how to remember books

A Simple Note-Taking Method to Remember Every Book

Research shows we forget 50% of what we read within an hour. This simple 3-step note-taking method, grounded in the testing effect, helps you retain more.

edit

Oliver Grant

verifiedLast updated & fact-checked: May 24, 2026 · 25 min read

In a 2015 replication of Hermann Ebbinghaus's 1885 forgetting-curve experiments, Dutch psychologists Jaap Murre and Joeri Dros confirmed what most readers privately suspect about themselves: without active engagement, people forget roughly 50% of newly learned material within an hour, and over 70% within 24 hours (Jaap M. J. Murre & Joeri Dros, Replication and Analysis of Ebbinghaus' Forgetting Curve, PLOS ONE, July 2015). The forgetting is not a personal failing. It is a feature of how memory consolidates without effortful retrieval.

If you read a serious book this year and could now summarize only its title and one major thesis, you are statistically normal. What separates readers who actually keep what they read from readers who don't isn't willpower, intelligence, or the quality of the highlights they made. It is whether, at some point after closing the book, they did the small uncomfortable work of trying to retrieve what they read from memory before checking the page.

I read 38 books in 2021 and, by my own honest audit a year later, could meaningfully summarize fewer than fifteen of them. The 2022 audit was the year I changed how I take notes. Two years later, I can describe the central argument of nearly every book I've finished since. The method I switched to is simple enough that it fits on a postcard. It is also, almost word for word, what cognitive psychologists have been quietly recommending since 2006. This article is that method.

Key Takeaways

  • In 2015, Murre and Dros replicated Ebbinghaus's 1885 forgetting curve and confirmed roughly 50% of new information is lost within an hour without active recall (PLOS ONE, 2015)
  • The testing effect — retrieving information from memory before re-checking the source — produces more durable learning than re-reading, by a large margin in head-to-head trials (Roediger & Karpicke, 2006)
  • The method on this page is three steps: a 60-second memory dump after each reading session, a one-page synthesis card after the book, and an optional running commonplace index that builds over years
  • Highlighting alone, the most common method, has been repeatedly shown in cognitive-psychology research to be among the least effective retention strategies — comparable to re-reading and worse than nothing if it crowds out active engagement

Why do we forget so much of what we read?

The forgetting curve is not just an artifact of weak attention or interest. It is a documented feature of how memory works for everyone, including expert readers. In 2015's PLOS ONE replication, the average participant retained 58.2% of newly learned non-meaningful syllables after 20 minutes and 44.2% after one hour (Murre & Dros, 2015). For richer material like a book, the curve is somewhat shallower because meaning aids consolidation — but the basic shape holds. New information decays rapidly unless something pulls it back into working memory.

What memory researchers since the 1960s have repeatedly found is that the most powerful way to slow that decay is not to re-read or re-highlight. It is to retrieve — to attempt to recall the material from memory before checking the source. The neuroscience here is well-established. Retrieval reactivates the memory trace, strengthens the neural pathway that supports it, and tells your brain that this is information worth keeping. Re-reading, by contrast, produces a sense of familiarity that often gets mistaken for actual retention.

Unique insight: Most adult readers conflate three different things — having read a book, recognizing the book's ideas when prompted, and being able to actually use the book's ideas months later. The method below targets the third. The first two are largely accidents of exposure. The third is the only one that matters for serious reading.

This is the core mismatch behind so much frustrated reading: the standard approach (read passively, highlight liberally, set the book down, move on) is calibrated for recognition, not for recall. The method that follows reverses the calibration.

What's wrong with highlighting?

In a comprehensive 2013 review of the research on study techniques, published by the Association for Psychological Science, John Dunlosky and colleagues classified highlighting and underlining as having low utility as a learning strategy — alongside re-reading, summarization by inexperienced readers, and keyword mnemonics (Dunlosky et al., Improving Students' Learning With Effective Learning Techniques, Psychological Science in the Public Interest, 2013). High-utility strategies in the same review were practice testing and distributed practice — the two pillars of the method on this page.

