Retrieval Is Learning

Without Retrievalism, Pupils Will Not Learn What We Teach

Retrieval is learning. Not in the literal, pedantic, cognitive-science sense. Of course, learning has two components: encoding and retrieval. But in the practical, educational, on-the-ground sense that matters for school improvement, retrieval is learning. It is the mechanism that makes knowledge stick. Encoding matters, deeply. But without systematic, spaced retrieval, its impact evaporates.

And that is the problem: schools obsess over encoding because it is visible, familiar, intuitive, and they neglect retrieval because it isn’t.

I believe wholeheartedly that this is one of the biggest structural errors in British education. It explains the endless cycle of reteaching, the illusion of progress in books, the summertime chorus of “I can do it in lesson, but I can’t do it in the exam” and the unshakeable sense that pupils leave school knowing far less than they were taught. It’s not because teachers or students are bad, lazy, or disengaged but because we build the entire system around only one half of learning.

Schools have the perfect structural conditions for spaced retrieval practice. We have pupils every day, for years, with predictable timetables, regular contact and enforced attendance. No other environment in a young person’s life is remotely as well-positioned to strengthen long-term memory. However, spaced retrieval practice is not treated as a fundamental part of learning. It is not given the same status as encoding. Teaching overshadows learning.

Learning Is Change in Memory

For me, learning is a sustained change in memory.
That change has two parts:

• Encoding: moving information from working memory into long-term storage.

• Retrieval: pulling stored information back into working memory, strengthening it as you do.

Everything else is metaphor.

I agree learning cannot happen without encoding. That part is obvious, uncontroversial and, therefore, safe. Schools put their weight behind it: explanations, modelling, scaffolding and worked examples. Encoding is the glamorous half of learning.

Retrieval is the quieter half: less visible, less intuitive, and hugely overlooked. But it is the half that creates durable learning, the half that decides which knowledge survives, and the half that makes pupils capable of thinking critically, solving problems, and applying knowledge flexibly.

This is why I’m comfortable saying teaching is encoding and learning is retrieval.
Not because it is technically perfect, but because it is practically perfect.

If a concept cannot be retrieved, it has not been learned in any meaningful, educationally relevant sense. If knowledge can be retrieved, it becomes the foundation for everything else we want pupils to do, from extended essay arguments to complex mathematical reasoning. The semantics aren’t the point.

Why Schools Put Their Weight Behind the Wrong Half

Encoding looks like learning.
Retrieval feels like learning.

Encoding produces tidy books, completed tasks, answers that seem secure.
Retrieval produces hesitation, struggle, gaps, uncertainty. When pupils forget, teachers assume the encoding was poor. When retrieval feels difficult, teachers assume it’s not working. When quizzes reveal gaps, teachers assume pupils “weren’t listening.”

But the research is remarkably consistent: most forgetting is not caused by bad teaching or bad students, it is biological, predictable and unavoidable without sufficient retrieval.

Encoding without retrieval is like planting seeds but never watering them.

The Missing Half of the System

Retrieval is rarely systematic. It is done sporadically, varying massively from school to school, department to department, teacher to teacher. I’ve spoken before about the failures of the most common forms of ‘retrieval’ we see in schools, primarily five-minute starters and spiral curricula. In most cases, these are essentially illusions of retrieval with no real impact. Teachers teach, pupils fail to retrieve weeks, months or years later, teachers reteach and the cycle continues.

Disadvantaged pupils suffer most. Subjects with high cognitive load (like maths) suffer most obviously. Essay-based subjects suffer quietly, because teachers don’t always realise how much foundational knowledge fluency their subject requires for genuine analysis. Context matters - subject, school type, pupil background, assessment paradigm - but retrieval matters everywhere.

This is precisely why I built the Spacing Calculator.

The Spacing Calculator – A Recap

Manually scheduling retrieval is almost impossible. I tried. Most teachers who understand retrieval theory try once and then give up. You cannot track retrieval intervals for hundreds of “atoms” (thank you Kris Bolton) with a pen, notebook and hard work.

