Feedback a student can use is short, specific, and arrives while they still remember the paper. The shape that works has three parts: one thing that genuinely worked, one habit to change, and one target the next paper can measure. A rank, a percentage and a colour-coded answer sheet are not feedback - they are information about the past. The difference is whether the student knows what to do differently when they sit down to the next paper.
This is written for the person who set the paper. Its companion, a 30-minute routine for analysing mock test results, covers how to read the numbers after a paper closes. This one covers what you say once you have read them. Students who want the other side of it - how to work through their own papers - should read how to take mock tests.
A Rank List Is Not Feedback
A rank answers a question no student asked: how did everyone else do. It moves when other people's scores move, so a student who genuinely improved can still slide down the list and read it as failure. Worse, it collapses a whole paper into one number, which is the one thing a diagnostic must never do. Two students on the same rank can have nothing in common - one attempted everything and guessed badly, the other answered half the paper almost perfectly and ran out of clock. Opposite problems, opposite advice.
MockSetu has already settled this in both directions, and it is worth knowing which way. A creator sees the top three students on an exam and nothing below that line. Every student, meanwhile, sees their own rank printed at the top of their review screen whether you mention it or not. So the choice was never whether to rank - it is whether you let a list of positions stand in for the comment, and you should not.
The Window Where Feedback Still Lands
Feedback stops working when the paper stops being vivid. While a student can still remember why option C looked right on that circuits question, a correction is a conversation and they argue back. Once that memory fades, the same correction is a lecture about a paper nobody can picture, and it gets filed under things the teacher said. Get it out before the memory goes, and before the next paper is built.
There is a practical constraint here. MockSetu sends nothing about a test: no email, no SMS, no WhatsApp notification when a paper closes or a result is ready. The product sends transactional account mail for signup and password reset, and that is the end of it. So the nudge to go and look at the paper has to come from you, in whatever group you already use, and it should carry one instruction: open your attempt review before you read anything I write.
That review screen is worth more than it looks. A student can reopen a submitted attempt question by question and see what they picked, what was correct, the marks the question awarded, and how long they spent on it. Next to their own time, the row carries the average time taken by the students who answered that question correctly, wherever there is data behind it - on a wide screen, at least, because the phone layout drops that line. A student who sees their own time dwarf that benchmark does not need you to tell them they have a pacing problem.
Say It to the Batch First, and Say It Once
Much of what a class needs after a mock is the same for every student in it, which makes the batch note the highest-return thing you write all week. Creator analytics are built for exactly that: they are aggregated and anonymous. You get a score distribution, section-wise accuracy and average time per attempted question, the questions most often left blank, the questions most often marked for review, and a common-misconceptions panel that names the wrong option the largest number of students chose. The only names anywhere on that page are the usernames of the top three.
Write one batch note that does four things, in this order.
- Name the two or three questions the class actually lost, by question number, and say what the error was - not that the question was hard.
- For the worst of them, name the wrong option most of the class chose and explain why it was attractive. A wrong answer that tempted the room is a teaching point; a wrong answer nobody picked is not.
- Name one question the class got right but slowly. Time lost on a question they can already do is the cheapest set of marks in any paper to recover.
- End with the single change you want to see in the next paper, stated so a student can tell afterwards whether they did it.
The Shape of a Comment to One Student
For an individual, three sentences beat three paragraphs. The shape is always the same.
- One thing that worked, specifically. "You finished the first section with time in hand" is feedback. "Good effort" is noise, and students know it.
- One habit to change, stated as a behaviour rather than as a subject. "You stayed on the first hard question too long before moving on" can be acted on next Sunday. "Revise thermodynamics" cannot.
- One target the next paper can measure, using their own last paper as the baseline - fewer blanks in this section than last time, or the first section finished before the halfway mark.
Three parts, and nothing else. The temptation is to add everything else you noticed, and it has to be resisted, because a long list of changes is easy to nod at and impossible to act on. Pick the one habit whose fix unlocks the most marks and let the rest wait for the paper after next.
A score tells a student where they finished. Feedback tells them what to do on Monday. A rank list does the first and quietly pretends it has done the second.
What to Say When the Score Has Fallen
A score that dropped is where feedback most often goes wrong, because the obvious reading - they studied less - is usually not the right one. Ask before diagnosing. A fall can come from a harder paper, from a deliberate change in strategy that will pay off later, from attempting more questions and therefore collecting more negative marks, or from a student sitting the paper at midnight after a full day of school.
Some of those deserve praise rather than correction. A student who has stopped guessing blindly under a negative marking scheme can score lower while the new habit settles and is playing the exam better. A student who attempted a section they used to skip entirely has done exactly what you asked. Your analytics cannot tell you which of these happened, because they are aggregated across the batch - but the student's own review screen can, and so can one question: what did you do differently this time?
Then, if something uncomfortable does need saying, say it plainly and keep it attached to the paper rather than the person. The attempt underperformed; the student is not a lesser human being this week. A falling score with no explanation offered is how a batch loses a student quietly, somewhere between the result and the next class.
Individual Feedback Without Report Cards
Here is the honest boundary, because competitors in this space will not draw it for you. MockSetu has no per-student report cards and no CSV or Excel export of results. There are no certificates, and nothing emails a student a summary of how they did. The creator view is deliberately a view of the batch - counts, averages, distributions, and a top-three leaderboard. Individual written feedback is work you do yourself, and no amount of clicking will generate it.
That is less bleak than it sounds, because the person holding the data you would have exported is the student. Make them bring it to you.
- Ask every student to send three things from their own review: their score, the question numbers they marked for review, and the one question they got wrong that they were certain they had right.
- That last item is the valuable one. It names a misconception the student can actually feel, and it is the one thing on the list no analytics page could have told you.
- Reply to the batch about patterns and to individuals about habits. Pattern feedback scales; habit feedback does not, so spend it where it will change something.
- Rotate the written comments. A handful of students each week, tracked in a sheet you keep yourself, means everybody gets a real comment on a schedule and nobody is quietly skipped for a term.
- Keep one line per student per paper. Creator analytics never give you a per-student history, so over a term that line is the only running record of an individual you hold - and it turns "you are inconsistent" into "this is the third paper running where you lost the last section to the clock".
One more thing to be straight with your batch about: published papers are public, and there is no proctoring of any kind - no webcam, no lockdown browser, no tab-switch detection. A score is a number the student produced under their own supervision. Say so in the feedback and frame the paper as a diagnostic they are running on themselves; the honest ones stay honest once they can see the point of it.
Make the Next Paper Carry the Feedback
The real test of feedback is the next paper, not the conversation about the last one. If you told the batch they bleed marks by lingering on the first hard question, the next paper should contain a section where that habit is punished and the fix shows up in the timing data. Build the paper so it checks whether the feedback landed.
Duplicating the previous paper is the cheap way to do that. A duplicate carries the marking scheme, the timing groups and the language links across, so the only thing you change is the questions - though the copy is best-effort, so open it and check the settings before you publish. A weekly test series is the format where this loop compounds, and how to conduct an online exam for students covers the mechanics of getting the paper in front of a batch without a bad test day.
To see the paper the way your students saw it, preview it from the editor and sit a few questions. A creator preview records nothing - no attempt, no responses, no analytics - so your own run never contaminates the batch numbers you are about to comment on. Everything a creator can build here is free and takes no card, and the analytics behind every feedback note come with each paper you publish.