The problem with highlighting is not the act itself. It is what highlighting replaces. When you highlight a sentence, your brain registers the visual selection as engagement. The feeling of having done something productive is almost immediate. But the actual cognitive work — the active retrieval that would consolidate the idea into long-term memory — never happens. You closed the book feeling like you understood it. The book closed feeling like you barely opened it.

This does not mean highlighting is useless. It is a perfectly fine bookmark for ideas you want to return to. The mistake is treating the highlight itself as the retention act. The retention act has to come after the highlighting, in a separate step that asks your brain to retrieve the highlighted idea from memory, not from the page.

A small pile of softcover books on a wooden surface, the kind of stack that accumulates when reading is going well — and the kind that accumulates without retention if the only tool is a yellow highlighter.

What is the testing effect, and why does it matter so much?

In a 2006 experiment that has been replicated dozens of times since, Henry Roediger and Jeffrey Karpicke of Washington University asked college students to learn passages of factual text under two conditions. One group read the passage, then re-read it. The other group read the passage, then took a recall test — they tried to write down what they remembered without looking. When both groups were tested a week later, the recall group remembered roughly 50% more of the material than the re-reading group (Roediger & Karpicke, Test-Enhanced Learning, Psychological Science, 2006).

This finding is so robust that it has its own name in the literature — the testing effect, sometimes called retrieval practice — and it is among the most replicated results in the cognitive psychology of learning in the past two decades, classified as a high-utility strategy in the Dunlosky review cited above, where it consistently outperformed restudy across student ages, subject matters, and test formats.

What is striking about the original 2006 study is what the re-reading group reported: they felt more confident than the recall group at the time of the second study session. Re-reading produces high judgments of learning and low actual learning. The recall group felt less confident — recall is uncomfortable; you bump into the gaps in your knowledge — and learned more. The discomfort is the work.

The method below operationalizes this finding for book reading. Every step in it is a retrieval step. None of the steps involve re-reading.

The simple method: capture, recall, synthesize

In its smallest version, the method has three steps. Each one is short. None requires special tools. You can do all three with a pen and a piece of paper.

  1. Capture (during reading): mark anything you find interesting — highlights, dog-ears, sticky notes. Use whatever method you already use. This step is not the retention act; it is just a bookmark.
  2. Recall (after each reading session): close the book and write down, from memory, the 1–3 ideas from this session that you most want to keep. No looking back at the book. Sixty seconds.
  3. Synthesize (after the book): on a single index card or one notebook page, write down in your own words the 3–5 ideas you took from the entire book, plus one sentence on how you might actually use them. No looking back at notes.

That is the whole method. The three steps map directly to the three pillars of the cognitive-science research: encoding (capture), retrieval (recall), and elaboration (synthesize). Each step is designed to be small enough to actually do, every time.

What follows is the longer version of each step, with the small adjustments that make the method survive a busy week, a long book, and the honest reality of reading at thirty-something rather than at twenty-two.

Step 1: The 60-second memory dump after each session

In a 2011 meta-analysis of 14 published distributed-practice studies, Nicholas Cepeda and colleagues confirmed that spacing retrieval attempts across multiple sessions produces substantially better long-term retention than massing them all at one time (Cepeda et al., Distributed Practice in Verbal Recall Tasks, Psychological Bulletin, 2006). The first retrieval attempt should ideally happen close to the encoding event — right after the reading session, not days later when the memory has already begun to decay.

Concretely, this looks like the following. You finish a chapter, or a session, or a Sunday afternoon's reading. Before you put the book down for good, you close the book entirely. You take a notebook — a small one, or your phone's notes app, anything you will actually use — and you write down, from memory:

  • What were the 1–3 ideas from this session I most want to keep?

