The deeper lesson is humility about our own evaluative processes. When you remember a vacation as wonderful, a meeting as productive, or a product as delightful — you are reporting a summary constructed from two moments, not a faithful record of the full experience. Your confidence in the evaluation is high, but the evidence base is thin. The remembering self does not announce that it has discarded ninety percent of the data. It presents its two-point summary as the complete truth. Understanding this is the first step toward designing experiences that serve both selves — and toward recognising that your own memories of experiences are Peak-End summaries, not objective histories.
The quote also illuminates a tension that every founder and product designer must navigate. The experiencing self and the remembering self want different things. The experiencing self wants comfort, consistency, and minimal friction throughout the journey. The remembering self wants intensity, contrast, and a satisfying conclusion. A perfectly smooth, frictionless experience may please the experiencing self in every moment but leave the remembering self with nothing to encode — producing the paradox of an experience that felt pleasant to live through but is remembered as forgettable. The most memorable experiences contain moments of heightened intensity — surprise, challenge, delight, even productive discomfort — that the experiencing self may not have preferred in real time but that the remembering self converts into the defining memory. The art of experience design is serving both selves: enough consistency and quality to retain the experiencing self through the journey, and enough peak intensity and ending craft to give the remembering self something worth keeping.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
The Peak-End Rule belongs in Tier 1 because it governs the mechanism by which every experience your company creates is converted into a memory — and memories, not experiences, drive all subsequent behaviour. Retention, referral, expansion, forgiveness after failure — every outcome that depends on a customer's (or employee's, or investor's) recollection of what happened is mediated by the Peak-End Rule. The company that understands this designs for memory. The company that does not designs for moments that will never be remembered.
What makes this model uniquely actionable is its specificity. Most cognitive biases tell you that human judgment is flawed. The Peak-End Rule tells you exactly which two moments in any experience determine the evaluation — and therefore exactly where to concentrate your design, investment, and attention. It transforms the vague mandate of "improve the customer experience" into the precise instruction of "improve the peak moment and the ending." That precision makes it one of the few cognitive models that translates directly into resource allocation decisions without requiring interpretation.
The most underappreciated application is in churn prevention. Most companies analyse churn by examining the moments before cancellation — the support tickets, the declining usage, the feature gaps. But the Peak-End Rule suggests that churn is often authored much earlier, at the moment when the customer's most recent experience ended poorly. A customer whose last support interaction ended without resolution, whose last product session ended with an error, or whose last invoice arrived with an unexplained increase has a negative end encoded as their most recent memory of the company. That memory is what they retrieve when the renewal decision arrives. The churn analysis should not begin at the cancellation event. It should begin at the last ending — the final moment of the most recent interaction — because that ending is the memory that will be weighed most heavily in the decision to stay or leave.
In fundraising, the Peak-End Rule explains a pattern every experienced founder recognises: the investor who seemed enthusiastic during the meeting but never follows up. The meeting itself may have contained strong content — compelling metrics, a clear market thesis, a differentiated product. But if the meeting ended weakly — a rambling close, no clear ask, an awkward goodbye in the lobby — the investor's memory of the meeting is authored by that ending. The enthusiasm from the middle of the meeting is not in the memory that determines whether the investor writes the follow-up email. The ending is. The founders who close rounds efficiently are the ones who design the last three minutes of every investor meeting with the same rigour they apply to the opening: a crisp summary, a specific ask, a confident close that leaves the investor's final memory as one of conviction rather than ambiguity.
Section 10
Test Yourself
The Peak-End Rule operates below conscious awareness because memory does not announce its editing process. When you remember an experience as wonderful or terrible, the evaluation feels like a comprehensive assessment of everything that happened. It is not. It is a two-point summary — the peak and the end — presented by the remembering self as the complete record. These scenarios test your ability to identify when the Peak-End Rule is shaping an evaluation that feels objective but is actually constructed from two moments rather than the full experience.
