The Grammar of Interaction: Tap, Hold, Drag, Swipe and What Each Says
Every playable ad chooses a gesture, usually without much deliberation, and that choice quietly tells the player what kind of product this is. Four gestures, four different sentences.
In short
- Gestures are not interchangeable. Each one describes a different relationship between the player and the object, and players read that description without noticing.
- A tap is a command: you ask, the screen answers. It is the fastest gesture to understand and the one that gives the player the least ownership.
- A drag is manipulation: the object follows your hand and you are responsible for where it goes. It produces the strongest sense of control and costs the most time.
- A hold is a commitment, which makes it the only gesture that can express risk or accumulation — and the easiest to get wrong, because nothing on screen tells you it is available.
- A swipe is a dismissal or a direction. It is the one gesture players already use for something else, which is both its advantage and its danger.
Most conversations about gestures in ad creative are about feasibility. Can we do a drag in this format, will a swipe conflict with the feed, does a hold work on every device. Reasonable questions, and they skip the more interesting one.
A gesture is not just an input. It is a small statement about what the player is in relation to the thing on screen — whether they are instructing it, handling it, committing to it or pushing it away. Players do not articulate that, but they respond to it, and a mismatch between the gesture and the product it represents is one of the quieter ways an ad ends up feeling wrong without anybody being able to say why.
Four gestures do nearly all the work in this format. They are worth taking one at a time.
Tap: the command
A tap is the shortest possible sentence. You indicate a thing, and something happens to it. The player is an instructor and the screen is the one doing the work.
That makes it the most legible gesture in the format, and the reason it is the right default. A stranger with no instructions will tap something before they try anything else, and a tap requires no explanation, no precision beyond hitting the target, and no time.
What it does not produce is much sense of authorship. The player pointed; the game acted. The result belongs to the system rather than to them, which is fine when the product is genuinely about choosing rather than doing — a puzzle where the interesting part is which piece you pick, a game where the skill is in selection rather than execution.
Where it goes wrong is when a tap is used for something that is physically continuous in the real product. If the game is about moving things, representing that as a tap tells the player the game is about choosing things, and the first minutes after install will not match what they expected.
Drag: the manipulation
A drag is the only gesture where the player and the object are in continuous contact. The thing follows the finger, tracks it, and stops where they let go. It is the closest a touchscreen gets to handling something.
This produces the strongest sense of control available in the format, and it is the reason drag-based mechanics feel more satisfying than their tap equivalents even when the underlying game is identical. The player did not request an outcome — they produced one, continuously, with their own hand.
It also costs the most. A drag takes longer than a tap, requires the player to be accurate for the whole duration rather than at a single instant, and suffers most from the occlusion problem: their thumb is on top of the object for the entire gesture.
Two rules make drags feel good rather than fussy. The object must track the finger exactly, with no lag and no easing, because any delay reads as the screen being slow rather than the player being imprecise. And the drop target should be generous — snapping to a nearby valid position feels responsive, while requiring exact placement feels like a test.
A tap asks the screen to do something. A drag does it yourself. Those are different claims about who is in control, and players hear the difference without noticing they heard it.
Hold: the commitment
A hold is the gesture with the most interesting properties and the worst discoverability.
What makes it distinct is that it has duration, and duration lets it express things no other gesture can: filling, charging, accumulating, risking. A hold is the only way to represent a decision that gets better the longer you commit to it and worse if you commit too long. Aim a shot, fill a glass, charge a jump — all of these are holds because all of them are about how much rather than whether.
The problem is that nothing on a screen looks holdable. A tap target and a hold target are visually identical, and a player given no instruction will tap and get a fraction of the intended result. So a hold almost always requires teaching, which in a format that cannot afford instructions means teaching through failure: the player taps, sees a small unsatisfying version of the outcome, and works out that more contact means more effect.
That works, and it costs a couple of seconds. Whether it is worth it depends on whether the product is genuinely about accumulation. If it is, nothing else represents it honestly. If it is not, the hold is an affectation that spends the format's scarcest resource.
Swipe: the dismissal
A swipe is the most loaded gesture available, because the player already uses it for something else on the same screen.
Outside your ad, a swipe means move on, get rid of this, show me the next thing. That association does not switch off because your ad arrived. It means a swipe inside a playable is read first as an attempt to leave and only second as a game action, which is an unusual amount of baggage for an input method to carry.
The upside is that it is the one gesture requiring no teaching whatsoever, and it expresses direction and force naturally — a flick has a speed and an angle that a tap does not. Games about launching, throwing, steering or sweeping are genuinely well served by it.
The risk is specific and worth stating plainly: in a feed placement, a swipe can be intercepted by the container rather than the ad, so the player performs your mechanic and the feed scrolls instead. That is not a design problem you can fix with better art. Test swipe mechanics inside a real placement before committing to one, because the gesture that works perfectly in a preview can be unusable in the environment the ad actually runs in.