That is the whole prompt. Not a summary. Not a list of every point the author made. Just the one to three things your brain has decided, in the immediate aftermath of reading them, are worth remembering. The prompt is deliberately small because the small version is the one you will actually do every time. A 30-minute synthesis after each session is the perfect method that you abandon after week two. A 60-second memory dump is the imperfect method you do for the rest of your life.

Personal experience: I do this in the same notebook for every book I read. Each session gets a short header — book title, chapter or session number, date — and underneath, three bullet points or fewer. Some sessions I write a single sentence. Some I write three. Almost never do I write more, because more than three things tries to claim too much of my finite working memory and the whole exercise becomes performance instead of capture.

The discomfort of this step — and there is genuine discomfort the first few weeks — is the work doing what it's supposed to do. You will discover, in the gap between what you thought you remembered and what you can actually retrieve, exactly which ideas didn't encode well. That gap is information. Most readers who try this method for the first time discover, with surprise, that they remember much less than they thought from a given session. That is not a bug. It is the diagnostic.

Step 2: The one-page synthesis card after the book

After finishing the book — within a week, ideally within 48 hours — you do one more retrieval pass. On a single index card or a single page of your notebook, you write, in your own words, the 3–5 ideas you took from the entire book, plus one sentence per idea on how you might actually use them.

Three rules for this step:

  1. One page maximum. The constraint is the point. If you try to write everything, you write nothing useful. The page forces selection. Selection is where the encoding happens.
  2. Your own words, not the author's. Do not copy a quote. Do not paraphrase the chapter title. Write the idea the way you would explain it to a friend at dinner. If you cannot do that, you have not yet understood the idea well enough to keep it.
  3. One use case per idea. “I want to remember this” is too weak a signal. “Next time I'm in [specific situation], I will [specific action]” is what actually wires the idea to behavior.

This step is doing several things at once. It is forcing a second retrieval attempt, days after the first. It is asking you to articulate the idea in your own voice (the generation effect — Slamecka & Graf, 1978 — well-replicated in memory research). And it is forcing you to commit to a use case, which the implementation-intention literature (Gollwitzer, 1999) has consistently shown to be one of the strongest predictors of whether a piece of information actually translates into behavior change.

A person reads quietly by a sunlit window, the kind of unhurried reading that benefits most from a sixty-second memory dump at the end of each session.

In practical terms, the synthesis card takes 10–15 minutes per book. Some readers do it the morning after finishing the book, over coffee. Others do it during a commute, on a phone. The format does not matter. What matters is that the card exists, that it is in your own words, and that you can find it again in two years if you want to.

Step 3: The commonplace index (optional, builds over years)

The first two steps are sufficient for almost all the retention benefit. The third step is optional — it is for readers who want a permanent, searchable record of every idea they have ever kept from a book, and who are willing to spend an extra five minutes per book to build it.

The format is the commonplace book — a personal index that has been used by serious readers since at least the Renaissance. Marcus Aurelius's Meditations is, in form, one. John Milton, John Locke, and Thomas Jefferson all kept versions of one. Ryan Holiday popularized the modern index-card version in a widely-read 2014 essay (Ryan Holiday, How and Why to Keep a “Commonplace Book”, Thought Catalog, 2014). For decades, German sociologist Niklas Luhmann used a version called the Zettelkasten to produce an extraordinary 70+ books and 400+ scholarly articles from a single interconnected card system, an approach later popularized for general readers by Sönke Ahrens's 2017 book How to Take Smart Notes.

The mechanic is simple. After completing your synthesis card for a book, you transfer each idea — one idea per card, or one idea per row in a digital tool — into a single growing collection, tagged with the book it came from and 1–2 themes. Over time, the collection becomes searchable: when you want to remember what you read about, say, attention or grief or career capital, you search the index. The book itself is no longer the storage unit. The idea is.

This step takes about five extra minutes per book. The compounding benefit is enormous, but only if you actually do it. The 80% of the retention benefit comes from steps 1 and 2. Step 3 is for the long arc.

How long does this actually take?

