Zeigarnik Effect

Law 13 / 20 Design Better Decisions

Unfinished tasks are often said to stick in memory; the evidence for that is weak, but people do tend to return to what they started.

  • Documented behavioral effect
  • 5 min read
  • Documented study
  • Reviewed by Ali Dezhami, November 8, 2026

The dilemma

Illustrative scenario

Daniel leads growth at a residential solar installer. Its online quote form has six steps, and most people who abandon it do so at step four, which asks them to upload a recent electricity bill.

The agency pitches a "Zeigarnik campaign." A progress bar will start at 20% before the visitor has done anything, and abandoners will get an email and a text within the hour: "You're 80% done! Your savings estimate is waiting." The pitch claims the brain "can't let go" of unfinished tasks.

Daniel's operations lead is wary. A past round of reminder texts brought back some quotes but also a spike in complaints and opt-outs. And call notes suggest many people stopped at step four simply because they did not have a bill to hand, not because they forgot.

The agency wants a yes by Friday so the campaign can run before the spring sales push. Should Daniel approve it?

The short answer

The Zeigarnik effect is the claim that people remember interrupted or unfinished tasks better than completed ones. It is named after the psychologist Bluma Zeigarnik, who reported it in 1927 from experiments in Kurt Lewin's laboratory in Berlin, in which participants worked through a series of short tasks and some were interrupted before they could finish.

How strong is the evidence?

Documented behavioral effect

An effect found in psychology experiments. Its size depends on conditions, and some findings replicate better than others.

This is a documented behavioral effect with weak general support. Zeigarnik's original result is real history, but later attempts to reproduce it gave mixed results, and the effect seemed to depend on conditions such as how motivated people were and whether an interruption felt like a personal failure.

A closely related idea is easier to defend. Maria Ovsiankina, also working in Lewin's circle, studied the tendency to resume interrupted tasks when given the chance. Memory and resumption are different claims: one says you will remember the unfinished task, the other says you will want to go back to it.

The most recent synthesis, a 2025 meta-analysis by Ghibellini and Meier (see the example), separates the two. For marketing, the practical lever is resumption, not a supposed memory itch: make it easy to continue where someone stopped.

This matters for abandoned-cart and abandoned-form reminders. They may work, but a working reminder does not show that the Zeigarnik effect is the cause; a reminder also helps through simple prompting, convenience, a link straight back to the task, or an added discount.

Investigating an example

Documented study

What a 2025 meta-analysis found

Ghibellini and Meier pooled experiments on both effects: better recall of unfinished tasks (Zeigarnik) and the tendency to resume them (Ovsiankina). They found no general memory advantage for unfinished tasks, concluding that the Zeigarnik effect lacks universal validity. They did find evidence for a tendency to resume interrupted tasks.

For a team designing forms, checkouts or onboarding, that is a useful shift. The defensible bet is not that customers will be haunted by an unfinished quote. It is that some of them will come back if returning is easy.

Limits: a meta-analysis can only combine the studies that exist, and most of them are laboratory tasks rather than purchases or sign-ups. It does not tell you how much any specific reminder will lift completion; only your own controlled test can do that.

A published study or record, linked below. Its limits are stated.

Choose your move

You are in the dilemma above. Which move would you make? Open each one to see its trade-offs.

A Approve the agency campaign: a progress bar that starts at 20% and "80% done" messages within the hour.

May recover some quotes quickly. But the progress shown is not real, the messages overstate how close people are, rapid texts risk complaints and opt-outs, and the theory behind it is weak.

B Fix step four first (allow an estimated bill or a later upload), save progress automatically, and send one opt-in reminder linking to the exact step, tested against a holdout.

Slower to build and less dramatic. It tackles the real reason people stop and measures true lift. Consent rules limit who can be reminded, so the reach is smaller.

C Send no reminders at all and simply save progress.

Respectful and simple, with no complaint risk. But it gives up a resumption nudge that may help people who meant to return, and without a test the team learns nothing.

No move wins in every business. The law tells you what to check, not what to choose.

How to apply it

  1. Find out why people stop. Use step-by-step drop-off, support questions and call notes. Missing information, price shock and simple distraction need different fixes.
  2. Make resuming effortless. Save entries automatically and link people back to the exact step they left, with their answers intact.
  3. Show honest progress. Progress indicators should reflect the real steps remaining. A fake head start misleads and erodes trust when people notice.
  4. Remind with consent and restraint. Use a clear opt-in, cap frequency, choose a sensible delay and make unsubscribing easy.
  5. Test against a holdout. Withhold reminders from a random slice of abandoners and compare completion, complaints and downstream quality.

What to measure

Completion rate for reminded abandoners against a holdout, time to resume, unsubscribe and complaint rates, and downstream quality (for the installer: quotes that become site visits). Guardrail: if the lift over the holdout is small while complaints or opt-outs rise, scale back frequency or stop.

Worked example

Worked example: reading a reminder test against a holdout (illustrative numbers)
GroupAbandoned quotesCompleted within 7 daysCompletion rate
One reminder sent 2,000 260 13.0%
Holdout: no reminder 2,000 190 9.5%
Difference (reminder effect) — 70 +3.5 points

Illustrative scenario, not data from a real company. Without the holdout, the team might credit all 260 completions to the reminder, although 190 of them would likely have happened anyway. Even the 70 extra completions do not reveal why the reminder worked.

Where it breaks

  • The memory advantage does not hold in general. "The brain can't let go" is not a sound basis for a campaign.
  • Resumption is a tendency, not a guarantee. People also abandon because of price, missing information or a changed mind, and reminders do not fix those.
  • Interruption can feel like failure, and in some studies that changed how people remembered or approached the task.
  • Manufactured incompleteness backfires. Fake progress, split-up content and guilt-based streak alerts can feel manipulative and drive opt-outs.
  • Ethics and law: reminders need consent and sensible frequency; misleading progress claims can cross into deceptive practice.

What if you apply it incorrectly?

A language-learning app splits each lesson into deliberately unfinished halves and sends "You left something unfinished!" notifications, citing the Zeigarnik effect. Sessions per user rise for a few weeks. Then notification opt-outs climb, reviews mention nagging, and the team cannot tell whether learners came back to learn or just to clear the alert.

Myth versus evidence

The social-video version

Leave something unfinished and customers' brains will force them to come back.

What the evidence supports

A 2025 meta-analysis found no general memory advantage for unfinished tasks; people do tend to resume interrupted tasks, so make resuming easy instead of manufacturing loose ends.

Five-minute field assignment

Open your form or checkout analytics and find the single step with the largest drop-off. Write down the most likely reason: missing information, price shock or distraction. Then test it yourself: start the flow, leave at that step, and come back tomorrow from another device. Do you land on that step with your entries saved, or start over?

Checkpoint

Two questions: one about the idea, one about a decision. Answer both correctly to master this law. Questions are drawn from a larger bank and graded on our server; a retry brings different questions.

Evidence ledger

Every factual claim in this lesson, where it comes from, and how far it has been checked.

Further reading

Which other law could explain this decision?

Editorial review: Ali Dezhami, November 8, 2026. Last substantive update: November 8, 2026. Our methodology