The Case Against Tutorial Hands

Look at enough playable ads and the pointing hand becomes invisible through sheer repetition. It arrives about a second in, taps a target, sometimes pulses, sometimes drags a dotted line across the screen. It is close to universal.

It is also close to indefensible, in a specific and unusual way. Ask people who make these ads whether the hand is good and you rarely get a defence. You get an explanation — the player has a second to understand, there is no room for instructions, the hand solves it in every language at once. All true. None of it is an argument that the hand is good, only that it is necessary.

That gap is the interesting thing, and it is what this piece is about. When an element is universal and unloved, it is usually not serving the purpose everyone assumes. Our position is that the hand is not a design choice at all. It is an instrument that measures something, and what it measures is whether the mechanic underneath it works.

Here is the reframing, stated plainly: an ad that needs a tutorial hand has already failed a test, and the hand is what failure looks like rather than what fixes it.

The test is legibility. A playable ad has to make a stranger act correctly within about a second, with no instructions and no prior knowledge. A mechanic that passes that test needs no hand — the screen shows one obviously incomplete thing, and a person's first instinct is the correct move. A mechanic that fails it needs something bolted on to tell the player what to do, and the hand is the cheapest available bolt.

Framed that way, the hand stops being a question about tutorial design and becomes a reading on the mechanic. Its presence says: at this point in this ad, a stranger would not know what to do. That is worth knowing, and it is exactly the information that gets discarded when the hand is treated as a standard component rather than as a signal.

This also resolves the puzzle from earlier. The reason nobody defends the hand is that there is no argument for it that is not, underneath, an argument about the mechanic being unclear — and nobody wants to make that argument about work they just finished.

If that is right, the obvious question is why the practice is universal rather than rare. The answer is not ignorance. It is that the two available fixes have wildly different costs, and only one of them can be done on a Thursday afternoon.

A playtest reveals that new players hesitate. There are exactly two responses. Redesign the mechanic so the correct action is obvious — which means new art, new layout, possibly a different mechanic entirely, and a return to the start of the process. Or add a hand, which takes an hour and demonstrably fixes the hesitation in the next test.

Both close the ticket. One closes it in an hour. The hand therefore wins every time it is compared against the alternative on cost, and it keeps winning, which is how something nobody advocates for becomes something almost everybody ships.

The trap is that the hand does genuinely work. Hesitation drops, the completion rate improves, the numbers move in the right direction. It is not a placebo, and that is precisely what makes it durable — a fix that works on the measured symptom will always beat a fix that addresses the cause, when the symptom is what gets measured.

The case against is not that the hand is ugly or that it takes up room. Those are trivial. The cost is that it changes what the player is doing, and it does so in the direction the format can least afford.

A playable ad sells by letting somebody discover something. They look at an unfamiliar screen, form a guess, act on it, and find out they were right. That sequence is short and it does real work — the person who worked out the mechanic has learned something about the product and has a small stake in having been correct. It is most of what distinguishes the format from video.

A hand pointing at the only valid target removes every step of that except the last. There is no guess, because the answer is on screen. There is no discovery, because nothing was concealed. The player performs a gesture they were instructed to perform, which is a smaller thing in every respect that matters, and the ad has spent its central advantage to save itself a second of confusion.

There is a second cost, less discussed and more practical. A hand teaches the player to wait for instructions. Once an ad has established that it will indicate the correct action, the player stops examining the screen and starts watching for the next prompt — which means every subsequent moment of the ad also needs one. The hand is not a single intervention. It sets a convention the rest of the ad has to keep.

Three qualifications, because a position stated this strongly is easy to overapply and we do not think the hand should be banned.

Some mechanics genuinely cannot be made self-evident, and are still worth showing. A gesture with no physical analogue, or a rule that only exists inside this product, may simply need indicating. That is a legitimate use — and it is also a reason to check whether the mechanic belongs in an ad at all, since a rule that needs explaining is expensive at this length.

A hand used as help rather than instruction costs much less. One that appears only after the player has hesitated, indicates a region rather than an exact point, and disappears once they have acted correctly preserves most of the discovery. That is the compromise our teardown of fifty ads landed on, and nothing here contradicts it. The argument in this piece is about where the conversation should start, not about which hand behaviour is best.

The evidence is not settled, and we are not pretending otherwise. This is a reasoned position about what the hand indicates, not a measured claim about conversion. It is entirely possible that hands improve outcomes on average precisely because most ads have the legibility problem described here — which would make them a correct local fix and still a symptom. Separating those two would take a controlled test rather than an argument, and that test is on our list.