Marking time
Total department hours logged against marking for one term, against the same term the previous year with the same number of papers. Self-reported by the six teachers on their own timesheets.
A science department, Standards 6–10
The department that got its Sundays back
Six teachers, forty scripts each, every fortnight. The marking was not the job — it was the thing standing between them and the job.
−83% marking time
28 min per paper, reviewed
96% agreement with teachers
8 min to a ready paper
Four days of marking to twenty-eight minutes
What they were doing,
What they were doing,
and what replaced it.
6 teachers · Standards 6–10 · fortnightly papers. Measured across one full term, against the same department’s prior year.
Before
A fortnightly paper meant a lost Sunday, and marking drifted depending on which teacher picked up which pile.
After
Eight minutes to set, twenty-eight to review forty scripts, and a written rubric that made two teachers mark the same script the same way.
“I did not mind the marking. I minded that the marking was the reason I never got to the part where I work out what to teach next.”
Head of science
How it actually went
Five moments,
Five moments,
in the order they happened.
Not a plan — what they did, including the parts that were not the plan. Pick a moment.
Wk 1
Write the rubric down
Before anything else. Six teachers, one afternoon, deciding in writing what earns the second mark on a three-mark question. Nobody in the department had ever written it down, and the first attempt disagreed with itself.
- One afternoon, six teachers
- First draft disagreed with itself
- Second draft is what shipped
Wk 2
First paper, marked both ways
Forty scripts marked by hand as usual, and marked again by the platform, without either seeing the other. The comparison is what convinced the two sceptics, in both directions.
- Forty scripts, marked twice
- Blind to each other
- Disagreements read individually
Wk 4
Reviewing, not marking
The change in the job. Scripts arrive marked against the rubric with the phrase that earned each point, sorted so the least confident ones are first. Twenty-eight minutes for forty.
- Low-confidence scripts first
- Every point traced to a phrase
- Twenty-eight minutes, not five hours
Wk 8
Setting stopped being a Sunday
Eight minutes from scoping chapters to a paper the teacher was willing to publish. The department stopped reusing last year’s papers, which had been the real cost of setting being slow.
- Eight minutes to a ready paper
- Last year’s papers retired
- Fortnightly became genuinely fortnightly
End
The Sunday
Marking time down 83% across the term. The department’s own framing: the hours did not go into more marking, they went into the reteaching conversation that marking had always crowded out.
- 83% less marking time
- 96% agreement with the marker
- The hours went somewhere useful
How each figure was measured
Four numbers,
Four numbers,
and where each came from.
Measured across one full term, against the same department’s prior year. A figure without a method is a figure you cannot check, so each one says what it counted and what it did not.
Per paper, reviewed
Median time for one teacher to review a marked set of forty scripts, including every override. The slowest was 51 minutes; the fastest 19.
Agreement with teachers
The proportion of AI-proposed scores the reviewing teacher left unchanged, across roughly 4,800 marked questions. Objective sections are excluded, since agreement there is trivially exact.
To a ready paper
Median from opening the generator to publishing, including edits. Excludes the term where the rubric was being written, which is a one-off.
What it cost
What got worse
What got worse
before it got better.
Every rollout has one, and a case study without it is marketing. These are theirs, in their own account.
The first fortnight was slower, not fasterThe rubric had to be written down
An afternoon of six teachers arguing about what earns a second mark, then a second draft because the first one contradicted itself. That is genuine work and it happened before a single script was marked. It is also the thing that made the marking consistent between teachers, which the department had been quietly failing at for years.
Two of six were not convinced at firstAnd the comparison is what settled it
They asked for forty scripts to be marked both ways, blind, before agreeing to anything. Roughly one script in twenty-five came back materially different, and reading those individually is what changed their minds — in both directions. Two of the disputed marks were the platform’s, and three were the teacher’s.
It does not work for extended writingAnd the department stopped trying
One teacher in the department marks extended argument, where the marking is the teaching. After a term she stopped using it for that and kept it for everything else. We would rather report that than average it away — the 96% figure excludes her subject entirely.
If any of that would be a problem where you are, it is worth raising on the first call rather than in the second term.
Two more
Other schools,
Other schools,
other numbers.
All six are measured the same way. See all six
Questions
What teachers ask
What teachers ask
about this one.
The ones that come up almost every time.
No. It marked first and the teachers reviewed, which is the whole design. The saving is in not reading forty scripts from scratch — the reviewing, and the final say, stayed with the department.
Written English composition. The rubric that worked for Science and Social Science could not capture what that department was actually rewarding, and they went back to marking it by hand. The 96% figure excludes it.
Their own moderation sample said no, but they ran that sample for two terms before trusting it — which is what we would suggest to anyone.
Ready when you are
Bring us last term’s results.
Book a free 30-minute demo. We’ll build the same breakdown from a paper you already ran, so the first number you see is your own.