Zeigarnik effect
Unfinished tasks are said to stay in mind more readily than finished ones.
Zeigarnik reported that waiters recalled unsettled orders better than settled ones, and that interrupted laboratory tasks were recalled better than completed ones. The proposed mechanism is a tension state that persists until the task closes. The replication record is weak, and modern accounts lean on goal activation: The moment a new user first gets the value the product promised. rather than tension.
How it shows up in software
Progress bars that start partly filled, incomplete profile prompts, draft states and half-finished checklists all cite this effect. What products actually rely on is a visible open loop the person can see and close, which is closer to goal reminding than to a memory phenomenon. Treat the visibility of the incomplete state as the working mechanism.
Using it well
- Make an incomplete state visible and closable in the same view, so the reminder arrives with the action attached.
- Save drafts automatically and show where they live, so an interrupted task can actually be resumed rather than merely nagged about.
- Cap open loops. One profile completion prompt is a reminder; five simultaneous incomplete indicators is noise.
- Let people dismiss a loop permanently. A checklist item that cannot be declined is not a task, it is a fixture.
Where it turns manipulative
- Manufacturing incompleteness that has no value to the user, such as a profile meter that will never reach 100 percent or a checklist with a permanently unchecked upsell.
- Cliffhangers in notifications that withhold information the person already has a right to see, so the only way to close the loop is to open the app.
- Leaning on an effect this shaky in a design review. If a decision needs evidence, cite the goal-reminding work, not Zeigarnik.
Where you have seen it
LinkedIn
Profile strength is shown as a partly filled meter with the next missing field named underneath.
Google Docs
Closing a document mid-sentence preserves the draft and the cursor position, so the open loop is resumable rather than lost.
Trello
Checklists on a card show a fraction complete and a progress bar that fills only as real items are ticked.
What the research says
- Zeigarnik, 1927Contested
Participants interrupted partway through simple tasks recalled those tasks roughly twice as often as tasks they had finished.
Small samples, procedures that are hard to reconstruct from the paper, and effect sizes that later work has not matched.
- Van Bergen, 1968Contested
A review of replication attempts found the effect appeared inconsistently and depended heavily on how interruption was framed to participants.
- Masicampo and Baumeister, 2011Mixed evidence
Unfulfilled goals intruded on unrelated tasks, and writing a specific plan for the goal removed the intrusion.
This supports goal activation rather than the original tension account, and comes from a literature with its own replication problems.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.