Skip to content
InsideUX Design
Psychology 16 min read

The Zeigarnik Effect: What’s Really Behind the Popular UX Rule

Unfinished tasks stick in your mind, says a popular UX rule. What research says today, what it means for progress bars, checklists and streaks, and where manipulation begins.

Picture a budgeting app. Let’s call it “Monthly.” Right after you sign up, a bar appears: “Your profile is 20% complete.” Below it sits a list of nine open items, from connecting your bank account and inviting your flatmates to turning on notifications. On the home screen, a little flame counts the days in a row on which you have logged your expenses. And if you haven’t logged anything by 10:30 pm, a push notification arrives: “Your streak is in danger!”

Ask the team why it is built this way and one name comes up quickly: the Zeigarnik effect. Unfinished tasks stick in your mind, so open items make people come back. That is what many UX collections, blog posts and workshop slides say.

The trouble is that it isn’t that simple. The effect Bluma Zeigarnik described almost 100 years ago was hard to replicate early on, and a 2025 meta-analysis finds no memory advantage for unfinished tasks on average. What does hold up fairly well is something else: the tendency to pick interrupted tasks back up. This article sorts out what is actually behind the popular UX rule, what remains useful for design and where “motivation” turns into manipulation. The “Monthly” app serves as a running example throughout. It is made up, but typical.

What Bluma Zeigarnik studied in 1927

Bluma Zeigarnik (1900 to 1988) was a Soviet psychologist who worked in Kurt Lewin’s research group in Berlin in the 1920s. Her study appeared in 1927 in the journal Psychologische Forschung as part of Lewin’s series “Untersuchungen zur Handlungs- und Affektpsychologie” (investigations into the psychology of action and emotion). Its title translates as “On the retention of completed and uncompleted actions.”

There is a popular anecdote about how it all started. Lewin and some colleagues are said to have noticed in a Berlin restaurant that a waiter remembered orders perfectly as long as they were unpaid, then quickly forgot them. In a 2020 historical review, psychologist Colin MacLeod points out that the story comes in different versions: in one, the waiter forgot the total of the bill, in another, who had ordered what. It makes a nice illustration, but it is not evidence.

The actual experiment worked like this. In the first study, 32 adults individually worked through 22 small tasks each, such as folding paper, doing multiplication, drawing a vase or counting backwards. Each task took a few minutes. The experimenter interrupted half of them, right when the person was most absorbed. Later, participants were asked to list the tasks they remembered.

The result: interrupted tasks were named much more often than completed ones. According to MacLeod, the ratio across Zeigarnik’s four main experiments hovered close to 2, meaning roughly twice as many interrupted tasks as completed ones were recalled.

The explanation came from Lewin’s field theory. When you set out to do a task, you build up a kind of inner tension, a “quasi-need.” That tension is only released once the intention is fulfilled. If the task stays unfinished, the tension stays, and so does the task in your memory.

How the effect made its way into UX

The idea is catchy, and it fits many patterns that are common in digital products anyway. Jon Yablonski’s collection Laws of UX sums the effect up like this: people remember uncompleted or interrupted tasks better than completed ones. Among other things, it recommends providing a clear indication of progress to motivate people to finish.

In practice, the “Zeigarnik” label shows up in four places in particular:

  • Progress bars in multi-step processes, such as “Step 2 of 4” in a checkout or application form.
  • Profile completeness meters like “Your profile is 64% complete.”
  • Onboarding checklists meant to guide people through their first steps in a product.
  • Streaks, meaning runs of consecutive days on which someone did something, familiar from language, fitness and habit apps.

The reasoning is usually the same: an open item creates tension, tension stays on your mind, so people come back to resolve it. That sounds plausible. But it blends two things that research now separates clearly: remembering unfinished tasks and the urge to finish them.

