How to make flashcards that actually work

A good flashcard tests exactly one thing, takes less than five seconds to read, and forces you to produce the answer from memory rather than recognize it. That's it. Most flashcards students make violate all three of these rules — they're long, they bundle multiple facts together, and they reward familiarity with your own phrasing rather than genuine understanding. The result is a deck that feels productive to review but doesn't actually transfer to exam performance. The rules below come from decades of cognitive science research on retrieval practice, memory encoding, and spaced repetition system design — primarily from Piotr Wozniak's twenty rules of formulating knowledge, Roediger and Karpicke's work on the testing effect, and Allan Paivio's dual coding theory.

Why making good flashcards is hard

The difficulty isn't technical — anyone can write a question on one side and an answer on the other. The difficulty is deciding what to put on the card. Turning a page of notes into flashcards requires you to identify which facts matter, separate them from the surrounding context, and formulate questions that test genuine understanding rather than surface recognition. This is a compression problem. You're taking dense, interconnected information and breaking it into atomic, independent units that can each be reviewed on their own schedule. Most students skip this step entirely — they copy sentences from their notes, paste them onto cards, and end up with a deck that tests whether they can remember their own handwriting. The act of creating a good flashcard is itself a learning activity. Craik and Lockhart's levels of processing framework (1972) explains why: the deeper you process information during encoding, the better you remember it. Deciding what's important enough to become a card forces deep processing that passive rereading never triggers.

Common mistakes students make with flashcards

Putting too much on one card. A card that asks "Describe the structure and function of mitochondria" is testing at least four separate facts. When you get it wrong, the spaced repetition system can't tell which part you forgot, so it reschedules everything — including the parts you knew. Piotr Wozniak calls this the minimum information principle: each card should contain the smallest possible unit of knowledge. If you need more than one sentence on the back, you probably need more than one card.
Writing cards you don't understand. Wozniak's first rule: do not learn what you do not understand. If you're making flashcards while reading for the first time, you're likely copying things you haven't fully processed. Read the chapter, understand the relationships between concepts, then make cards. Memorizing something you don't understand is like storing a phone number for a person you've never met — it won't stick and it won't be useful.
Testing recognition instead of recall. Flipping a card and thinking "oh yeah, I knew that" is not the same as producing the answer from memory. Roediger and Karpicke (2006) showed that active retrieval — actually generating the answer before checking — strengthens memory traces far more than passive review. If your flashcard routine involves mostly nodding at answers you recognize, you're practicing the illusion of competence, not actual learning.
Making cards from someone else's deck without editing them. Pre-made decks save time, but they skip the encoding benefit of creating cards yourself. If you use them, at minimum review each card and ask yourself: do I understand this? Is this the way I would ask this question? Rewriting a pre-made card in your own words recovers some of the generation effect you'd otherwise lose.
Dropping cards too early. Kornell and Bjork (2008) found that students routinely remove cards from their stack that they haven't actually mastered. Getting a card right once doesn't mean you know it. The spaced repetition system exists precisely to test you again later when you think you've already learned it. Let the algorithm decide when you're done, not your confidence.

How to actually make good flashcards

Start with material you've already read and understood. Never create flashcards on a first pass through new content — you'll make cards about things that turn out to be unimportant, or worse, cards that encode your misunderstandings. One fact per card. Ask yourself: if I get this wrong, is there only one possible reason? If yes, the card is atomic. If no, split it. "What are the four chambers of the heart?" is a bad card — it's testing a list. "Which chamber receives deoxygenated blood from the body?" is a good card — it tests one fact with one unambiguous answer. Keep the front short. The question side should take under five seconds to read. Long, wordy question stems reduce the cognitive benefit of flashcards because you spend your mental effort parsing the question instead of retrieving the answer. If your question is a paragraph, rewrite it. Make the answer specific. A back side that says "it's complicated" or "there are several factors" teaches you nothing. If you can't write a clear, concise answer, you don't understand the concept well enough to make a card about it yet. Go back to the source material. Use images when possible. Paivio's dual coding theory (1971) and the picture superiority effect show that people remember images roughly twice as well as words alone. A card with a labeled diagram of the heart will outperform a text-only card about heart anatomy. This doesn't mean every card needs a picture — but for anything visual or spatial (anatomy, geography, diagrams, processes), images dramatically improve retention. Add context, not clutter. A card that says "Q: What is apoptosis? A: Programmed cell death" is a dictionary entry, not a flashcard. Adding why it matters — "Q: Why is defective apoptosis central to cancer? A: Cancer cells evade programmed death signals, allowing uncontrolled proliferation" — turns it into something worth remembering. The context should clarify, not pad. Watch for interference. Similar cards will confuse each other in your memory. "What does the mitral valve do?" and "What does the tricuspid valve do?" will bleed together unless you add distinguishing context or visual cues. Wozniak identifies interference as probably the single greatest cause of forgetting in experienced spaced repetition users. The fix is making similar cards visually or contextually distinct from each other.