The four gestures, and the sentence each one speaks:
| Gesture | What it says, and what it suits |
|---|---|
| Tap | "Do this." The player instructs and the screen performs. Fastest to understand, weakest sense of ownership. Right for games about choosing. |
| Drag | "I am moving this." Continuous contact and full responsibility for the outcome. Strongest sense of control, slowest to perform. Right for games about handling. |
| Hold | "I am committing to this." The only gesture with duration, so the only one that can express charging, filling or risk. Needs teaching. Right for games about accumulation. |
| Swipe | "Away, or that way." Needs no teaching and carries direction and force. Collides with the feed's own meaning. Right for games about launching or steering. |
Choosing the one that tells the truth
The practical question is not which gesture performs best. It is which gesture describes the product accurately, because an ad that uses the wrong one produces installs that do not survive contact with the real thing.
The test is simple: what does the player physically do in the actual game, and what is the shortest honest version of that? A merge game is a drag, because merging is moving something onto something else. A card game is a tap, because playing a card is a choice rather than a manipulation. A game about timing is a hold, because timing is duration.
Where teams go wrong is optimising the gesture for the ad in isolation. Taps convert well because they are instant and require nothing, so there is a steady pull toward representing everything as a tap. The ad gets faster, the install numbers improve, and the product turns out to be a different kind of game than the one demonstrated.
Four checks before settling on a gesture.
- 1
Name what the player does in the real product
In one verb. Move, choose, aim, launch, sort. The gesture should be the touchscreen version of that verb, and if the honest answer is a verb your ad does not use, the ad is describing a different game.
- 2
Use one gesture, not two
A playable ad has time to teach exactly one interaction. An ad that requires a tap for one thing and a drag for another has spent its budget on tuition rather than demonstration, and the second gesture is usually the one nobody discovers.
- 3
Check it against the placement
Swipes can be swallowed by a scrolling feed and holds can be interrupted by the container. Whatever you choose, run it inside a real ad slot rather than a preview window before the creative is finalised.
- 4
Make the gesture discoverable without a sentence
A tap is self-evident, a drag needs a visual hint of movement, a hold needs to be taught through a first attempt that under-delivers. Whichever you pick, the teaching has to happen in the first two seconds and without words.
When in doubt, drag
If the product genuinely involves moving things and you are choosing between a tap and a drag on grounds of speed, the drag is usually worth the extra second. It produces more sense of control, represents physical mechanics honestly, and gives the player something their hand did rather than something they requested. The exception is any placement where precision is unrealistic — a small target on a moving feed, for instance, where a tap is simply more reliable.
Where this is convention rather than law
The meanings described here are learned, not innate. A swipe means dismissal because a decade of interfaces taught it, and a different decade could teach something else — conventions on touchscreens have already shifted more than once, and the pull-to-refresh gesture that everyone now understands was baffling when it appeared.
There is also genuine variation between audiences. Players who spend hours in mobile games read gestures more fluently than people who use a phone mainly for messages, and an ad served to a broad audience meets both. The safest assumption is the less fluent reader, which is another argument for the tap as a default and for teaching anything more complex through a first attempt.
What is not convention is the relationship between duration and meaning. A gesture with duration can express accumulation and one without it cannot, regardless of what anyone has learned. That part is structural, and it is why the hold survives despite being the hardest gesture to discover.
Questions this raises in practice.
Should an ad ever use more than one gesture?
Rarely, and usually not. The format has time to teach one interaction properly, and a second one tends to be discovered by a small fraction of players while confusing the rest. If a product genuinely needs two, the honest options are to demonstrate the more representative one and accept that the ad is partial, or to build two separate creatives.
Why do so many ads use a drag when the game is a tap game?
Because drags hold attention slightly longer and feel more involving, so they can improve the numbers an ad is judged on. It is the same drift described above, running in the opposite direction, and it has the same consequence — the player arrives expecting a physical game and finds a selection one.
Is a hold worth the teaching cost?
Only when duration is the point. If the product is about charging, filling, aiming or risking, nothing else represents it and the two seconds are well spent. If the hold is decorative, it is the most expensive possible way to add nothing, because the player who taps gets a broken-feeling result and may leave before the mechanic is understood.
What about double taps and multi-finger gestures?
Neither belongs in a playable ad. A double tap is undiscoverable without instruction and frequently registers as two separate taps; multi-finger gestures assume a grip that one-handed players do not have. Both are reasonable in an app somebody chose to install and unreasonable in an ad somebody did not.
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
Why a Button Needs to Feel Pressed
A response to a tap is not decoration. It is the only proof the player has that they caused something — and if it arrives late enough, they stop believing they did.
8 min readDesigning for the Thumb
A phone screen is not a small canvas. It is a surface with a curved reachable region, a blind spot that moves, and a corner almost nobody can comfortably touch.
11 min readA Story in Eight Seconds
Eight seconds is not enough time to tell a story. It is more than enough to imply one that is already happening and let the player finish it themselves.
8 min read