What the research says today

Early doubts

The fact that the effect is not as robust as UX slides suggest is not news. In his review, MacLeod traces how mixed the replication attempts were back in the 20th century. As early as the 1950s and 1960s, researchers noted that the effect was far from reliable. In her 1968 doctoral thesis, after running careful replications of her own, the researcher van Bergen summed it up: of the studies intended to show a Zeigarnik effect, fewer than half actually did.

According to MacLeod, the effect appears to depend heavily on individual differences, such as achievement motivation, and is hard to demonstrate with unselected participants.

The 2025 meta-analysis

A meta-analysis combines the results of many individual studies statistically. In 2025, Romain Ghibellini and Beat Meier of the University of Bern did exactly that for the Zeigarnik effect and its sibling, the Ovsiankina effect. Their paper appeared in Humanities and Social Sciences Communications and is open access.

They searched two psychology literature databases and ended up including 59 publications: 38 on the Zeigarnik effect, 20 on the Ovsiankina effect and one covering both. Because the studies report their results in very different ways, the authors analyzed several measures separately. The key findings on recall:

  • Dividing the average number of recalled interrupted tasks by the average number of recalled completed tasks gives a value of 0.99 across 38 studies. Both are remembered practically equally well.
  • The share of interrupted tasks among all recalled tasks is around 49%, if anything slightly below what you would expect with no effect at all.
  • Only one measure, which averages ratios per person, shows a slight advantage for interrupted tasks. According to the authors, that measure is heavily distorted by single extreme values. For the same reason, they consider Zeigarnik’s own figures inflated.

Their conclusion is clear: the supposed memory advantage for interrupted tasks is “certainly not universal.” The Zeigarnik effect lacks universal validity.

So why did it work for Zeigarnik? The authors suspect that the effect depends on the situation and the person. Zeigarnik herself described how conscientiously her participants worked, out of a sense of duty, ambition or interest. Ghibellini and Meier also assume that university experimenters carried a lot of authority at the time. Both strong task involvement and a weighty situation, they argue, are less common today. They present this explicitly as an assumption, not an established finding.

The authors are open about the limits of their analysis, too: some measures rest on only a few studies, methods vary widely, and the search was deliberately narrow. So the meta-analysis does not end the debate. It does shift the burden of proof, though. Anyone claiming today that open tasks are generally remembered better is on thin ice.

The Ovsiankina effect: resuming rather than remembering

The second part of the analysis is more interesting for design. Maria Ovsiankina, also from Lewin’s Berlin group, studied behavior rather than memory in 1928. Her question: what do people do when they get the chance to return to a task after an interruption? Her participants often resumed the task on their own, without being asked.

Across 21 studies, Ghibellini and Meier found a weighted resumption rate of 67%, compared with a chance rate of 50%. Their conclusion: the Ovsiankina effect represents a general tendency. Here, too, they urge caution. Because the studies were designed so differently, the figure is indicative rather than absolute.

For UX, this distinction matters. Whether someone remembers a half-finished application form is uncertain. That they are more likely to continue it when it is visible and easy to get back to is well supported. Incidentally, the authors also mention what gets in the way of focused work on a task today: phone notifications, work emails and a tendency to multitask. Digital products are part of the very problem they try to solve with progress bars.

What really drives progress bars and streaks

If the Zeigarnik effect is shaky as an explanation, that doesn’t make progress indicators useless. There are simply better-fitting explanations, and they have stronger evidence behind them.

The goal gradient: In 2006, Ran Kivetz, Oleg Urminsky and Yuhuang Zheng studied a café reward program along the lines of “buy ten coffees, get one free.” Customers bought coffee more often the closer they got to the free reward. People put in more effort the smaller the remaining distance to a goal seems. That explains why “one more step” motivates, without needing a special memory phenomenon.

Endowed progress: The same year, Joseph Nunes and Xavier Drèze showed that artificial advancement increases persistence. At a car wash, some customers received a loyalty card with eight spaces, others a card with ten spaces and two stamps already filled in. Either way, eight washes were needed. According to the authors’ results, those with the apparent head start were more likely to keep going. Kivetz and his team found a similar effect with a twelve-stamp card that came with two bonus stamps.

