What to Write on Revision Cards (So They Actually Work)

The question of what to write on revision cards comes up because most students start with the wrong unit. They think in terms of topics ("the water cycle", "World War One causes") and try to fit an entire topic onto one card. That produces a card with a title and a wall of text on the back, which is really just a shrunk-down page of notes.

A revision card works when it holds one retrievable fact, framed as a question. The front asks something specific enough that there is only one correct answer. The back gives that answer and nothing else. If you can't state what single thing the card is testing in one sentence, it isn't ready to write yet.

Before writing a single card, go through your notes and mark only the material you would struggle to recall without a prompt. Skip anything you would recognize instantly if you saw it again. Cards built from material you already know well waste review time you'll want later for the material that actually trips you up.

The front of a revision card should ask a real question, not display a term waiting for a definition. "Mitochondria" on the front and "produces ATP" on the back tests recognition: you see the word, the answer looks familiar, and you mark it correct even if you couldn't have produced that answer unprompted. "What organelle produces ATP through oxidative phosphorylation?" on the front forces you to generate the answer instead of just confirming it.

The back should hold the answer and, where it genuinely helps, one short clarifying detail, not a full paragraph. If your answer runs past two sentences, the card is usually covering more than one fact and should be split into two cards.

Watch for two problems that sneak in even when the front is phrased as a question. First, self-revealing cards, where the wording of the question gives away the answer ("What is the process where plants use sunlight, called photo-something, to make food?"). Second, cards where the back lists multiple unrelated facts, which makes it easy to convince yourself a partial answer counts as correct.

What to write on revision cards shifts depending on the subject, even though the underlying rule (one question, one retrievable answer) stays the same.

For science subjects, cards work best around processes, mechanisms, and cause-effect relationships rather than isolated vocabulary. Instead of "Define photosynthesis", write "What are the two main stages of photosynthesis, and where in the cell does each occur?" For math, put a worked problem type on the front and the method or formula on the back, not just the formula in isolation: "How do you find the derivative of a product of two functions?" tests more than "State the product rule."

For history and other essay-based subjects, cards work better around arguments and evidence than dates alone. A card that asks "What were two long-term causes of World War One, and how did each contribute?" pushes you to reconstruct an argument, not just recall a year. For languages, keep vocabulary cards to one word or phrase per card, and add a second card type for grammar patterns used in a full sentence, since knowing a word in isolation doesn't test whether you can use it correctly.

Across every subject, the same test applies: if the answer side is a single fact you could state in one breath, the card is doing its job. If it reads like a paragraph from a textbook, it's still notes, not a revision card.

The most common mistake is copying sentences directly from a textbook or slide deck onto a card. Copied text tests whether the wording looks familiar, not whether you understand or can reproduce the idea in your own terms. If you can copy a definition without engaging with what it means, the card won't help you when the exam phrases the question differently.

A second mistake is cramming an entire topic onto one card because splitting it feels inefficient. A card with five bullet points on the back isn't testing recall, it's testing whether you can read a list. Each bullet point that stands on its own as a fact deserves its own card.

A third mistake is writing cards only for definitions and skipping application. A deck full of "What is X?" cards will get you through recognizing terms, but exams usually ask you to apply a concept, compare two things, or explain why something happens. Mix in cards that ask "why" and "how", not just "what".

Finally, avoid vague fronts that accept too many correct-sounding answers, like "What is inflation?" A vague front lets you convince yourself a loosely related answer counts as correct, which defeats the purpose of testing recall in the first place.

One. A revision card should test exactly one fact, one relationship, or one step in a process. This is the rule most decks break first, usually because splitting cards feels like it will double the deck size and the review time along with it.

In practice, splitting a bloated card into two focused ones doesn't cost you review time, it saves it. A single overloaded card takes longer to read and grade than two clean ones, and a card with three facts on the back often gets marked "correct" when you only actually recalled one of them. That false sense of mastery is worse than not reviewing the material at all, because it convinces you a topic is solid when it isn't.

A useful check: read the back of the card out loud. If it takes more than one breath or covers more than one idea, split it. A card that asks about the causes of a historical event, for example, works better as separate cards for each cause than as one card listing all of them, especially if the exam might ask about a single cause on its own.

Working out what to write on revision cards takes longer when you're starting from a blank card and a stack of raw notes, a recorded lecture, or a scanned textbook chapter. That conversion step, deciding what's testable and phrasing it as a question, is where most of the time in building a deck actually goes.

Notelyn generates a first-pass deck directly from your source material, whether that's typed or pasted notes, a PDF, an uploaded or recorded audio lecture, or a video and podcast link. The material is transcribed or extracted, summarized, and turned into flashcards you can review and edit, instead of you retyping content into a separate app by hand.

The generated deck is a starting point, not a finished one. Go through it the same way you would a hand-written deck: split any card testing more than one fact, rewrite cards that are too broad or that reveal their own answer, and add application questions the AI is less likely to generate on its own, since it defaults toward definitional questions that map directly onto the source text. For a longer look at building an editing pass into your workflow, see our guide on flashcards for retrieval practice.

Knowing what to write on revision cards comes down to a handful of checks you can run on any card before it goes in your deck. Run through this list, and most of the weak cards that turn a deck into copied notes get caught before they cost you review time.

Before adding a card, check: does the front ask a specific question rather than display a label? Does the back hold exactly one fact, stated in one breath? Is the front free of wording that gives away its own answer? Would answering this require you to recall the fact, not just recognize it? And across the deck as a whole, are there application questions mixed in with the definitional ones?

A deck built around these checks takes a little longer to write than one filled with copied sentences, but it's the difference between a deck that prepares you for the exam and one that only feels like preparation. Whether you build the deck by hand or start from a Notelyn-generated draft, the same short list decides whether the time you spend reviewing actually pays off.