Curiosity Gaps: The Oldest Trick in Advertising, Rebuilt for Touch
Curiosity is the most reliable lever in advertising and the most thoroughly exhausted. Two decades of headlines that promised an answer and delivered a slideshow taught everyone to spot the shape of a withheld thing — which is the problem any ad using it now has to solve first.
In short
- Curiosity works on a gap: you have to know enough to feel that something specific is missing. Knowing nothing produces no curiosity at all.
- The feeling is genuinely uncomfortable, which is why it drives action — and why a gap held open too long converts into irritation rather than interest.
- Two decades of clickbait taught people to recognise a withheld answer on sight. The text version of the trick is largely spent.
- What touch changes is who closes the gap. A headline makes you request an answer from someone else; an interaction lets you find it out yourself.
- The rule that survives all of this: open a gap you intend to close, in the place you opened it.
Curiosity is not a general appetite for information. It is much more specific than that, and the specificity is the useful part.
You do not feel curious about things you know nothing about. Nobody is consumed by a burning need to know the export figures of a country they cannot name. Curiosity requires that you already have most of a picture and can feel the shape of the piece that is missing — which is why the sensation is so often described as a gap rather than as an appetite.
That has an immediate practical consequence. To make someone curious you have to give them something first. A question that lands in an empty space produces nothing; a question that lands next to three facts they already hold produces an itch.
This is the oldest working idea in advertising, and it has been running more or less unchanged since advertising began. What has changed, dramatically and recently, is how much people trust it.
The gap is uncomfortable, and that is the mechanism
It would be pleasant to think curiosity is a positive feeling — the joy of learning, the pleasure of discovery. Sometimes. But the version that drives behaviour is closer to an itch.
An open question produces mild, genuine discomfort. It occupies attention slightly. It makes the unresolved thing more available than it should be. And the relief when it closes is disproportionate to how much the answer mattered — which is why people click things they do not care about and feel faintly annoyed at themselves afterwards.
Understanding it as discomfort rather than delight explains the failure mode that catches most advertising. If the gap is a pleasure, holding it open longer is generous. If the gap is an itch, holding it open longer is an imposition, and there is a point where the person stops wanting the answer and starts wanting to be left alone.
That point arrives much sooner than most creative assumes. And it arrives sooner still for anyone who suspects the answer is not coming at all.
Curiosity is an itch, not an appetite. Holding a gap open is not building anticipation — it is asking someone to keep not scratching.
How the open web burned it out
For most of its history this technique was limited by cost. Opening a curiosity gap required buying a page, a spot or a billboard, which meant the people using it had something to protect and a reason not to squander the audience's patience.
Then distribution became free and the constraint disappeared. What followed is worth setting out as a sequence, because the shape of the decay is the thing worth learning from.
The same trick, running for a century, with the cost of using it falling to nothing at the end:
Early print
Scarce and mostly honest
A gap cost money to open, so it was usually followed by something. The advertiser had a name to protect and only one shot at the page.
Broadcast
The gap becomes a format
Teasers before the break, questions answered after it. The audience accepted the structure because the answer reliably arrived a few minutes later.
Early web
The cost collapses
Publishing a headline becomes free and can be tested. The optimisation loop rewards whatever gets clicked, and a gap gets clicked more than a summary does.
Peak clickbait
The answer stops arriving
Headlines are written to open a gap and pages are built to postpone closing it. The economics reward the click, not the resolution, so resolution is where the cutting happens.
Now
The shape itself is the signal
People recognise a withheld answer at a glance and read it as a warning rather than an invitation. The technique did not become less effective psychologically — it became a reliable marker of something not worth opening.
That last row is the situation any advertisement using curiosity now inherits. The underlying psychology is untouched: a gap still produces an itch, exactly as it did in 1950. What changed is that the presentation of a gap has acquired a second meaning, and the second meaning arrives first.
This is the same pattern as the dismissal reflex — a cue that fires before content is assessed. Curiosity-shaped copy now triggers it.