The critical distinction is between an evaluation that reflects the full experience — weighting duration, average quality, and variability — and an evaluation driven by the Peak-End Rule — weighting only the most intense moment and the final moment while neglecting everything else. Both produce evaluations that feel equally valid to the person reporting them. The difference is in what data the evaluation actually draws from. When the evaluation would change dramatically if you altered only the ending — keeping everything else identical — the Peak-End Rule is the dominant mechanism.
The diagnostic question in every case is: does the evaluation track the average quality of the full experience, or does it disproportionately reflect the most intense moment and the final moment? When the two diverge — when a mostly good experience is remembered negatively, or a mostly difficult experience is remembered positively — the Peak-End Rule is authoring the judgment.
Pay particular attention to duration neglect — the companion phenomenon. If a longer experience is rated equally or more favourably than a shorter one with the same average quality, duration neglect is operating. If the length of a meeting, a project, a customer interaction, or a wait has less influence on the retrospective evaluation than you would expect, the Peak-End Rule has overridden the duration signal. The most useful diagnostic is the mismatch test: compare the objective timeline (what happened, in what order, for how long) with the subjective evaluation (how the person remembers and rates it). Where the two diverge, the Peak-End Rule is the most likely explanation.
The counterfactual test is also revealing: ask yourself whether changing only the ending — while keeping the entire rest of the experience identical — would change the evaluation. If a terrible ending would convert a positive evaluation into a negative one, or if a brilliant ending would redeem an otherwise mediocre experience, the Peak-End Rule is the dominant evaluative mechanism. The same test applies to peaks: would removing the single most intense moment change the evaluation more than removing any five ordinary moments combined? If so, the peak is doing disproportionate work in the memory that generates the judgment.
The counterfactual test is also revealing: ask yourself whether changing only the ending — while keeping the entire rest of the experience identical — would change the evaluation. If a terrible ending would convert a positive evaluation into a negative one, or if a brilliant ending would redeem an otherwise mediocre experience, the Peak-End Rule is the dominant evaluative mechanism. The same test applies to peaks: would removing the single most intense moment change the evaluation more than removing any five ordinary moments combined? If so, the peak is doing disproportionate work in the memory that generates the judgment.
Is the Peak-End Rule shaping this judgment?
Scenario 1
A restaurant serves an outstanding appetiser, a good but unremarkable main course, and a dessert that arrives thirty minutes late, partially melted, with an unapologetic server. On the drive home, the diner tells their partner: 'That place was disappointing — I wouldn't go back.' The appetiser was the best dish they'd had in months.
Scenario 2
A SaaS company's annual user conference runs for three days. Day one has strong sessions. Day two features a product announcement that generates a standing ovation — the audience genuinely believes the new feature will transform their workflow. Day three's closing session runs overtime, the Wi-Fi fails during the final demo, and attendees leave the venue confused about next steps. Post-event surveys rate the conference a 6.2 out of 10.
Scenario 3
A consulting engagement runs for six months. The first five months involve competent but unremarkable work — the deliverables are adequate, the communication is professional, the insights are incremental. In month six, the lead consultant delivers a final presentation that reframes the client's entire strategic challenge, provides a genuinely novel insight the client had never considered, and closes with a clear, actionable roadmap. The client rates the engagement 9 out of 10 and immediately signs a renewal.
Section 11
Top Resources
The Peak-End Rule literature spans cognitive psychology, behavioural economics, experience design, and service management. The strongest foundation begins with Kahneman for the discovery and theoretical framework, extends to Redelmeier for the empirical evidence, and deepens with practical applications in customer experience and product design. The combination of theoretical understanding (why does memory privilege peaks and endings?) and operational application (how do I design experiences that produce favourable memories?) is what transforms the Peak-End Rule from a psychological curiosity into a competitive advantage in product, service, and leadership design.
For practitioners, the most immediately valuable resources are those that translate the cognitive science into design methodology — providing frameworks for identifying where peaks should be engineered, how endings should be crafted, and what baseline quality must be maintained across the full experience. Start with Kahneman for the mechanism, move to Redelmeier for the foundational evidence, advance to the Heaths for the practical design framework, then deepen with the original experimental papers that established the phenomenon's boundary conditions and robustness.