Example study session: 45 minutes of flashcard creation

Minutes 0–15: Read one section of your textbook or lecture notes. Don't make any cards yet. Just read for understanding. Highlight or underline key facts, but only after you've read the full section. If you can't explain the main ideas without looking, reread. Minutes 15–30: Go back through your highlights and create flashcards. For each key fact, write a question that requires recall — not recognition. One fact per card. Keep the front under one sentence. If the material is visual (a diagram, a process, a map), take a photo and use image occlusion or include the image on the card. Aim for 10–20 cards from a typical lecture's worth of material. More than that and you're either testing trivial details or bundling multiple facts. Minutes 30–40: Quiz yourself on the cards you just made. This first review, even though the material is fresh, kickstarts the spaced repetition process. Pay attention to which cards feel confusing or ambiguous — those need to be rewritten, not just reviewed again. If a card has two possible correct answers, the question is bad. Fix it now. Minutes 40–45: Review any cards from previous days that your spaced repetition system has scheduled. These are the older cards that are approaching the edge of your forgetting curve. The effort of recalling them now — when it feels hard — is exactly what strengthens the memory. With Lexie, you can photograph your notes and the AI generates these cards for you, but the review discipline is the same regardless of how the cards were created.

Key facts

Each card should test exactly one retrievable fact (Wozniak, minimum information principle)
Active recall produces 50% better retention than passive review at one week (Karpicke & Blunt, 2011)
People remember roughly twice as many images as words (Paivio, 1971; picture superiority effect)
Students who created their own flashcards scored significantly higher on exams (Senzaki et al., 2017)
67% of college students use flashcards, but most use them ineffectively (Wissman, Rawson & Pyc, 2012)
Interference between similar cards is the leading cause of forgetting in experienced spaced repetition users (Wozniak)
A good target is 10–20 new cards per day; more than that creates an unmanageable review backlog within a week

Frequently asked questions

Making your own cards has a measurable learning advantage. The act of deciding what matters and formulating a good question forces deep processing of the material, which strengthens encoding before you even start reviewing. Senzaki et al. (2017) found that students taught to create their own flashcards scored significantly higher on exams. That said, high-quality pre-made decks (like AnKing for medical students) can save enormous time. The best approach is to use pre-made decks as a starting point, then edit, delete, and add cards based on your own understanding. A deck you've personally curated works better than one you've blindly imported.

Aim for 10–20 well-designed cards per hour of content. More than that usually means your cards are too granular — you're testing trivial details that won't appear on any exam. Fewer than that usually means your cards are too broad — you're bundling multiple facts together. The minimum information principle is the guide: each card should test one atomic fact. If you find yourself writing 50 cards for a single lecture, you're probably including things that aren't worth memorizing. Focus on the concepts that connect to other concepts, the facts your instructor emphasized, and the material you'd need to recall under exam pressure.

Both work, but for different purposes. Basic question-and-answer cards are better for conceptual understanding because they force you to generate the answer entirely from memory. Cloze deletions — where a word or phrase is blanked out of a sentence — are faster to create and work well for vocabulary, definitions, and factual recall. The risk with cloze deletions is pattern matching: you learn to recognize the shape of the sentence rather than understanding the content. For language learning and terminology, cloze cards are excellent. For understanding mechanisms, relationships, and processes, question-and-answer cards produce deeper learning.

Yes, but you need to adapt the format. For problem-solving subjects like math and physics, create cards that test the setup or first step of a problem rather than the full solution. "What method would you use to solve this type of integral?" is a better card than "Solve this integral." For essay-based subjects, create cards that test key arguments, evidence, and connections: "What is the strongest counterargument to utilitarianism?" forces you to recall and organize ideas that you'll later assemble into an essay under exam pressure. The flashcard becomes a building block, not the final product.

Turn your notes into practice questions in seconds

Lexie uses active recall and spaced repetition to help you actually remember what you study. Snap a photo of your notes and get instant practice.

Download for iOS Download for Android