The Real Cost of "Just One More Quiz": Where Assignment Prep Time Actually Goes

Ask a teacher where their grading time goes and most can give you a rough number. Ask where their prep time goes and the answer is usually a shrug, even though it's the same kind of hour: unpaid, unscheduled, and gone before Sunday night. Assignment prep time rarely gets counted the way grading does, and that's worth fixing, because it's not a small line item.

The hour nobody's counting

The grading-time conversation has a number attached to it — the roughly 90 hours a semester teachers report saving once the mechanical part of marking is handled elsewhere, the figure this blog has covered before when breaking down where essay-grading time actually goes. That number is real, and it's useful. But it's also about one side of the job. Nobody's running the equivalent stopwatch on the evening spent building next week's quiz, or the free period that turned into typing out twenty matching pairs by hand.

Prep time doesn't show up in that conversation because it doesn't happen in one visible block the way a grading session does. It's scattered — fifteen minutes here writing a prompt, forty minutes there formatting a gapped-text passage, a whole planning period lost to a matching exercise that needs to hit a specific vocabulary set. Scattered time is harder to notice and easier to just absorb into "that's the job." It still adds up to real hours, and unlike grading, almost none of it involves the judgment calls that actually need a teacher — it's mostly production work: deciding on content, formatting it correctly, checking it against a level.

What each task type actually costs you to build

Not all assignment types cost the same to prep, and it's worth being specific about why, because "building an assignment" isn't one task, it's five different ones wearing the same name.

A writing prompt is the cheapest to produce and the easiest to get wrong in a way that doesn't show until students start submitting. Writing one sentence that reads like a proper prompt — clear task, appropriate register, pitched at the right CEFR level — takes minutes. Writing one that actually produces the range of language you want to assess takes longer, because a vague prompt gets you vague writing back, and you don't find that out until you're grading.

A quiz costs more, mechanically. Every question needs its stem, its distractors if it's multiple-choice, and an answer key that's actually correct — not "probably correct," since a wrong key means regrading later. Fifteen or twenty questions, each checked, is a real chunk of a planning period, and that's before formatting them for whatever tool you're using.

A gapped-text task is its own kind of work: you need a passage first, then you need to decide which words to remove and why, then you need distractors plausible enough to test real understanding rather than just testing whether a student can spot the odd one out. Getting the gap placement right — testing grammar or vocabulary on purpose, not by accident — takes more thought than either of the above.

A matching exercise looks simple and isn't. Building a clean set of pairs (vocabulary to definition, collocation halves, sentence halves, whatever the subtype) means every pair has to be unambiguous on its own, with no pair accidentally matching two different prompts. That's slow, careful work, and it's easy to lose twenty minutes just double-checking that column two doesn't have a second right answer hiding in it.

A transformation drill — word-level or sentence-level — has the highest hidden cost of the group, because the exercise isn't just the prompt, it's the model-answer list behind it. A sentence-transformation task in particular needs a genuinely complete set of acceptable answer variants, since scoring runs against exactly what you've entered. An incomplete answer key here doesn't just mean a mistake to fix later — it means legitimate answers getting marked wrong at scale.

None of these costs are the same, and none of them are the part of teaching that actually needed a teacher's expertise to decide "this is a good task." Most of it is production: typing, formatting, checking.

Where AI generation removes a step — and where it just moves it

This is the part worth being honest about, because it's not "AI generation eliminates prep time," it's narrower than that.

For a writing prompt or quiz, generation genuinely removes a step. Set the topic and CEFR level, generate, and what comes back is a first draft you're reviewing rather than producing from a blank page — the difference between writing and editing, which is a real time saving, not a marginal one.

For gapped-text and matching, generation still saves time, but it moves some of the work rather than deleting it. You're no longer hand-writing the passage or hunting for a matching set, but you are reviewing gap placement, checking that distractors are genuinely plausible, and confirming pairs are unambiguous — the same checks you'd do on your own work, just applied to a draft instead of a blank page. That's faster, but it's not zero-review faster.

For transformation drills, the honest version is closer to "moves the work than removes it." Generation gets you a first pass at the sentences and a starting model-answer list, but because scoring depends on that list being complete, the review step here matters more, not less — you're the one who has to decide whether the generated variants actually cover what a student might reasonably write.

And if what you're starting from is a worksheet you've already built, sitting as a PDF from a previous term, the fastest path usually isn't regenerating it at all. It's importing the file directly, which preserves a properly structured test as-is rather than reinterpreting it — a different kind of time saved than generation, and worth knowing which situation you're actually in before you pick a starting point.

What this adds up to

None of this is a claim that prep time disappears. It's a claim that it's uneven — some assignment types are genuinely fast to build with help, some just get faster to review, and knowing which is which is what actually saves an evening rather than just shuffling the same hour to a different task. Assignments that scale is the shorthand for that idea: levelled prompts, quizzes, and reusable structures generated in seconds so the time that's left goes to the checking that actually needs your judgment, not the typing that doesn't.

---

Start Free — build your next assignment in minutes, not an evening. Start Free — no credit card required.