Zeigarnik Effect
Unfinished tasks stay in memory longer than completed ones. An open question creates cognitive tension that demands resolution. This is why prompts work.
The Zeigarnik effect is the engine behind every open loop — every question that won't leave you alone, every mantra that pulls you back to what matters, every prompt that creates forward motion.
The Mechanism
| State | What happens | Example |
|---|---|---|
| Open loop | Mind keeps returning to the unresolved task | "What percentage of your pipeline found you?" |
| Cognitive tension | Discomfort from incompleteness drives action | You start counting your inbound vs outbound leads |
| Resolution seeking | Person engages to close the loop | They click the depth link, book the call, fill the grid |
| Completed | Tension releases, memory fades | The question is answered — but a new one opens |
The skill is not closing loops. The skill is opening the right ones.
Why Prompts Work
A prompt is a deliberately opened loop. "Why does this matter?" creates tension that only your own answer can resolve. The question stays active in working memory until you respond — and the response itself opens the next question.
| Instrument | How it uses Zeigarnik |
|---|---|
| Mantra | Short phrase opens a loop that resets orientation under pressure |
| Tight Five | Five questions, each opening a loop the next one deepens |
| Prompt Deck | Each slide ends with a prompt — the open loop pulls the reader to the depth link |
| Questions section | Every page closes with Socratic questions that create forward motion |
A good question is a Zeigarnik device. It manufactures the exact cognitive tension that makes someone lean forward.
Pitch and Prompt
In a pitch, each slide has two layers:
| Layer | What it does | Zeigarnik role |
|---|---|---|
| Pitch (bullets) | Confirms what they already feel | Builds tension — "yes, that's my problem" |
| Prompt (question) | Opens a loop they can't close | Creates the pull — the question stays active |
| Depth (link) | Offers resolution | The only way to close the loop is to go deeper |
The pitch is the berley. The prompt is the hook. The depth link is the net.
Stacking with Other Biases
Zeigarnik is strongest when paired:
| Bias | How it stacks |
|---|---|
| Confirmation | Confirm what they suspect (pitch), then open a loop about what they haven't considered (prompt) |
| Loss aversion | Name what they're losing (pitch), then ask what they'd gain (prompt) — the loss stays open |
| IKEA effect | The prompt invites co-creation — their answer becomes their asset |
| Hyperbolic discounting | Make the near-term loop vivid — "What would change in your next 30 days?" |
For Agents
Agents need prompts too. A well-structured system prompt is a Zeigarnik device — it opens loops the agent must resolve through its work. The Tight Five works as alignment architecture for both humans and phygital agents because the mechanism is the same: open the right loops, and the work follows.
Risk
Too many open loops create anxiety, not action. The balance:
- 1-3 loops = productive tension, forward motion
- 5-7 loops = cognitive overload, scattered attention
- 10+ loops = paralysis, anxiety, burnout
The Tight Five is five, not ten, for this reason. Miller's Number under pressure.
Context
- Tight Five — The five prompts that use this effect
- Mantra — The simplest Zeigarnik device
- Prompt Deck — The instrument that renders open loops as slides
- Questions — Eternal feedback loop
- Persuasion — The rhetoric framework
- VVFL — Feedback loops that compound
Links
Questions
If unfinished tasks stay in memory, what happens when you deliberately structure every page to end with an open question?
- How many open loops can you hold before tension becomes anxiety — and is that number different for agents than for humans?
- If a mantra is a Zeigarnik device, what distinguishes one that resets you from one that haunts you?
- When someone clicks the depth link to close the loop, does the resolution create a new loop — and should it?