The research on how to remember what you read points to one habit: close the book and write down, or say out loud, everything you can recall, then check what you missed. Repeat that two days later, a week later and a month later. Recalling from memory and spacing the sessions out are the only two techniques that a 2013 review of ten common study methods rated highly. Rereading and highlighting feel productive and fade fast.
The gap is large. In a 2006 experiment at Washington University, students who studied a short passage once and then practiced recalling it three times remembered 61% of its ideas a week later. Students who spent all four sessions rereading it remembered 40%, even though they had read it about four times as often.
Below are the four studies behind that advice, with their sample sizes and limits, then one book, The Psychology of Money, worked through in four sessions with the exact prompts we'd use.
Why you forget most of a book
Most forgetting happens soon after you learn something, and then it slows down. Hermann Ebbinghaus measured this on himself in 1879 and 1880, learning lists of nonsense syllables and timing how long they took to relearn, and published the results in 1885. In 2015, Jaap Murre and Joeri Dros repeated Ebbinghaus's forgetting-curve experiment as closely as they could and got a similar curve.
The subject was Dros himself, then 22, who spent about 70 hours learning and relearning lists over 75 days. The measure was "savings": how much less time relearning a list took than learning it the first time. Relearning saved 37% of the original time after one hour, 28% after nine hours, 32% after one day, 23% after two days, 17% after six days and 4% after 31 days.
Two details matter for readers. Savings at one day were higher than at nine hours, and in all four datasets the paper compares, Ebbinghaus's included, the one-day or two-day point sits above the smooth curve. The authors note that sleep research would predict that bump, but say it isn't established for this kind of task. The month figure is less settled than it looks: Dros's 4% sits far below Ebbinghaus's 21%, partly because Dros learned more slowly as the weeks went on. Corrected for that, the authors put it near 14%.
The limits are big. This is one person learning syllables built to mean nothing, while a book's ideas come attached to stories and arguments. So these percentages are not the share of a book you'll keep. What carries over is the shape: a steep early drop, then a long, slow tail. That's why the first review should come the same day or the next, not whenever you next pick the book up.
What the research says works
If you want to retain what you read, two techniques have the strongest evidence: pulling material out of memory, and spreading that practice over time. Two popular habits, rereading and highlighting, do much less than they feel like they do.
Recall beats rereading (Roediger & Karpicke, 2006)
Henry Roediger and Jeffrey Karpicke tested recall against rereading with 120 Washington University undergraduates. The material was two short science passages from a TOEFL prep book, one on the sun and one on sea otters, each under 300 words. After studying a passage once, students either studied it again or took a recall test: a sheet headed with the passage's title, and the instruction to write down as much as they could remember. Nobody told them what they got right.
On a final test given right after, the rereaders were slightly ahead, recalling 81% of the passage's ideas against 75%. Two days later the order flipped: 68% for the tested group, 54% for the rereaders. After a week it was 56% against 42%. The tested group remembered as much after a week as the rereaders did after two days.
A second experiment, with 180 students, widened the gap. Students who studied the passage once and took three recall tests remembered 61% a week later. Students who spent all four sessions studying it remembered 40%, even though they had read the whole passage 14 times on average, against 3.4 times for the recall group. The rereaders were also the most confident they'd remember it a week later.
Daniel Kahneman describes the same trap in Thinking, Fast and Slow: what feels familiar and easy to process tends to feel true. Rereading makes a page familiar, and familiar passes for known. Psychologists call the fix retrieval practice, and study apps call it active recall. For books, it means closing the book before you try to remember it.
The study has limits. The participants were undergraduates reading passages of a few hundred words, and the longest delay was a week. A whole book is a bigger job, and the study can't tell you how much of one you'll keep. The authors also note that tests with feedback, where you check your answers, may help more than the feedback-free tests they used.
Spacing beats cramming (Cepeda et al., 2006)
Nicholas Cepeda and four colleagues pooled a century of research on spaced practice: 839 measurements from 317 experiments in 184 published articles. Spreading study sessions apart beat bunching them together almost everywhere they looked. Only 12 of 271 direct comparisons showed no benefit from spacing, or a cost.
