Most writing instruction has a feedback problem, and it is not what teachers usually think. The problem is not the quality of the comments. It is not the time it takes to write them. The problem is structural: feedback that does not produce a revision is not feedback. It is a comment. And comments, however precise, do not move student writing forward on their own.
This distinction matters more than it sounds. When a teacher returns a marked-up essay and the student files it away, the instructional loop never closed. The student read the comment, or didn't. They may have agreed with it, or felt defensive about it. But without a revision cycle tied directly to that feedback, there is no observable evidence that learning occurred. The teacher has no way to know whether the comment landed, and the student has no mechanism to act on it.
Why the loop stays open
The loop stays open for a predictable set of reasons. Time is the most obvious one. A teacher with 90 students and a five-paragraph essay assignment might spend three to four hours writing substantive comments. By the time papers are returned, the class has moved on. Asking students to revise now means double the grading on top of new content. Most teachers make a reasonable call: move forward. The revision cycle gets cut.
A second reason is structural. Writing instruction is frequently organized around products rather than process. Students produce a draft, the teacher evaluates it, a grade goes in the book. The grade signals completion. Revision implies the grade was provisional, which creates grade-management complications most schools have not resolved.
A third reason is feedback literacy. Students who have not been explicitly taught how to read and act on feedback will not do so automatically. Telling a student their thesis is unclear does not tell them how to fix it. Telling them their evidence is underdeveloped does not show them what development looks like. Without that bridge, the comment produces awareness, not action.
A comment that produces awareness without action has done only half the job. The other half requires the student to write again.
What closed-loop feedback looks like
Closed-loop feedback has three components that open-loop feedback lacks. First, it is tied to a specific revision opportunity. The student does not just receive the feedback; they are expected to act on it and resubmit. Second, it is verifiable. Someone, or something, can see whether the revision addressed the feedback or not. Third, it has a low enough friction cost that students actually complete the cycle.
This is where most writing instruction falls apart on the third point. If the revision cycle requires waiting for teacher office hours, or submitting to a different platform, or navigating a process that took three days to set up, the friction will kill compliance for the students who need revision most. Those students have already demonstrated they struggle with the work. Adding logistical barriers on top of instructional barriers is compounding the problem.
The feedback loop has to close in a single session or close to it. The student submits. They receive feedback that is specific enough to act on, with at least one concrete example of how to improve a specific sentence or section. They revise. They resubmit. The revision is visible to the teacher without requiring the teacher to re-read every word.
What visibility actually changes
Teacher visibility into revision is underrated as an accountability mechanism. When a student knows their teacher can see whether they revised substantially or made cosmetic changes, the behavioral dynamic shifts. Not because of surveillance, but because the work has an audience beyond the grade. The student's process is now part of the record.
This is distinct from a teacher reading every draft manually. The difference between "the teacher can see I revised" and "the teacher read every word of my revision" is significant for workload. A dashboard that classifies revisions as substantial, moderate, surface-level, or none gives a teacher actionable signal in thirty seconds per student. That signal tells them where to invest their direct instructional time.
The students who received feedback and did not revise are the students who need a teacher's attention. The students who revised substantially are the students who demonstrated the loop closed. Both groups are identifiable without the teacher having to read everything.
The goal of writing instruction is not more feedback. It is feedback that produces revision, and revision that can be seen. If neither is happening, the feedback is producing something else. It may be producing useful information for the grade book. It is not producing better writers.
The structural conditions for closed-loop feedback are achievable. They require feedback that is specific enough to act on, a low-friction mechanism for resubmission, and a way for the teacher to verify the loop closed. None of those three things require more time from teachers. They require different organization of the time teachers already spend.