Most “learn vocabulary fast” guides are listicles. Twenty-one tips, each one a paragraph long, none tested against anything. The research literature actually agrees on a much shorter answer: three techniques, stacked, applied daily. This post is the protocol, not the theory.
Why most vocabulary advice fails
The popular techniques are not the effective ones. In 2013, John Dunlosky and colleagues reviewed ten common study strategies for Psychological Science in the Public Interest and rated each on how well the evidence backed it. Highlighting, rereading, and summarization all came back low utility. Those are the techniques students use most often. The two HIGH-utility techniques across populations and materials were practice testing and distributed practice (Dunlosky 2013, AFT digest).
The problem isn’t that learners are lazy. It’s that the popular techniques feel productive. Re-reading a chapter is fluent and reassuring. The trouble is that fluency at study does not predict recall at test, which Karpicke and Roediger called a metacognitive illusion in their 2008 Science paper: students’ predictions of their own performance were “uncorrelated with actual performance” (Karpicke & Roediger 2008, Science).
Daily volume targets fail for the same reason. “Learn 100 words a day” looks ambitious; in practice, without a forgetting strategy, most of those words are gone in a week.
Technique 1 — Encounter words in context, not lists
A word you meet inside a sentence you mostly understand is easier to acquire than the same word in an isolated list. This is the core of Stephen Krashen’s input hypothesis: adult learners acquire language best when input is “just beyond” current competence — what Krashen calls i+1 (Krashen, Principles and Practice in Second Language Acquisition).
How many encounters does it take for a word to stick? A Cambridge study on incidental L2 vocabulary from reading reports that “after eight exposures, L2 readers recognized the form and the meaning of 86% and 75% of the target nonwords, respectively” (Studies in Second Language Acquisition). Earlier work converges on the same window: 8–10 meaningful encounters for reliable acquisition. A 2025 meta-analysis confirmed similar gains from captioned video (Kurokawa 2025, Language Learning).
Practical sources of input at your level: graded readers, subtitled shows with the target language on-screen, simplified news (NHK Easy for Japanese, News in Slow Spanish), and conversation transcripts. Comprehensible input on its own is not enough; a 2025 critique notes the input hypothesis underspecifies which words get acquired and at what rate (Frontiers in Psychology 2025). It is the right first step, not the whole protocol.
Technique 2 — Force retrieval, don’t re-read
This is the technique most learners skip and the one with the strongest evidence. Karpicke and Roediger taught university students 40 Swahili–English pairs in four conditions. After a one-week delay, the bottom-line finding: “repeated studying after learning had no effect on delayed recall, but repeated testing produced a large positive effect” (Karpicke & Roediger 2008, Science).
In plain English: once you understand a word, re-reading the definition does almost nothing for next week’s recall. Trying to produce the word from a prompt, and sometimes getting it wrong, does almost everything.
The practical drill for vocabulary:
- Cover the L2 word. Read or hear the L1 prompt (or context sentence). Try to produce the L2 word out loud before flipping the card.
- If you blanked, mark it Again. Don’t soften your self-rating to “Hard” just because the answer felt close. The scheduler needs an honest signal.
- Don’t be discouraged by errors. Effortful retrieval is what builds the memory; getting it right too easily means the gap was too short.
Technique 3 — Space your reviews, do not cram
The third leg is when you review. Cepeda and colleagues ran over 1,350 learners across multiple delays in 2008 and mapped what they called a “temporal ridgeline of optimal retention”: the optimal gap between study sessions was “about 20% of the test delay for delays of a few weeks, falling to about 5% when delay was one year” (Cepeda et al. 2008, Psychological Science).
The plain-English version: if you want to remember a word in two weeks, review it in two or three days. If you want it for a year, longer gaps work better. There is no single magic interval. The right gap depends on how long you want the memory to last.
This sits on top of an earlier Cepeda meta-analysis covering “839 assessments of distributed practice in 317 experiments located in 184 articles,” showing that distributed practice reliably beats massed practice, and that the optimal interstudy interval grows with the retention interval (Cepeda et al. 2006).
You cannot apply this manually past about 50 cards. That is where a spaced repetition system (SRS) comes in: it tracks the right next-review date per card so you do not have to. Our explainer on spaced repetition walks through SM-2 and FSRS in plain English.
A daily protocol that stacks all three
The three techniques only work because each one feeds the next. New input gives you words worth keeping. Retrieval turns those words from passive recognition into something you can produce. Spaced reviews keep the produced words alive long enough to show up again in fresh input. It is a loop, not a checklist.
Pick one slot in your day. 30 to 40 minutes is enough.
- 10–15 min — input. Read or listen to something one notch above your current level. Mark the unknown words you actually want to keep.
- 5 min — capture. Type the 5–10 most useful new words into your flashcard tool, with the L1 prompt and an example sentence. The sentence is what makes Technique 1 do its work later.
- 10–15 min — retrieval. Clear today’s review queue. Rate honestly. “Again” today is not a failure; it is the algorithm catching the card before you actually forgot it.
- Weekly — re-encounter. Re-read last week’s input source. Seeing the words you captured show up again in their original context is the cheapest reinforcement available.
That is the whole protocol. Three techniques, one daily slot, one weekly pass.
Try it without the planning overhead
The honest framing: this works on paper. Sebastian Leitner described it in 1972 in his book So lernt man lernen, as a five-compartment cardboard box: cards move forward on correct recall and back to box one on errors (Wikipedia: Leitner system). The University of York’s study guide still recommends the Leitner box as an effective low-cost SRS (York Subject Guides). You do not need an app to start.
What an app removes is the scheduling tax. Maintaining intervals for 300 paper cards by hand is hostile to actually doing the daily protocol. That is why we built Peek — language-only, no sign-up, scheduling handled. You type the words you collected from this morning’s input, rate the recall on each one tonight, and the next review date arrives in your queue when it should.
The speed in “learn vocabulary fast” is not about how many words per day. It is about how many words you still remember in six weeks. The protocol above is the only honest answer.