The progression from theory to evidence to application is deliberate: understanding why memory privileges peaks and endings (Kahneman) makes the empirical demonstrations more interpretable (Redelmeier, Kahneman et al.), and both make the design frameworks actionable rather than formulaic (Heath and Heath, Hsieh).
The definitive treatment of the Peak-End Rule within the broader framework of dual-process cognition and the experiencing-self versus remembering-self distinction. Chapters 35–38 provide the theoretical architecture: why memory edits experience, how duration neglect operates, and why the remembering self — not the experiencing self — makes decisions about the future. Kahneman's ice-water experiments and colonoscopy studies are presented with the precision of a discoverer explaining his own findings. Essential for understanding why the Peak-End Rule is not a communication trick but a structural feature of human memory.
The original colonoscopy study that provided the most rigorous empirical demonstration of the Peak-End Rule. Redelmeier and Kahneman collected real-time pain ratings from 682 patients and compared them with retrospective evaluations. The finding — that retrospective ratings correlated with peak intensity and end intensity but not with duration — established the Peak-End Rule as a robust empirical phenomenon with direct clinical and design implications. The paper's methodology, which combines moment-by-moment experience sampling with retrospective global evaluation, remains the gold standard for Peak-End research.
The most actionable translation of the Peak-End Rule into a practical design framework for leaders, founders, and experience architects. The Heath brothers identify four elements of defining moments — elevation, insight, pride, and connection — and provide a systematic methodology for engineering peaks in customer experiences, employee journeys, and educational settings. The book's case studies span hospitals, hotels, schools, and technology companies, demonstrating that peak design is a learnable discipline, not a matter of inspiration or luck.
The ice-water experiment that produced the most counterintuitive demonstration of the Peak-End Rule: participants preferred ninety seconds of pain to sixty seconds because the longer trial ended with marginally reduced intensity. The paper's finding that people choose more total pain when the ending is slightly better is the strongest evidence that the remembering self evaluates by peak and end rather than by integration, and that this evaluation — not the experiencing self's real-time preference — drives future choice. A concise, essential paper for understanding duration neglect.
Hsieh's account of building Zappos is the most detailed business case study of a company that operationalised the Peak-End Rule before the term entered business vocabulary. The book documents how Zappos designed its customer service to create peak moments (unexpected upgrades, personal connections) and positive endings (effortless returns, farewell interactions that felt genuine rather than scripted). The result was customer loyalty and word-of-mouth that defied the unit economics — because the Peak-End Rule converted a few extraordinary moments into memories that drove repeat behaviour at scale. Essential reading for any founder who wants to understand how customer experience translates into competitive advantage through the mechanism of memory rather than through the mechanism of average quality.
A comprehensive review of how the Peak-End Rule and related memory biases interact with digital customer journeys — where the "ending" is often the post-purchase email sequence, the app's session-close screen, or the final interaction in a support chat. The paper synthesises research on how digital experience design can engineer peaks (personalisation, surprise features, achievement notifications) and endings (session summaries, progress indicators, anticipation-building previews) to shape remembered evaluations in environments where the customer's attention span is measured in seconds. Particularly valuable for product managers and growth teams designing digital funnels where the Peak-End Rule's leverage is highest.
Tension
Systems Thinking
Systems thinking — the discipline of understanding how components interact within a whole — creates productive tension with the Peak-End Rule by insisting that every part of a system matters, not just the peak and the end. The Peak-End Rule says that memory privileges two moments and discards the rest. Systems thinking says that the "rest" — the middle of the experience, the operational infrastructure, the consistent baseline quality — is what enables the peak and the end to function. A restaurant that engineers a spectacular dessert (end) and a stunning presentation of the main course (peak) but serves cold appetisers and employs rude hosts will lose customers before they reach the moments the Peak-End Rule can leverage. The tension is essential: the Peak-End Rule identifies where to concentrate premium investment for memory formation, but systems thinking identifies the baseline quality that the entire experience must maintain to keep the customer engaged through to the end. The resolution is a two-tier design philosophy — uniform baseline quality across all touchpoints (systems thinking) with concentrated excellence at the peak and end (Peak-End Rule). Neither framework alone produces optimal outcomes.