Streaks as a goal in themselves: In seven studies published in 2023, Jackie Silverman and Alixandra Barasch showed that highlighting an intact streak makes people repeat a behavior more than highlighting a broken one. What mattered was not people’s actual behavior but how the log presented it. The streak becomes a goal of its own. The drop after a broken streak was stronger when people blamed themselves and weaker when they could repair it.

That makes for a more honest picture. Progress indicators probably don’t work because unfinished things stick in memory. They work because people like reaching goals that feel close, because a head start motivates and because a displayed streak turns into a goal in itself. These are powerful levers. And that is exactly why they can also be used against people’s interests.

What this means for design

1. Plan for interruptions and make it easy to pick up again

The most robust finding in this research is the Ovsiankina effect: people like to continue interrupted tasks when they get the chance. For design, that mainly means creating that chance. Save drafts automatically, offer to continue where someone left off the next time they open the app, and show the current state so nobody has to guess what is already done.

In “Monthly,” it looks like this:

  • Before: If you abandon the bank account setup, you land on the home screen the next time you open the app. Your half-finished progress is gone.
  • After: A calm card sits at the top of the home screen: “Connect your bank account: 2 of 3 steps done. Continue.” One tap leads straight to the open step, with everything entered so far still in place.

This is not a psychological trick. It is simply good interaction design. It takes work off people’s hands instead of making them feel guilty.

2. Show progress honestly

Progress indicators help people orient themselves: how long is this process, and where am I right now? That is their real purpose. To serve it, they have to be accurate. A bar that says “Step 2 of 4” and then unfolds three more sub-steps destroys trust.

Endowed progress is a borderline case. If people see 20% after signing up because their name and email address really have been captured, they are getting true information. If they see 20% when nothing has happened, they are being misled. The rule is simple: only count what is actually done.

3. Checklists that deliver real value

An onboarding checklist works well when every item benefits the person using it. It shows what the product can do and leads to the moment when it helps for the first time. It becomes a problem when it mainly works through the company’s goals: upload your contacts, allow notifications, rate the app.

  • Before: Nine items, including “Invite contacts,” “Turn on notifications” and “Try Premium.” The bar stays stuck at 78% until you try Premium.
  • After: Three items: “Log your first expense,” “Set a monthly budget,” “Adjust categories.” Each one can be skipped. After the third, the list disappears and makes room for the monthly overview.

The ending matters, too. A checklist that can never be fully completed is not a guide, it is a standing order. Let it disappear once it has done its job, or let people hide it.

4. Streaks with a safety net

Streaks can help build habits, as Silverman and Barasch also show. But their results make clear where the risk lies: a broken streak is demotivating, especially when people blame themselves. If you display a streak, plan for what happens when it breaks.

Useful options include taking a break or repairing a streak, and a display that doesn’t just show a zero after a gap but also the overall achievement: “You logged expenses on 24 days this month.” In “Monthly,” the streak now counts weeks instead of days, and one day off per week doesn’t break it. For a budgeting app, that is plenty.

5. Make completion visible

If open items tie up attention, the reverse also holds: clear completion is a relief. A confirmation after closing the month, a checkmark next to finished steps, a summary of what has been saved. This is not derived from Zeigarnik’s study. It is a matter of feedback, which Jakob Nielsen made the first of his ten usability heuristics: the design should always keep people informed about what is going on. People should know they are done, and then be left in peace.

Where it tips into manipulation

Every mechanism described here has a dark side. They all target motivation, and motivation can be used in both directions. At these points at the latest, a motivational pattern becomes a dark pattern:

  • Fake progress: A bar that starts at 80% so people don’t abandon the rest. Or a process that keeps going after the supposedly final step.
  • Guilt as a notification: “Your streak is in danger!” at 10:30 pm, or copy that makes people feel bad for declining an option.
  • Open items as leverage for business goals: A profile that never becomes “complete” without a phone number, a contact list or a rating.
  • Artificial interruptions: Content that deliberately cuts off mid-sentence or mid-episode to pull people into a subscription or the next session.
  • Fear of loss with no way out: Streaks that can only be saved by paying.

