We build a maths app. So when we publish a list of "the best maths apps for kids," you should be suspicious of it — and you'd be right to be.

Here is something more useful instead: the checklist we would use on our own product, and on anyone else's. It's the set of questions we asked ourselves while building CREST Champs, and further down we'll score ourselves against it in public — including the point where we currently fall short.

You don't need to be technical to use it. Ten questions, ten minutes sitting next to your child while they use the app. That's the whole audit.

The wrong question about screen time

Most screen-time advice is a number: so many minutes per day, per age. But the guidance behind the headlines is more interesting than the headlines. Both the American Academy of Pediatrics and the World Health Organization, in their guidance on children and screens, put weight not just on minutes but on what is on the screen and whether an adult is involved — content and co-use matter, not raw duration alone.

Which changes the question. Not "how long was my child on the tablet?" but "what was the screen doing to — or for — my child?" Thirty minutes of passively watching a cartoon character count is not the same substance as fifteen minutes of answering questions pitched at exactly the right level. One is consumption. The other is retrieval — and retrieval is where learning lives.

The question isn't whether your child gets screen time. It's whether the screen is asking them anything.

There is research support for the optimistic half of this, too: reviews and meta-analyses of maths apps and game-based learning generally find positive effects on learning when the app is practice-based and adapts to the child — not when it's videos with a quiz sprinkled on top.

The 10-point checklist

Sit beside your child for one session and ask these, in order.

1. Does the child answer things, or mostly watch things? Count the minutes. If more than half the session is animation, narration or reward screens, it's television with homework's reputation.

2. Does the difficulty adapt? Watch what happens after five correct answers in a row. If the questions don't get harder — and easier after a rough patch — every child is getting the same conveyor belt, and it's the wrong speed for almost all of them.

3. What happens after a wrong answer? This is the most revealing question on the list. A red X and the same question again is not teaching. A good app treats a wrong answer as information: it explains the specific slip, or routes the child somewhere — an easier rung, a different representation, a prerequisite skill. (Wrong answers are far more diagnostic than parents are told; they're rarely careless.)

4. What are rewards tied to? Effort and streaks are healthy anchors. If the sparkly things are tied to purchases — or the app manufactures frustration that a purchase relieves — the design is aimed at your wallet through your child.

5. Are there ads — any at all? In a children's product, our line is simple: zero. An ad network in a kids' app means an attention market operating on your child, and often a data pipeline behind it.

6. What data is collected, and can you actually read the policy? Wherever you live, there is probably a legal floor most parents don't know they're entitled to.

  • In the US, COPPA requires verifiable parental consent before collecting personal information from children under 13.

  • In the UK and EU, GDPR and the Age-Appropriate Design Code push services toward data minimisation and designing in the child's best interests, not the advertiser's.

  • In India, the DPDP Act 2023 requires parental consent for children's data and prohibits behavioural tracking of children and targeted advertising at them.

Different laws, one shared essence: parental consent, no behavioural tracking of children, no ads targeted at children. If an app's privacy policy is unreadable, assume that's a choice.

7. Can a session be short and still feel complete? A well-designed session has an ending — a natural "done for today." If the app is engineered to be endless (one more level, one more chest, a penalty for leaving), it's optimised for retention, not learning.

8. Does it report to parents in skills or in vanity metrics? "Subtraction within 20 is shaky" is a report. "247 stars, 96 minutes this week" is a scoreboard. Only one of them tells you what to do next.

9. Is the maths curriculum-anchored, or a random question pile? Questions should map to the skills your child's school system actually builds, in a deliberate order — because maths is sequential, and most struggles trace back to one specific missing skill, not a general shortage of questions.

10. The exit test. After the screen goes off, does your child ever want more maths — asks you a number question at dinner, points out a pattern, wants to beat yesterday? An app that only generates appetite for the app has taught the wrong thing.

A note on scoring: points 1–3 are about learning, 4–7 about the business model, 8–10 about respect for your family. Most apps pass one group and quietly fail another — a beautifully adaptive app that mines data, or a scrupulously private one that's really a worksheet with sound effects (and static worksheets have their own limits). The checklist works because it forces all three conversations at once.

A mother we spoke to ran this checklist on her son's "educational" app — the one that had bought the family four months of guilt-free tablet time. Question 3's answer turned out to be: wrong answers were simply skipped, never revisited. Question 5's answer arrived on its own, in the form of a candy advert between levels. Ten minutes with a checklist did what four months of "he's on a learning app, it's fine" hadn't.

Scoring ourselves, in public

Fair is fair. Here is where CREST Champs stands on its own checklist.

Ads (point 5): zero. No ad networks, no sponsorships, no "free with ads" tier. Nothing on the screen is trying to sell your child anything.

Data (point 6): no behavioural tracking. We run no analytics trackers, no advertising pixels, and no third-party trackers of any kind, for users of any age — and student work is never sent to third-party AI services. The privacy policy is written to be read by a parent, not a lawyer.

Adaptivity (points 2 and 3): yes, and it's the whole product. Every answer adjusts what comes next. A wrong answer isn't a red X; it's a signal that routes practice toward the prerequisite that's actually missing.

Reports (point 8): skills, not stars. Parents see which specific skills are secure and which are shaky — "subtraction within 20 needs work," not a minutes counter.

Where we're still improving: offline access. Practice currently needs a connection, and we know plenty of family maths time happens on trains, in waiting rooms and in places where the signal doesn't reach. We're not going to pretend that's a feature. It's on the list.

The mechanics we deliberately don't use

One more section, because most app reviews skip it. The persuasive-design toolbox — variable-reward loops, streak-loss dread, countdown timers on "limited" rewards, escalating chests — is extremely effective. That's the problem. Those mechanics are built to maximise time-in-app, and time-in-app is not learning; it can be its enemy, since short daily sessions beat long compulsive ones for actual retention.

So we use streaks and points the boring way — as a gentle reason to show up tomorrow — and we let sessions end. A ten-minute session that finishes cleanly, at the right difficulty, is worth more than an hour a child couldn't put down.

Ten minutes, once

Whatever app your child uses — including ours — run the checklist once, sitting beside them. You're not looking for a perfect score; almost nothing scores ten out of ten (see above). You're looking for the dealbreakers: ads at children, wrong answers that vanish, sessions built to never end.

And if you'd like to see how we answer the ten questions in practice, CREST Champs' daily adaptive practice is free: ten minutes a day, questions that adapt to your child, wrong answers treated as clues rather than failures, and a parent report that names skills instead of stars. Bring the checklist. We wrote it to be used on us.