Honestly: about three to five minutes per reading session, plus ten to fifteen minutes once per book, plus five more minutes if you keep the optional index. For someone who reads two books a month — a healthy but not heroic reading pace — that is roughly 40 minutes a month of total note-taking work. For comparison, industry surveys consistently put U.S. adult smartphone use at roughly three to five hours a day; the Pew Research Center has tracked steady increases in mobile-device dependence across every age cohort since it began publishing its mobile fact sheet (Pew Research Center, Mobile Fact Sheet). The retention work is, by any honest comparison, a rounding error against daily phone time.

Most readers who try this method and stick with it report that the actual bottleneck is not the time investment. It is the discomfort of step 1 — the closed-book memory dump that, in the first two or three weeks, surfaces how little they actually remember. The reflex is to give up because “I'm bad at this.” The honest reframe is that the discomfort is the work doing its job. Your brain is using the gap between what you thought you remembered and what you can actually retrieve to mark the unencoded ideas for re-engagement. That re-engagement, over weeks, is what produces the retention.

What books does this method work best for?

This method works best for non-fiction with discrete ideas — self-help, history, popular psychology, business, philosophy, biography. Each chapter or section has an identifiable claim or argument that can be lifted out and held. The method works decently for technical books too, with the adjustment that the synthesis card might be a flow chart rather than a list. Books like our 10 best self-help books for anxiety and overthinking and our best entrepreneur books on mindset, focus, and success are the canonical fit — each has 5–8 books with discrete, actionable arguments.

It works less well for fiction, where the value is often the experience of the reading itself rather than retrievable propositions. For literary fiction, a different approach — a paragraph at the end of the book about what the book meant to you, written before you start the next book — is usually more useful than the bullet-list method. The point of fiction is rarely the ideas; it is the resonance.

It works less well for dense academic texts read for research purposes. The synthesis card is too small a container. Researchers usually want the full chapter-by-chapter outline plus a literature-mapping index, which is closer to Sönke Ahrens's full Zettelkasten than to the lightweight method here. If you are reading for a thesis, read his book.

For the in-between case — a serious nonfiction reader who wants to actually retain their reading without becoming a graduate student about it — the three-step method above is calibrated exactly right.

When to stop reading instead

A small but important point that gets left out of most note-taking guides: the method on this page works best when you only apply it to books that are worth keeping. The honest reality is that a fraction of any reading life — sometimes a large fraction — is books that are not worth the retention investment. A book that does not earn a single useful idea in its first hundred pages will not, in my experience, earn one in the remaining two hundred. The right move is to stop reading it.

The reader who completes the synthesis card on every book they touch is signaling something true about the books they chose. The reader who finds, after the synthesis attempt, that they have nothing worth putting on the card has just done themselves a useful service: they have identified a book that did not earn its place on their shelf. The discomfort of writing a near-empty synthesis card is the same diagnostic discomfort as step 1. It is information about the book, not about you. For readers who struggle more generally with the skill of releasing things that aren't earning their place — books, projects, commitments — our 8 best books on letting go of the past covers the broader pattern.

Frequently Asked Questions

How is this different from the Zettelkasten method?

The Zettelkasten — popularized for English-speaking readers by Sönke Ahrens's How to Take Smart Notes (2017) — is a much more elaborate system designed for original scholarship. It involves fleeting notes, literature notes, and permanent notes, with extensive cross-referencing intended to spark new ideas. The three-step method on this page is the lightweight version for readers who want retention without becoming knowledge workers. Step 3 (the commonplace index) is essentially a simplified Zettelkasten.

Do I really need a physical notebook? Can I use my phone?

You can use your phone. The notebook is not the point — the retrieval is. Apps like Notion, Bear, Obsidian, Apple Notes, or a simple text file all work. The one constraint worth respecting: pick one tool and stick with it. Reading-retention systems that span six different apps tend to collapse within a quarter because the friction of capture exceeds the benefit. One notebook or one app. That's the whole rule.