The spacing calculator changed everything.

You input:

• your atomised list of knowledge

• the date it was taught

• the date it was successfully retrieved

And the system outputs a personalised, expanding-interval calendar using a simple multiplier. It turns cognitive science into timetabled reality, gives retrieval rhythm and gives effective encoding a chance to actually show its impact.

Unstoppable Learning gave me, what I believe to be, a near-perfect, complete system of encoding knowledge. The Spacing Calculator allowed me to see the sustained impact over time.

The Results

I’m aware that internal data means very little to anyone outside of the school it’s from. However, this feels worth sharing.

Our most recent year 11 mock results are staggering.

· More students achieved a grade 9 in this mock than have ever achieved a grade 9 in the actual exams.

· There were more of all of the top grades when compared to the previous year’s mocks (and even when compared to the previous year’s real exam data!)

· Given even similar progress to previous years from now until their exams this would likely mean 10 or more pupils achieving a grade 9.

· This is from a cohort with similar prior attainment compared to the previous year and from a school which has produced fewer than 10 grade 9s in maths since 2017.

This is not a marginal gain. This is what happened when I stopped leaving memory to chance. Durable learning should not be a miracle. It should be the predictable outcome of a purposeful system.

The Semantics…

In response to “retrieval is learning”, I hear:

“Retrieval isn’t sufficient for learning.”
“Encoding matters too.”
“Learning is more than recall.”
“We need critical thinking, not just memory.”

All true. All irrelevant.

Not because the ideas are wrong, but because the objections miss the educational stakes. Retrieval is not rote recall. Retrieval builds structure, interest, critical thinking, problem-solving, motivation and identity. This is the point I made in my Skin in the Game piece: when you remember something, it becomes part of you, and that personal ownership transforms your curiosity.

In the most technical sense, what I believe is: the quality and quantity of retrieval is proportional to the quality and quantity of long-term change in memory. So, arguing whether retrieval is learning or strengthens learning or is fundamental for learning or whatever else spoils the point and distracts everyone from the simple, urgent reality: we do not retrieve enough. I wouldn’t want Retrievalism, an idea which could have such a profound impact on so many students, to be bogged down in the semantics of my own exaggerated claim.

The Equity Argument

Retrieval is the most equitable practice in education. Like most great thinkers of history, privileged pupils can rely on background knowledge built through family talk, museum trips, books at home, revision schedules, private tutors etc. Disadvantaged pupils almost solely rely on school for their education.

With systematic retrieval built into schools pupils could remember all kinds of knowledge and draw it into their working memory at will. Imagine a quote that has been regularly retrieved and a pupil is now able to pull it into their working memory with very little effort. They can now engage with the great authors of history when they’re brushing their teeth or when they are eating dinner. They can engage with it in different mind states, different contexts and with different people than they otherwise would, expanding their understanding of the quote far beyond what could be achieved in a single lesson.

Without systematic retrieval their experience of education is consigned to the classroom, with it they are free to engage with education any time. Without systematic retrieval, their knowledge leaks away. With retrieval, it compounds. It may not level the playing field, but it will drastically raise the floor. It will give pupils the one thing that opens up everything else: secure, usable knowledge.

Secure knowledge builds interest, motivation, confidence, wellbeing, intellectual identity, problem solving, critical thinking and all the wonderful things we hope for for our pupils. And retrieval builds secure knowledge.

What I Want

I want school leaders to start giving retrieval the same status as encoding.

I want leaders to stop pretending starter activities and spiral curricula are sufficient.

I want teachers to experience the joy of watching pupils grow their memory systematically, not accidentally.

I want disadvantaged pupils to have access to the same durable schemas as their more privileged peers.

I want those active in the debate to focus on what matters.

I want pupils to access the highest forms of understanding by building secure knowledge over time.

I want learning that lasts.

And so, I will keep saying it:

Retrieval is learning.

Not because it is technically complete. But because it is practically true.
And because until the system accepts it, pupils will not learn what we teach.

Learning survives through retrieval.