Why this decayed and other techniques did not
Most advertising techniques degrade slowly, through familiarity. Curiosity degraded quickly and thoroughly because the incentive pointed directly at abuse: the click was measurable and instant, the disappointment was not measured at all, and no individual publisher paid for the audience's growing scepticism. Any mechanism with that structure — immediate measurable benefit, deferred and collectively borne cost — will be exhausted at whatever speed the optimisation loop runs.
What touch actually changes
So what is different about a curiosity gap you can put your finger in?
The honest answer is: one thing, and it is not the thing people usually claim. Interaction does not make curiosity stronger. It changes who is responsible for closing the gap.
A headline creates a gap and then asks you to request the answer from someone else. You click, you arrive somewhere, and whether you get what you came for is entirely up to a party whose interests are not yours. Every clickbait disappointment has that structure — the asymmetry is the whole problem.
An interaction creates a gap and hands you the means to close it. What happens if I drag this there. The answer is not supplied by the advertiser on request; it is produced by the thing on the screen in response to what you did. You are not trusting anyone. You are looking.
That is a smaller claim than the usual pitch for interactive advertising, and it is the one we would actually defend.
The same curiosity gap, opened two ways. The psychology is identical; the trust relationship is not.
A gap you must be told the answer to
- You ask; someone else answers.
- Whether the answer comes is their decision, not yours.
- Every previous disappointment is relevant evidence.
- The delay between question and answer is dead time you must accept.
- Being wrong about the payoff costs you something and teaches you to trust less.
A gap you can reach into
- You act; the screen responds.
- The answer is produced by your own action.
- Previous disappointments matter less, because nobody is being trusted.
- Question and answer are separated by about a fifth of a second.
- Being wrong costs a tap, and the correction is itself informative.
The fourth row is worth dwelling on. The interval between opening a gap and closing it is where scepticism lives — it is the window in which a person can wonder whether they are being led somewhere. A headline's gap stays open for the length of a page load, which is long enough to reconsider. An interaction's gap closes almost immediately, and a question answered that fast never becomes a decision about whether to trust anyone.
Which also means the advantage is spent the moment an ad makes someone wait. An interactive ad that opens a question and then plays four seconds of animation before answering it has converted itself back into a headline.
Asked
A gap opened by copy, closed by someone else's decision
Reached
A gap opened by a screen, closed by the person's own action
Waited
Any delay between the two, which is where the scepticism gets in
Gaps that close
The practical work here is mostly about resolution rather than provocation. Opening a gap is easy and most advertising is already good at it. Closing one properly is where the technique either rebuilds its credibility or spends the last of it.
Five rules for using curiosity now that everyone is wise to it.
- 1
Give them the three facts before you ask the question
A gap needs context on both sides. Show enough that the missing piece is specific — this board, these pieces, one move — and the question forms on its own. A question asked before the player understands the situation is noise, because there is nothing for it to be a gap in.
- 2
Close it where you opened it
The gap an ad opens should be answerable inside the ad, not on the store page. Deferring the answer to after an install is exactly the clickbait structure people learned to distrust, and it is recognised as such even when the product is good.
- 3
Answer faster than they can get suspicious
The window where a person decides whether they are being strung along is somewhere under a second. Respond inside it and the question never becomes a judgement about your intentions. This is the single biggest advantage interaction has, and it is easy to squander with an animation.
- 4
Never make the answer the reward for an install
Withholding the resolution until after a download is the version of this that broke the technique in the first place. It converts reasonably well once per person, and it is the transaction people are most reliably annoyed about afterwards.
- 5
Let one gap close before opening another
Stacked unresolved questions do not compound — they read as evasion. A gap that closes cleanly earns the credibility to open the next one, which is how a short ad can hold attention without ever feeling like it is stalling.
The difference between a question and a tease
It is worth naming the line precisely, because the two look almost identical at the moment they are deployed.
A question is opened with the intention of answering it. A tease is opened with the intention of being paid to answer it. Same copy, same mechanism, same itch — and the person on the other end cannot tell which they are getting until the moment of resolution arrives or fails to.
