ScoreFluent
Plan the class from evidence, not from the next chapter. Before the lesson, tally last week's errors by type, count how many students share each one, and choose the one to three that most of the group shares and that one lesson can fix. Teach those, end with a short task on new material, and check next week whether the fix held. Leave the problems only a few students have for individual practice.
The method in one view
- Collect: last week's written and spoken work.
- Tally: errors by type, using the exam's own criteria as categories, and count students rather than errors.
- Choose: the one to three problems that most students share and one lesson can fix.
- Teach: open with the group's own errors, anonymised, then practise on fresh material.
- Check: the same tally next week. Is the error still near the top?
Why the next chapter isn't always the right lesson
A course book puts topics in a sensible order for an average student. Your class is not an average student. In a mixed group, the next unit is right for some, already known by others and too early for a few. Class time is the scarcest thing in exam prep. It is better spent on what most of the room gets wrong than on what the syllabus reaches next.
A score sheet doesn't solve this either. A list of bands tells you who is behind. It doesn't tell you what to teach on Monday.
Step 1: collect and tally the evidence
Take the last week of work: essays, short responses, recorded speaking. Tally errors by type, not by student. Use the exam's own criteria as your categories, such as IELTS's four writing criteria, MET's five, or PTE's speaking traits, then go one level down: "articles", "past tense in narratives", "unsupported reasons", "run-on sentences".
Count students, not errors. One student making the same mistake twenty times is an individual problem. Twelve students making it twice each is a class problem.
The tally is the slow, repetitive part of the job. It is also the part that is easiest to hand to software.
Step 2: decide what goes in class
| What you see | Where it goes |
|---|---|
| Shared by many students, teachable in one lesson (e.g. paragraph planning, linking words) | This week's class |
| Shared by many, but a long-term habit (e.g. a pronunciation pattern) | A short warm-up drill every lesson for several weeks |
| Shared by a few students | Individual practice outside class |
| One student, one-off | A note on their work, no more |
Three focus points are the most a single lesson can hold. One is often better.
Step 3: build the lesson around the group's own errors
A simple 60-minute shape:
- Warm-up (10 min). Show five or six real sentences from the group's work, anonymised, and ask pairs to find and fix the problem. Students recognise their own mistakes faster than a textbook's, and nobody is singled out.
- Teach (15 min). One explanation of the pattern: why it's wrong and what the right form does.
- Guided practice (15 min). Controlled exercises on that pattern only.
- Fresh task (15 min). A short exam task on a new prompt, where the pattern has to come up.
- Exit check (5 min). Each student checks their own fresh task for that one pattern only.
The fresh task matters most. Correcting a sentence you've been shown is easy. Getting it right in a new answer under time pressure is the skill the exam measures.
Step 4: check that it held
Next week, run the same tally. If the error has dropped down the list, move on. If it's still near the top, change how you teach it, for example with a different explanation, more controlled practice or a model answer. More of the same practice rarely moves it.
What to hand to software, and what to keep
Marking every attempt against the criteria, tallying errors and keeping the record over weeks are repetitive, and software can do them every day without getting tired. Deciding which three problems matter for this group and this exam date, explaining them so they make sense, and noticing the student who has gone quiet are teaching. Software can tell you that most of the class leaves out articles. It can't tell you why one student stopped speaking up in class.
Where ScoreFluent fits
In an institute using ScoreFluent, every student attempt is marked against the exam's criteria and each student gets a weakness profile. The institute console ranks the weak areas that affect the most students, which is the tally in Step 1, kept up to date for you. A per-student view shows where each student struggles by skill and task type. Students' answers stay private to them. The console shows patterns, not individual scripts.
It does not plan the lesson. That stays with you. Institutes can try it with five seats free for 15 days, without a card. See ScoreFluent for institutes. For the student side of the same method, see why a mock score doesn't raise your score.
Questions institutes ask
How do I decide what to teach in an exam-prep class?
Use last week's work. Tally the errors by type, using the exam's own criteria as categories, and count how many students share each one. Teach the one to three errors that most students share and that one lesson can realistically fix. Leave individual problems for individual practice.
How do I handle a mixed-level exam class?
Split the evidence into what the group shares and what only a few students need. Spend class time on the shared problems, and give the rest targeted practice outside class. A tally makes that split visible, so you are not guessing from the loudest students.
How often should I re-plan from student errors?
Weekly works for most courses. A fresh tally each week shows whether last week's focus held. If the same error is still at the top, change how you teach it rather than adding more of the same practice.
What part of exam-prep teaching can software do?
The repetitive parts: marking each attempt against the criteria, tallying errors and keeping the record over time. Deciding what matters for this group, explaining why, and noticing a student who has stopped trying stay with the teacher.
What does ScoreFluent show teachers?
ScoreFluent marks every student attempt and builds a weakness profile for each student. The institute console ranks the weak areas that affect the most students, and a per-student view shows where they struggle by skill and task type. It shows patterns, not students' individual answers.