How Many New Words Should You Learn Per Day?

The Short Answer

For most learners, 5 to 15 new words a day is the sustainable range. The popular default of 20 sits just above it, and it quietly assumes you have half an hour for reviews every single day. Any flat number is only half an answer, though. The reliable way to choose is backwards: decide how many minutes of review you can do daily, including your bad days, and derive the new-word number from that.

The direction matters because every new word is a small loan. A spaced repetition system will show it to you again tomorrow, then in a few days, then in a couple of weeks, and those payments stack on top of the payments from everything you added before. Add 10 new words a day and within two months you owe roughly 50 reviews a day. The sections below show that math, turn it into a formula for your own number, and cover the release valve that keeps the whole system stable.

Why a Flat Number Misleads You

Search this question and you will find confident answers of 7, 10, 12, and 20 words a day. They disagree because the question is incomplete. The right number depends on your future review workload, and the future is exactly what a day-one decision cannot see.

The first two weeks always feel easy. On day one you learn 10 words and review nothing. A week in, you are reviewing 20 or so words a day and it feels like nothing. But batches from different days keep overlapping, and the pile keeps growing for months. This is why people who start at 40 or 60 new words a day often feel great in week one and quit in week four. They did not lose discipline. They signed up for a workload nobody showed them. Anki's own manual warns about exactly this pattern: users who study hundreds of new cards in their first excited days, then get buried by the reviews that follow.

None of this is an argument against spaced repetition. Spacing reviews out beats cramming in one of the most consistently replicated findings in memory research, confirmed across hundreds of experiments in Cepeda and colleagues' 2006 meta-analysis, and actively recalling material beats re-reading it when memory is tested days later, as Roediger and Karpicke showed in 2006. The forgetting curve the whole system fights is real too: researchers replicated Ebbinghaus' original experiment in 2015, 130 years later, and found the same steep early decline. The system works. It just has arithmetic, and the arithmetic deserves to be planned for.

The Math: What 10 New Words a Day Actually Costs

To make review debt concrete, here is the review ladder LingoUnify uses. Each correct answer moves a card up one stage: 1 day, 3 days, 7 days, 14 days, 30 days, 60 days, 120 days. A word you always answer correctly gets reviewed on days 1, 4, 11, 25, 55, 115, and 235 after you add it. Seven reviews in eight months, then once every four months until you retire it. Most flashcard systems follow a similar shape.

Now add 10 new words every single day. On day 4, yesterday's words come back for their 1 day review while your first batch returns for its 3 day review: two batches land on the same day. That overlap is the whole story, and it keeps deepening:

And that table is the floor, because it assumes you answer everything correctly. Every miss demotes the word a step and shows it again in 10 minutes, so real numbers run higher. Anki's documentation makes the same point with its own scheduler: a steady 20 new cards a day works out to roughly 200 reviews a day once the system is warmed up.

The takeaway is a rule of thumb: whatever new-word number you pick, your eventual daily reviews will be roughly 5 to 10 times that number. Ten a day is a 15-minute daily habit a few months from now. Forty a day is an hour or more, every day, indefinitely.

Pick Your Review Capacity First, Then Derive the Number

Turn the multiplier around and the planning becomes simple:

1. Count your honest minutes. Not your motivated-week minutes. The number of review minutes you can do on a bad Tuesday.
2. Convert minutes to reviews. Most people get through 6 to 8 simple word cards a minute. Ten minutes is roughly 70 reviews.
3. Divide by 8 to 10. That plans for the heavier end of the multiplier range, which leaves margin for misses and bad weeks.

In practice:

If those numbers look small, run the yearly math. Eight new words a day is nearly 3,000 words a year. For scale, Nation and Waring's classic review estimates that native speakers, who live inside the language all day, add only about 1,000 word families a year, and that 3,000 to 5,000 word families provide the basis for comprehension of ordinary texts. Word families are counted differently from flashcards, one family covers several forms, but the point survives the unit change: a modest daily number is close to the fastest pace anyone sustains for years.

Which words should fill those slots is its own question, and we covered it separately in What Words Should You Actually Save When Learning a Language?

The Release Valve: Words Have to Graduate

The piece that makes the math work is graduation: review debt only stabilizes if words can leave the system. If every card repeats forever, your daily pile creeps upward forever, no matter how disciplined you are. Even SuperMemo, the system that invented modern spaced repetition, ships overload tools for postponing and deprioritizing material because its designers know demand can outgrow any human's day.

So retiring words you genuinely know is not cheating. It is how the math is supposed to work. If you can read a word in a fresh sentence and understand it without thinking, reviewing it again in four months buys you almost nothing, and it costs a slot your newer, weaker words need.

LingoUnify builds this distinction in, which is worth explaining since this article uses its numbers. Saving a word while reading gives it the Learning status: it gets highlighted in texts, keeps its translation a tap away, and costs you zero reviews, because saving a word does not create a flashcard. Review debt starts only when words go into a deck. And when a word is truly yours, one tap on Learned retires its card permanently and switches the highlight to green, though you can always bring it back if you were too optimistic. The honest limits: LingoUnify has no automatic per-day cap on new cards, so pacing stays your job, and its intervals stop growing at 120 days rather than stretching to years the way Anki's scheduler can. If you prefer Anki's algorithm, you can export your words as Anki-ready text files and keep the same collection there.

Already Buried? How to Get Out

A backlog of several hundred reviews is a scheduling problem, not a moral failure. The way out:

• Stop adding new words immediately. New cards are the only input to the debt. Pause them and the pile stops growing within days.
• Chip, do not binge. Twenty minutes a day on the backlog beats one heroic three-hour Sunday that burns you out again.
• Retire aggressively while you clear. A backlog is full of words you actually know. Graduating them is the fastest legitimate way to shrink the pile.
• Do not delete the deck. The cards you struggled with are a map of what needs work. Deleting everything throws away the map along with the debt.

When the pile is clear, restart new words at whatever number your capacity math gives you. It is probably lower than the one that buried you.

Mistakes That Create Review Debt

• Choosing a number on your most motivated day. Day-one energy is the worst possible input to a decision about month three.
• Making a card for every unknown word. Some words only deserve a quick lookup. Rare words you met once in one novel usually earn nothing.
• Never retiring anything. A deck where nothing graduates is a treadmill that only speeds up.
• Treating a missed day as failure. Spaced repetition is forgiving. The cards wait, tomorrow is slightly heavier, and the schedule self-corrects. One missed day costs almost nothing.
• Letting reviews replace reading. Flashcards maintain words. Meeting those words in real sentences is what makes them usable, so protect your input time even on heavy review days.

What to Do Next

1. Time yourself tonight. Review for five minutes, count the cards, and you have your personal pace per minute.
2. Run the formula. Honest daily minutes, times pace, divided by 8 to 10. That is your new-word number. Write it down.
3. Audit your current decks once. Retire everything you can read without thinking. Most decks carry more dead weight than their owners expect.
4. Set a review target you can see. In LingoUnify you can set your daily goal to card reviews, anywhere from 5 to 500 a day. A visible target keeps the debt visible too.
5. Re-derive in two months. Your pace improves and your capacity changes. Treat the number as a setting you revisit, and expect it to change.