Tension
Goodhart's Law
Goodhart's Law — when a measure becomes a target, it ceases to be a good measure — creates a critical tension with naive applications of the Peak-End Rule. If an organisation decides that only peaks and endings matter, it will begin optimising exclusively for those moments and neglecting the rest of the experience. The metric (post-experience satisfaction, driven by peak and end) becomes the target, and the organisation hollows out everything between the peaks to fund increasingly elaborate crescendos and finales. A hotel that invests everything in the lobby and the checkout experience but neglects room cleanliness is optimising for the Peak-End metric at the expense of the actual experience. Over time, customers detect the pattern — the experience feels manipulative rather than genuine, and the peak itself loses its emotional power because it sits atop a degraded foundation. Goodhart's Law warns that the Peak-End Rule is a description of how memory works, not a prescription for how to allocate resources. The prescription must balance peak-and-end investment against the baseline quality that makes peaks feel earned rather than manufactured.
Leads-to
[Halo Effect](/mental-models/halo-effect)
The Peak-End Rule is the experience-level mechanism that feeds the broader Halo Effect at the brand level. When a customer has a peak moment with a product — the first time a feature solves a problem they thought was unsolvable, the support interaction that exceeded every expectation — and the ending of their interaction is positive, the Peak-End Rule generates a favourable memory of that specific experience. Over repeated interactions, these favourable memories aggregate into a global positive impression — a halo — that radiates across every dimension the customer evaluates. The product is rated as more reliable, the company as more trustworthy, the brand as more premium — not because every interaction justified these ratings, but because the peaks and endings that dominated memory created a consistently positive evaluative signal. The leads-to relationship is causal: the Peak-End Rule determines which moments enter memory, and the accumulated memories form the halo. Companies that consistently engineer positive peaks and endings build halos faster and more durably than companies that deliver higher average quality but leave peaks and endings to chance.
Leads-to
Framing Effect
The Peak-End Rule shapes what is remembered; the framing effect shapes how what is remembered is interpreted. The leads-to relationship operates through narrative construction: after an experience, the remembering self does not merely recall the peak and end — it constructs a story that connects them. The peak becomes "the moment everything came together" or "the moment everything fell apart," and the ending becomes "and it all worked out" or "and it never recovered." This narrative construction is a framing operation — the same peak and end can be framed as a triumph or a disaster depending on the interpretive lens the remembering self applies. The practical implication is that organisations can influence post-experience framing by designing the ending to provide the narrative frame. A project retrospective that ends with "here is what we accomplished and here is what we learned" frames the peak moments — including the difficult ones — as progress. A retrospective that ends with "here is what went wrong" frames identical peaks as failures. The Peak-End Rule determines which moments are available for framing. The framing effect determines how those moments are interpreted. Together, they author the story the customer, employee, or stakeholder tells about what happened.
The rule has profound implications for how leaders manage organisational transitions. Layoffs, reorganisations, leadership changes, office moves — every organisational transition is an experience with a peak and an end. The peak is often unavoidable — the moment of announcement, the emotional intensity of uncertainty. But the ending is entirely within the leader's control. A layoff that ends with a generous severance conversation, a personal phone call from the CEO, and a genuine offer to help with placement is remembered differently from one that ends with a form email and a locked laptop. The people who leave will carry the ending into every Glassdoor review, every conversation with potential hires, and every interaction with the company's brand for years. The people who stay will carry the ending into their assessment of whether leadership can be trusted during the next transition. The Peak-End Rule means the final moments of a difficult transition have more influence on long-term organisational health than the transition itself.
