Peak-End Rule, Zeigarnik Effect & Experience Memory

The full lesson

Users don’t remember experiences the way they actually happened. Afterward, the mind reconstructs them from a few vivid fragments. Those reconstructions — not the real-time flow — drive satisfaction scores, word-of-mouth, and whether someone comes back.

Two psychological principles govern this process. The Peak-End Rule explains how people summarize completed experiences. The Zeigarnik Effect explains why incomplete ones linger in the mind. Together they form one of the most practically useful frameworks in UX design.

The Peak-End Rule: Memory Is Not an Average

Psychologist Daniel Kahneman and colleagues showed in a landmark 1993 study that people judge past experiences almost entirely on two moments: the most emotionally intense point (the “peak”) and how the experience ended. The total duration barely matters — Kahneman called this duration neglect.

In the original study, patients undergoing colonoscopies rated longer procedures as less unpleasant than shorter ones — if the longer procedure ended more gently. The extra minutes of discomfort paradoxically improved the remembered experience.

What This Means for UX

Every product interaction has a peak and an end. The peak can be positive (a moment of delight) or negative (a moment of frustration). The end is whatever the user experiences last before they close the tab or put down their phone.

Here’s what that means in practice:

• A painful peak poisons the memory of an otherwise smooth flow. A checkout form that fails validation three times will be remembered as “a nightmare” — even if onboarding was effortless.
• A strong ending rewrites the story. A well-crafted confirmation screen, a warm empty-state after deletion, or a helpful summary email after a complex task can lift overall satisfaction.
• Duration neglect means you can add helpful friction. A short clarification step or a brief “you’re all set” animation adds time to the experience. But it can still improve how users feel about it — if it lands at the right moment.

The Zeigarnik Effect: Unfinished Tasks Occupy the Mind

In 1927, researcher Bluma Zeigarnik noticed that waiters could recall every detail of open orders — but forgot them almost immediately after delivery. She called this the Zeigarnik Effect: the mind keeps incomplete tasks active, and that activation fades once the task is done.

This isn’t just a memory quirk — it’s an attentional mechanism. Incomplete loops create a low-level mental tension that keeps the task “alive” in working memory until it’s resolved. That tension can motivate someone to return and finish. But when too many unresolved loops pile up, the result is anxiety and cognitive overload.

Productive Tension vs. Cognitive Burden

The effect cuts both ways in product design:

The design question isn’t whether to use incomplete loops — it’s whether the tension those loops create serves the user or wears them down.

Designing the Peak: Creating Positive High Points

If the peak moment dominates memory, you need to find it and shape it deliberately. The peak isn’t always where designers expect it to be.

How to Locate the Peak

• Journey mapping with emotional valence. Annotate each touchpoint with its expected emotional intensity — positive or negative. Focus on friction, relief, surprise, and accomplishment, not just visual polish.
• Behavioral data over self-report. Session recordings and event funnels show where users spend extra time or abandon. These are candidate negative peaks. Post-task surveys alone can’t reliably find them — users often can’t accurately describe their own frustration in hindsight.
• Prototype testing with valence probes. After a task, ask “What moment stood out most?” rather than “How satisfied were you?” The first question surfaces specific scenes. The second collapses them into a number.

Designing Positive Peaks

A peak doesn’t need to be a fireworks animation. Positive peaks are usually moments of relief, competence, or surprise. Here are practical ways to create them:

• Milestone celebrations. Completing onboarding, a first successful export, a first team member invited — these moments should feel earned, not random.
• Error recovery moments. When a user recovers from a failure with your help, the relief can become a positive peak. A clear, specific error message with a one-click fix is more memorable than a flawless flow the user barely noticed.
• Unexpected delight. An empty state that acknowledges context (“No messages yet — start a conversation by clicking New Thread”) creates a small positive surprise instead of a generic placeholder.

Designing the End: The Recency Effect in Practice

The end of an experience carries outsized weight in how people remember it. “End” is context-dependent — it might be the last screen before checkout completes, the last email in an onboarding sequence, or the last thing a user does before canceling.

Principles for Strong Endings

1. Provide clear resolution. Ambiguous endings create open loops (the Zeigarnik Effect) that feel negative rather than motivating. A good confirmation state answers three questions: what happened, what comes next, and where to go if something went wrong.

2. Summarize accomplishment. A post-task summary like “You exported 1,240 rows in 3 formats” tells users they succeeded and reinforces the value of the interaction.

3. Don’t end on friction. Upsell screens, survey modals, and permission requests placed right after task completion contaminate the ending. Move them to a subsequent session or a neutral interstitial moment.