What makes the tease corrosive is not that it is unpleasant. It is that it is indistinguishable from the honest version in advance, which means each instance makes the honest version slightly harder to use. Everyone who opens a legitimate question is borrowing against a credibility that teasers have already drawn down.
The test we would apply is whether the person would feel the exchange was fair if they could see your intentions from the start. A gap you were always going to close survives that. A gap engineered to remain open until money changes hands does not, and the fact that it converts is not a defence — it converts because it is indistinguishable from the honest thing, which is precisely the objection.
A small test that works surprisingly well
Write down the question your ad opens and the moment it closes. If you can point to both, you have a curiosity gap. If you can only point to the first, you have a tease and should expect it to perform well once and poorly thereafter. Teams are often surprised by how many of their creatives fail this on inspection — not through cynicism, but because the opening is the part everyone works on and the closing is assumed to happen somewhere later.
Where this runs out
Two limits worth marking on the argument.
The claim that interaction rehabilitates curiosity is a structural argument rather than a measured one. It follows from a single observation — that a gap you close yourself requires trusting nobody — and the logic holds, but it describes why the format should behave differently rather than proving that it does. Use it as a reason to design a certain way.
There is also a real chance the reprieve is temporary. If interactive advertising adopts the deferred-answer structure at scale — gaps that only close after the install — there is no obvious reason the same decay would not happen again, on the same schedule, for the same incentive reasons. The format is not immune to the economics that broke the headline. It is simply earlier in the cycle.
And curiosity is not sufficient on its own. It gets attention to a place; it does not make the thing in that place worth wanting. An ad that opens a perfect gap onto a product nobody has any use for has produced a moment of interest and no customer, and that is the most common way this technique is wasted.
Everyone who opens an honest question is borrowing against credibility that somebody else already spent. The only way to pay it back is to close the gap where you opened it.
Questions this one tends to raise.
Is a curiosity gap the same as the Zeigarnik effect?
Closely related but not the same, and the difference is what is unfinished. A curiosity gap is about missing information — you want to know something. An interrupted task is about a missing action — you were about to do something. Information can be supplied by anyone, which is why curiosity depends on trusting the supplier. An intention belongs to the person who formed it, which is why it travels better.
Does clickbait still work?
On the click, often yes, which is why it persists. What it no longer does is produce anything downstream — the person arrives already suspicious, leaves quickly, and adjusts their expectations of the next similar thing. It is a technique that borrows from future performance to improve a present number, and it is usually chosen because those two things are measured by different people.
How do I open a gap without any text?
Visually, and usually better than with text. A partially covered object, a mechanism mid-motion, a board one move from resolving — each poses a specific question without a word. This is the most underused version of the technique in interactive advertising, and it has the advantage of not carrying the copy-shaped cue that people now distrust on sight.
Can an end card use a curiosity gap honestly?
Yes, if it is closing a gap rather than opening one. An end card that resolves what just happened and shows where it continues is fair. One that introduces a fresh question answerable only after installing is the deferred-answer structure, and it is the specific move that taught people to distrust this whole family of technique.
Is there a limit to how many gaps one ad can use?
In practice, one or two that actually close. The constraint is not attention but credibility — each unclosed question makes the next one less believable, and an ad that stacks them reads as stalling. A single question, answered fast, buys more of the remaining time than three questions left hanging.
Editorial
The Play Ads Maker team
We build the editor behind this site, which means we spend our days inside the format — its constraints, its export limits, and the question of what actually makes someone tap. We write about what we learn doing it.
The fastest way to understand the format is to build one
The editor is free to open, runs in the browser, and needs no code. Start from a template and have something playable within the hour.
Keep reading
The Zeigarnik Effect
Interrupted tasks stay active in the mind in a way completed ones do not. The useful part for advertising is not the ad itself — it is what survives the walk to the store.
9 min readWhen an Ad Stops Feeling Like One
People consistently report that playable ads feel less like advertising. They are right, and the reason is structural. The harder question is what follows when the defences people built against persuasion never switch on.
10 min readWhy People Skip Ads
Skipping is not usually a judgement about your ad. It is a habit that fires before anyone has looked properly — which is why making the ad better so rarely changes the number.
9 min read