The interaction between Peak-End and customer onboarding is where the rule generates the most measurable ROI. The onboarding experience is the customer's first complete journey with the product — and its peak and end will determine whether the customer becomes an active user or a churned trial. The peak should be the "aha moment" — the first time the customer achieves a meaningful result with the product. Slack's peak is the first message exchange with a teammate. Stripe's is the first successful payment processed. Canva's is the first design that looks professional. Every product has an aha moment, and the onboarding should be designed as a corridor to that moment with minimum friction. The end of onboarding — equally critical — should not be a generic "setup complete" screen. It should be a moment that affirms the customer's decision: a summary of what they accomplished, a preview of what comes next, a sense of momentum and progress. The default end of most onboarding flows is an empty dashboard. That is an ending designed by engineers, not by experience architects. The companies that convert the highest percentage of trials to paid users are the ones that design both the aha-moment peak and the onboarding ending with the same intentionality they apply to the product's core features.
The rule has a counterintuitive implication for how companies should handle service failures. A service failure that is resolved brilliantly can produce a stronger positive memory than a service interaction with no failure at all — because the resolution creates a peak (the moment of unexpected generosity or competence) and a positive ending (the problem solved) that a frictionless interaction never generates. This is the service recovery paradox, and the Peak-End Rule is its mechanism. The customer who experienced a problem and then experienced an extraordinary resolution has two vivid data points for the remembering self to encode. The customer who experienced a smooth, unremarkable interaction has none. The strategic implication is not to manufacture failures — but to invest disproportionately in recovery capabilities, because recovery moments are peak-and-end generators that frictionless interactions are not.
The Peak-End Rule also explains why some companies generate disproportionate word-of-mouth relative to their product quality. Word-of-mouth is driven by memory, not by experience. A customer who had a uniformly good but unremarkable experience has nothing memorable to share. A customer who had an uneven experience but one extraordinary peak moment — the support agent who went impossibly far beyond the expected response, the feature that solved a problem they had given up on, the personalised touch they did not expect — has a story. Stories are how word-of-mouth propagates, and stories are constructed from peaks and endings. The companies that generate the most organic referrals are not necessarily the ones with the highest average product quality. They are the ones that produce the most memorable peaks — because peaks become stories, and stories become referrals.
The Peak-End Rule reframes the entire discipline of brand building. A brand is not a logo, a colour palette, or a tagline. A brand is the aggregate memory that customers, employees, and partners carry about their interactions with the company. And that aggregate memory is constructed from peaks and endings — not from averages. The company that delivers one transcendent moment per customer interaction builds a stronger brand than the company that delivers consistently adequate interactions with no memorable peaks. This is why brands like Apple, Disney, and Ritz-Carlton command premiums that their average product quality alone cannot justify. The premium is not for the average. It is for the peaks and endings that the remembering self has encoded and that the brand now evokes. Understanding this reframes brand investment: every dollar spent on engineering a peak moment or crafting a positive ending is a dollar invested in the memory that constitutes the brand. Every dollar spent on raising average quality from adequate to slightly-above-adequate is a dollar invested in a dimension that memory will largely discard.
One critical warning: the Peak-End Rule can be weaponised to disguise a deteriorating experience. A company that maintains spectacular peaks and polished endings while hollowing out the middle of the experience — reducing support quality, cutting feature investment, allowing technical debt to accumulate — can sustain strong satisfaction scores for a period, because the metrics are measuring memory, not experience. The experiencing self notices the decline. The remembering self does not — yet. But eventually, the degraded middle becomes so poor that customers abandon the experience before reaching the engineered ending. When that happens, the carefully designed end never occurs, and the customer's final memory is the moment of frustration that caused them to leave mid-journey. The Peak-End Rule protects experience quality when applied honestly — investing in peaks and endings while maintaining baseline quality. It enables decline when applied cynically — using peaks and endings to mask deterioration that the experiencing self endures but the remembering self never encodes.
Scenario 4
An employee leaves a company after four years. Years one and two were excellent — strong mentorship, rapid promotion, meaningful projects. Year three involved a difficult manager transition that created six months of frustration. Year four was stable but unremarkable, and the employee's last week included a perfunctory exit interview, no farewell from the team, and a final day that ended with returning a laptop to IT. The employee's Glassdoor review describes the company as 'a place that doesn't value its people.'
