A six-year-old will not read your instructions. Many are only starting to read at all, and the ones who can will not spend that effort on a menu when there is a bright button right there to press. Design for that user and a surprising amount of ordinary interface design stops working.
The Interactive Kids Games Hub on my projects page is over 100 games in one place: classics like Tic-Tac-Toe and Snake next to educational ones like Math Quiz, ABC Clicker and a Solar System Quiz. Every one of them was held to the same bar: bright visuals, and controls simple enough that a young child can start without reading anything or asking anyone. That bar sounds modest. It rules out a lot.
Fun or educational is a false choice
Children's games tend to be one or the other. The fun ones teach nothing, and the educational ones get closed after two minutes, usually because adults designed them by reasoning about children instead of watching them play. Holding both kinds to one bar refuses that split: games built for young learners, not adult software adapted down to them.
Bigger targets than seem reasonable
Small hands are not precise, and on a tablet the finger covers the very thing it is pressing. A near miss that lands on the wrong button is worse than a miss on nothing. So: large targets, generous space between them, and nothing important at the edge of the screen, where the hand holding the tablet will press it by accident.
Some gestures are off the table altogether. Double-tap, long-press and precise dragging all assume motor control that many young children do not have yet, and hover does not exist on a touchscreen. One tap should do one thing.
Pictures and sound do the explaining
If the user cannot read the label, the label has to be something else. For a game like Snake, that means arrows rather than the word "up". For a planet quiz, a picture of the planet beside its name. For counting, a group of dots next to the digit.
Sound matters more than it does for adults. A spoken question gets past reading entirely, and a short sound on every tap confirms that something happened. Keep it pleasant, and give the grown-up in the room an obvious way to turn it off.
Answer every tap, immediately
An adult will wait a second to see whether a button worked. A child presses it again, and again. If nothing visible happens at once, the screen looks broken, and five presses later the game is somewhere nobody intended. Every tap should change something on screen straight away, even if the real work takes longer, and the code should expect the same button to be pressed several times in a row.
No dead ends
An adult who gets stuck looks for a back link. A child who gets stuck simply stops, and wanders off. Every screen needs a visible way forward and a visible way back to the start, both as pictures. No dialog boxes that ask a question in words. No step that needs typing. When a game ends, "play again" should be the biggest thing on the screen.
Nothing that punishes
This is where educational games most often go wrong. A wrong answer earns a red cross and a harsh buzzer, a countdown runs out on a child who counts slowly, and the lesson learned is that the game is unpleasant. A wrong answer is part of learning: show the right one gently and let the child try again. Timers belong in games where speed is the point, not in games where thinking is.
Losing is not the same as being punished. In Tic-Tac-Toe or Snake, losing is part of the game, and children understand that perfectly well. What matters is what losing costs. If starting again is one big button and no waiting, losing is fine. If it means finding the way back through three menus, it is not.
What a children's game should never ask for
Young children cannot reliably tell an advert from the game around it. A bright box that moves is something to tap, and one stray tap can take a child out of the game and into an app store or a website nobody checked. Adverts do not belong in games made for children.
The same goes for data. A game for a young child has no reason to ask for a name, an email address or a login, and no reason to track who played what. Many countries have specific rules about children's data, and the simplest way to respect all of them is to collect nothing. Data you never gathered cannot leak or be sold, and the parent handing over the tablet does not have to read a privacy policy first.
The best children's game asks for nothing: no reading, no help, no account and no patience.
Adults put up with a confusing screen because they want what is behind it. A six-year-old does not. If the game answers, they stay; if it does not, they leave without a word. No tester is more honest.
