It's one of the most-used terms in education technology, and one of the least explained. Here's what it actually means for how your child practises maths.
25 August 2026 · 6 min read
Most maths apps now describe themselves as "adaptive" somewhere on their homepage. It's become a label almost as common as "fun" or "engaging" — which makes it a fairly meaningless one unless it's explained.
So it's worth being precise about it. What does adaptive learning actually mean, what does it not mean, and how would you tell the difference between a platform that's genuinely adaptive and one that's just using the word?
The word gets shortened, in most people's minds, to "the questions get easier or harder." That's part of it, but it's a thin version of the idea — and on its own, it isn't really adaptive at all. A simple ladder that bumps the difficulty up after a correct answer and down after a wrong one is closer to a difficulty slider than a learning system. It reacts to whether your child got something right. It doesn't ask why.
A genuinely adaptive system uses a learner's responses to decide what should happen next — not just how hard the next question should be, but what it should be about, what kind of support to offer alongside it, and when to move on versus when to go back and repair something first.
That distinction matters because a score alone doesn't tell you much. If your child gets 6 out of 10 on a fractions exercise, you know four answers were wrong. You don't know whether that's four different misunderstandings, one misconception repeated four times, or a handful of careless slips under time pressure. Those need completely different responses, and "here's another fractions worksheet" only happens to be the right one for some of them.
We wrote about this in more detail in why more maths practice isn't always the answer — the short version is that a wrong answer is evidence, not just a mark.
Yes, and it's worth naming properly. When a learner gets stuck, the useful response isn't to reveal the correct answer, and it isn't to say nothing and let them keep guessing either. It's to give just enough support to help them get unstuck under their own steam — a hint that nudges without giving the method away, a worked example that walks through the reasoning using their own numbers, or an explanation that names the idea they've missed. That's scaffolding: temporary support that's there while the understanding is still forming, and that gets withdrawn as it becomes secure. It's the same principle a good tutor uses sitting next to a child — offer the smallest hint that gets them moving again, not the whole answer.
The reason this matters for adaptive learning specifically is that scaffolding only works if it's tied to the actual problem in front of the child. A generic worked example from a textbook, using different numbers and a different context, asks the child to do a second piece of translation work on top of the first — figure out how someone else's example maps onto their own question. A worked example built from their own numbers removes that step, so the explanation lands on the actual reasoning that went wrong.
If a child struggles with a genuinely hard question, handing them another question at the same level straight after usually just produces a second wrong answer. A responsive system can step back to the prerequisite skill underneath, confirm that's solid, and then return to the original level of challenge — not to make things permanently easier, but to make sure the child is actually ready when they get back there.
Put together, this is less a single feature and more a loop that repeats every time a child answers a question:
Assess → Understand → Practise → Adapt → Reassess → Progress
Each answer feeds the next decision. That's really the whole definition in one line: a system where what a learner does next depends on what they've just shown they know.
If you're weighing up an online maths platform for your child, it's worth looking past how many questions it has and asking what it actually does with the answers:
A platform that repeats the same worksheet with different numbers when something's wrong isn't adaptive, even if the marketing says otherwise. We wrote more about spotting that pattern in why "drill and kill" maths apps get it wrong.
Rather than giving every learner the same sequence of questions, the system uses each answer to help decide what the learner should practise next — where a mistake points to a specific misconception, it can offer scaffolding built from the child's own numbers, temporarily step back to a prerequisite skill when one's missing, and increase the challenge again once that's solid.
Whether a child is working on number basics, fractions, 11+ preparation or GCSE maths, the principle stays the same: start from what they actually know, find what needs attention, and build from there.
Don't just practise more. Practise what matters next.
Try a free Grasp Maths practice session and see how the engine responds to what your child actually knows.