The more useful finding is about the gap. The best gap between sessions grew with how long the memory had to last. In studies where the final test came more than a month later, sessions a day apart beat sessions on the same day, and spreading them over weeks or months did better still. For memories needed six months or more later, gaps of at least a month worked best.
Many spaced repetition apps use gaps that expand (a day, then three, then a week). The review found no reliable difference between expanding gaps and equal ones: 62.0% against 58.6% correct, averaged over 22 comparisons in 18 studies. If you use spaced repetition for books, that's good news: the exact days matter less than coming back at all.
Most of the studies used word lists and word pairs, not books, and only 4% of the comparisons tested memory more than a month later. The authors say as much, and call for more research over the weeks, months and years that matter in real learning.
What helps less than it feels like: rereading and highlighting (Dunlosky et al., 2013)
John Dunlosky and four colleagues rated ten study techniques by how well the evidence held up across learners, study conditions, materials and kinds of test. Two received their top rating, high utility: practice testing and distributed practice, the two techniques above. Elaborative interrogation (asking yourself why a stated fact is true), self-explanation and interleaved practice received moderate ratings.
Five were rated low utility: summarization, highlighting, the keyword mnemonic, mental imagery for text and rereading. Low doesn't mean useless. It means the benefits weren't consistent across learners and tasks. The authors chose to review rereading and highlighting because most students report relying on them, and they suggest practice testing in place of rereading.
Two caveats. The review is about students learning course material, not adults reading for work. And "summarization" there means learners writing their own summaries, which is different from reading a summary someone else wrote. We look at that question in whether book summary apps are worth it.
If you highlight anyway, make each highlight do some work. Turn it into a question in the margin or your notes app, then answer the question from memory on Day 2 below. That moves the highlight from a low-rated habit into practice testing. This is our suggestion, built from the review's findings rather than tested in it.
If you want the long version of this research, read Make It Stick: The Science of Successful Learning (2014) by Peter C. Brown, Henry Roediger and Mark McDaniel. Roediger is the same researcher behind the 2006 study above, and the book turns the findings into advice you can use. It's the book we'd hand anyone who wants the full story. Knotd doesn't carry it.
How to remember what you read in a book: one book, four sessions
Here is the method on a real book: The Psychology of Money by Morgan Housel (2020). Its chapters are short, and each one is built around one idea and a story, which makes it easy to test yourself on. The method works on any nonfiction book with an argument, and the self-improvement shelf is a good place to practice.
The schedule follows the studies. You recall once right away, then at two days and a week (the two delays at which Roediger and Karpicke's tested students pulled ahead), then at a month, because the gap should grow with how long you want to keep the book.
| When | What to do | Example prompts |
|---|---|---|
| Day 0, the day you finish | Blank page, book closed. Write all you remember, then open the book and add what you missed in another color. | What is the book's main claim, in one sentence? Which stories do I remember, and what was each one for? |
| Day 2 | Answer five questions from memory, then check each one. | The five questions below. |
| Day 7 | Explain one idea to someone, or use it on a real decision. | Explain "room for error" to a friend using your own budget. Before a big purchase, ask who you hope will notice. |
| Day 30 | A mixed quiz across the whole book. | A friend sells all their funds after a market fall: which idea fits? What does Housel call the rare events that drive most results? |
Day 0: the blank page
This step is adapted from Farnam Street's blank-sheet method. As of October 2026, fs.blog suggests writing what you already know about a subject on a blank sheet before you start, adding new ideas in a different color after each reading session, and reviewing the sheet before the next one.
We add one rule from the research: write from memory first, and open the book only to check. With the book closed, this is practice testing. With it open, it turns into summarizing, which the Dunlosky review rated low.
Day 2: five questions
Answer these on paper or out loud before you look anything up:
- Why does Housel treat Warren Buffett's fortune as a lesson about time more than skill?
- How is staying wealthy different from getting wealthy?