A useful test: would the person agree to the mechanism if they knew exactly how it works? A progress bar that honestly shows how far along an application is passes this test. A bar that artificially starts at 80% so nobody drops out does not. You’ll find more patterns and remedies in the article Spotting and Avoiding Dark Patterns. Where the line runs between a helpful push and paternalism is covered in the article on nudging in UX design.

There is also a very practical argument. Mechanics that put people under pressure can lift metrics like daily activity in the short term. But they don’t build trust, and people who feel nagged turn off notifications or delete the app. So don’t judge whether progress bars and streaks really help by clicks alone. Also look at opt-outs, uninstalls and what people say about them in usability tests.

Limits and pitfalls

A name is not a justification. “That’s the Zeigarnik effect” sounds scientific, but it doesn’t replace testing. If a checklist is supposed to improve activation, check it with an A/B test or at least with real usage data rather than citing a lab study that is almost a hundred years old.

The lab is not everyday life. Even the better-supported effects come from specific situations: loyalty cards, lab experiments, self-reports. How strongly they work in your product depends on your audience, the context and the task itself.

More open items do not mean more motivation. A list of twelve open tasks is more likely to overwhelm than to spark curiosity. The article Cognitive Load: Why Less Is Often More in Interface Design explains how much mental effort an interface demands.

Not everyone reacts the same way. Even the early Zeigarnik research showed how much personality and motivation come into play. For some people, a streak is an incentive. For others, it is a source of stress. Give people the option to hide such elements.

Conclusion

The Zeigarnik effect is one of the best-known “laws” in UX design, but one of the weakest. Bluma Zeigarnik’s 1927 study was original and influential. Its central result, that unfinished tasks stay in memory better, proved hard to replicate early on, though, and the meta-analysis by Ghibellini and Meier finds no such advantage on average. The Ovsiankina effect, on the other hand, is well supported: people like to continue interrupted tasks when they get the chance.

For design, that is a useful correction. Progress bars, checklists and streaks can help, but for different reasons than often claimed: because close goals motivate, because orientation reduces effort and because an easy way back in brings people back. After the redesign, the “Monthly” app has fewer open items, an honest progress bar, a streak with a safety net and no more guilt-inducing push notifications. It lets people pick up where they left off, and it leaves them alone once they are done.

Checklist

  • Arguments based on the Zeigarnik effect have been questioned and replaced by tests or data.
  • Interrupted processes are saved and can be resumed with one tap.
  • Progress indicators only count what is actually done, and the number of steps is accurate.
  • Every item on an onboarding checklist benefits the person using the product.
  • Checklists can be skipped, hidden and fully completed.
  • Streaks have a safety net: a pause, a repair option or a display of overall achievement.
  • Notifications inform rather than create guilt or fear of loss.
  • No progress is tied to business goals like contacts, ratings or subscriptions.
  • Completed processes get a clear confirmation.
  • Success is measured not just by activity but also by opt-outs and feedback.

Sources

Back to “All articles”

You might also like

Cognitive Load: Why Less Is Often More in Interface Design

Psychology

Every interface asks for mental effort, but much of it is unnecessary. What cognitive load is, how to spot it and seven practical ways to reduce it.

Spotting and Avoiding Dark Patterns: Why Interface Tricks Cost You

Psychology

Fake countdowns, hidden cancellations, cookie banners without a "Reject" button: why dark patterns cost trust, what EU and German law say, and what fair alternatives look like.

The Von Restorff Effect: Why What’s Different Stands Out

Psychology

What stands out from its surroundings sticks in memory. What Hedwig von Restorff found in 1933 and how to use emphasis in interfaces deliberately and accessibly.