4. Match tone to context. A delightful, celebratory ending in a legal document signing flow can feel jarring and undermine trust. Calm, confident closure is often better than animation.

Zeigarnik in Retention and Re-engagement

The Zeigarnik Effect is the psychological engine behind many widely-used retention patterns. Understanding the mechanism lets you deploy them ethically — and recognize when they cross a line.

Legitimate Applications

Progress bars and completion meters. LinkedIn’s profile strength indicator and GitHub’s contribution graph both create incomplete-loop tension that motivates return visits. The loop resolves when the user takes action. The product benefits, and so does the user.

Saved state and draft persistence. Auto-saving a partially completed form and surfacing it on the next visit (“Continue where you left off”) reactivates the Zeigarnik loop productively. The user has a genuine reason to return.

Streaks and learning cadences. Duolingo’s streak mechanic and GitHub’s contribution streaks leverage Zeigarnik tension around daily incomplete loops. They work because the behavior they encourage — language practice, code commits — genuinely benefits the user.

Exploitative Patterns to Avoid

The same mechanism powers dark patterns that create anxiety without any user benefit:

• Artificial incompleteness. Showing a profile as “40% complete” when the missing fields are optional or irrelevant creates tension without value.
• Uncontrollable notification badges. Badge counts that climb even when the user has reviewed all content (algorithmic feeds that keep generating “new” items) create open loops that can’t be closed.
• Cliffhanger UX. Forcing users to start a process before revealing its cost or time commitment — and completing it after they’ve invested effort — exploits Zeigarnik tension to push through decisions users might later regret. This pattern is increasingly actionable under deceptive-patterns regulations.

The ethical test is simple: does closing the loop benefit the user, or only the product?

Experience Memory vs. Real-Time Experience

Kahneman’s broader framework distinguishes two selves:

• The experiencing self lives in the present moment, tracking how things feel second by second.
• The remembering self evaluates experiences after the fact and decides what to do next.

Most UX metrics measure the remembering self. Net Promoter Score, satisfaction ratings, App Store reviews — all are post-hoc reconstructions governed by the Peak-End Rule. Real-time behavioral data (time on task, error rates, scroll depth, session length) is a closer proxy for the experiencing self.

The Practical Gap

A product can have high real-time friction — long but predictable flows — and still earn excellent memory ratings if the peak is positive and the ending is strong. Conversely, a smooth flow that ends abruptly or ambiguously can generate low satisfaction scores despite low measurable friction.

This is why validated scales like SUS, UMUX-Lite, and SEQ often diverge from task-completion metrics. They measure different selves. Using both gives a fuller picture than relying on either alone. The modern best practice is mixed-method triangulation: behavioral event data for the experiencing self, validated post-task scales for the remembering self, and targeted qualitative probes to locate peaks.

Applying Both Principles Together

The Peak-End Rule and the Zeigarnik Effect work together naturally. Here’s a repeatable design framework for applying both:

1. Map the emotional arc. Identify every touchpoint. Annotate expected valence and intensity. Find the likely peaks — positive and negative — and the endings.
2. Redesign negative peaks first. These have the highest leverage on remembered experience. A painful error state, a confusing modal, a failed payment — fix these before adding delight elsewhere.
3. Engineer the ending. The last screen, email, or moment in every key flow deserves the same deliberate attention as the hero flow’s first screen.
4. Use Zeigarnik tension intentionally. Decide which incomplete loops you want to create — onboarding steps, streaks, drafts — and make sure each one resolves in a way that benefits the user.
5. Close loops that should be closed. Notifications, action items, and incomplete tasks that users can’t meaningfully resolve should be resolved for them or suppressed — not left open as persistent cognitive debt.

Common Misconceptions

“Longer experiences are worse experiences.” Duration neglect means the length of an experience is nearly irrelevant to its remembered quality. A lengthy but well-resolved onboarding with a strong ending can be remembered more positively than a short onboarding that ends in confusion.

“Smooth flows equal positive memories.” Frictionless flows with no emotional peak are often simply forgotten. Memorable experiences — positive or negative — require emotional intensity somewhere. The goal isn’t to eliminate all friction; it’s to make sure the peaks are positive.

“Zeigarnik loops always drive engagement.” Only when the user can meaningfully close the loop and gains something from doing so. Open loops that can’t be closed — or that close to no user benefit — generate anxiety and churn, not engagement.

“Post-task surveys measure the experience.” They measure the remembering self’s reconstruction. They’re valuable, but they should always be triangulated with behavioral data from the experiencing self.