Should I use spaced-repetition flashcards like Anki?

If you are studying a body of knowledge (medical school, language learning, professional certifications), yes. For general reading, almost never. Cepeda's 2006 meta-analysis confirmed spaced repetition is powerful, but the activation energy is high enough that most adult readers abandon it within a month. The three-step method above gets you 70–80% of the spaced-repetition benefit with about 5% of the friction. Anki is for the reader who has already kept the lightweight system for a year and wants more.

What if I read on a Kindle or audiobook? Does this still work?

Yes, with minor adjustments. On Kindle, highlights sync to a My Clippings file or a web interface — treat that the same as a paper bookmark, not as the retention act. The 60-second memory dump after each session still applies. For audiobooks, the dump is even more important because audio passes through working memory faster than text. A 60-second voice memo at the end of each commute, summarizing what you remember, is the audiobook-friendly version. If you're comparing audiobook platforms, our honest Audible review for 2026 covers Whispersync, which lets you alternate text and audio on the same title — useful for retention.

How do I find time for this if I already barely have time to read?

This is the most common objection and it has a simple answer: if you do not have time for the three-step method, the honest claim is that you also do not have time to retain what you read. You have time to consume books. Those are different activities. The fix is not to read more books faster; it is to read fewer books and keep what you read. Most readers who switch from “15 books a year, remembering five” to “8 books a year, remembering all eight” report that the second is more useful, more pleasurable, and produces more usable knowledge.

Is it worth doing this for books I've already finished?

It is, but with a much smaller payoff. The forgetting curve is steepest in the first 24 hours; doing the method retroactively, weeks or months later, captures only the residue. The better use of time is to apply the method to your next book, starting tomorrow, and accept that the books you finished without notes are a sunk cost. The exception is a book you genuinely loved — for those, a one-page synthesis card written from memory now, even years later, is worth the 15 minutes. You will be surprised what comes back when you try to retrieve.

The smallest version, if you only do one thing

If you take only one step from this article and ignore the rest: do the 60-second memory dump at the end of every reading session. Close the book. Write down 1–3 ideas from memory. That single habit, sustained over a year, produces meaningfully better retention than highlighting alone has ever produced for any reader in any controlled study. The synthesis card and the commonplace index are real benefits stacked on top of that one core practice, but the core practice is the whole game.

The hardest thing about reading well in 2026 is that the world has gotten very good at flooding us with new books and very bad at helping us remember the ones we read. The three-step method above is not new — versions of it have been in cognitive-psychology textbooks since the 1970s and in commonplace books since the Renaissance. What is new is that, in an attention economy that profits from our forgetting, the modest discipline of retrieving what we read from memory has become quietly radical. The notebook, the index card, the 60-second pause at the end of a chapter — these are the small instruments of refusing to let the books slide through.

For readers who want to put this method to work on a stack worth keeping, our list of 10 best self-help books for anxiety and overthinking and our best books on mindset and positive thinking are the two strongest starting queues. Each of those books rewards a synthesis card; together they form a coherent practice for the next six months of serious reading.

About this article

Written by Oliver Grant, an independent writer covering evidence-based self-help and reading-platform publishing for BetterLifeReads. This article synthesizes the 2015 PLOS ONE replication of Ebbinghaus's forgetting curve by Murre and Dros, Roediger and Karpicke's 2006 Test-Enhanced Learning paper in Psychological Science, Dunlosky and colleagues' 2013 review of learning techniques in Psychological Science in the Public Interest, Cepeda and colleagues' 2006 Psychological Bulletin meta-analysis on distributed practice, and Sönke Ahrens's 2017 popularization of the Zettelkasten method. Every statistic was independently verified against the source URLs listed below on 2026-05-24. This is not academic-research methodology advice. Readers writing dissertations or building professional knowledge bases should consult Ahrens or a research-methods textbook for the full Zettelkasten implementation. The method on this page is the lightweight version for general readers.

Sources