- What is the "man in the car" paradox?
- Why save money with no goal in mind?
- What does Housel mean when he says reasonable beats rational?
Then check each answer against the book. A good answer to the first sounds like this: Buffett started investing as a boy and never stopped, and most of his fortune arrived after his mid-60s, so the years did more of the work than the rate of return. Anything you got wrong or left blank goes on the list for Day 30.
Day 7: explain it or use it
Explaining an idea forces you to rebuild the argument in order, and using it ties the idea to something you'll meet again. For the man in the car, the test is your next large purchase: ask who you hope will notice it, and whether they'd think about you or picture themselves in your place. There's more on this step in how to apply what you read.
Day 30: a mixed quiz
Mix questions from across the book, out of order, and include some that describe a situation instead of naming the idea. You then have to work out which idea fits, which is closer to how you'll meet these ideas at work and at home. The nearest technique in the Dunlosky review is interleaved practice, mixing different kinds of problems in one session, which it rated moderate. Start with the questions you missed on Day 2, then add items like these:
- A friend sells all their index funds after a 20% fall. Which idea would Housel point to? (Volatility is the fee you pay for long-term returns, not a fine for a mistake.)
- Housel argues that a few rare events drive most outcomes in investing. What does he call them? (Tails.)
- True or myth: getting rich and staying rich take the same skills. (Myth. Housel says getting rich takes risk-taking and optimism, and staying rich takes humility and some fear of losing it.)
If you'll use the book for a year or more, add one more session a few months out. Cepeda's review found that gaps of a month or longer worked best for memories needed six months or more later.
How to remember what you read without taking notes
You don't need a notebook to practice recall. You need to produce the idea without looking at it.
Recall out loud on a walk. In Deep Work, Cal Newport suggests "productive meditation": holding one problem in your head while you walk or commute. Use the same time for one chapter. Say its main point and the example behind it, as if someone had asked. Roediger and Karpicke's students wrote their recall rather than spoke it, so this is an adaptation, but it asks your memory for the same thing.
The teach-it test. Explain one idea to a friend or colleague in plain words, without the book. The places where you stall, or reach for the author's phrase because you can't say it your own way, are the parts you haven't learned yet. Go back to those pages and nothing else.
Audio first, recall after each chapter. If you listen, pause at the end of each chapter and say its point in one sentence before the next one starts. With audio there's no page to glance back at, so that pause is your check that you followed. Our piece on audiobooks versus reading looks at what else changes when you listen.
Tools for active recall and spaced repetition
A pen and a calendar are enough for everything above. These tools take over parts of it. The facts below come from each tool's own site, as of October 2026.
Anki is a free, open-source flashcard program for Windows, Mac and Linux, with a free Android app (AnkiDroid) and a paid iPhone app (AnkiMobile), according to its site as of October 2026. You rate how well you recalled each card, and Anki schedules the next review for around when you'd otherwise forget it. It suits readers who are happy to write their own cards: turn your Day 2 questions into cards and let Anki handle the spacing.
Readwise imports your highlights from Kindle, Apple Books, Instapaper and paper books (by scanning) and brings some back each day by email and in its app, using spaced repetition, according to readwise.io, which also offered a 30-day free trial as of October 2026. It suits people who read full books and highlight a lot, because it turns highlights into something you review.
Knotd is our app, so weigh this paragraph accordingly. It has the key ideas from 250+ nonfiction bestsellers, including The Psychology of Money, as narrated audio and text. Each book has a Key Cards deck you swipe through, from the question the book answers to one thing to try, and a nine-question Mixed review with three choices per question, a hint if you're stuck and the reason behind each answer. That covers the Day 30 quiz, though key ideas are not the whole book.
A checklist to keep
- Day 0: close the book, write everything you remember, then check.
- Day 2: answer five questions from memory before you look.
- Day 7: explain one idea to someone, or use it on a decision.
- Day 30: take a mixed quiz across the whole book.
- Recall before you reread, every time, and reread only what you missed.
- Turn each highlight into a question you can answer later.
