Why most students plateau at 6.5
Band 6.5 is a strange place to be stuck. You can hold a conversation. You can read a Hindu editorial. You can write a five-paragraph essay in under 40 minutes. The exam isn't unfamiliar any more — and yet, every attempt comes back the same.
After grading a few thousand attempts, the pattern is consistent: students at 6.5 are practising the wrong things, in the right order. They're doing more of what got them from 5.5 to 6.5 — vocabulary lists, full mock tests, generic templates — when the move from 6.5 to 7.5 is a different category of work entirely.
The move from 5.5 to 6.5 is about volume. The move from 6.5 to 7.5 is about precision.
Volume tactics — more practice tests, longer word lists, watching every YouTube IELTS playlist — give diminishing returns past 6.5. They were the right strategy a year ago. They are not the right strategy now. Precision tactics are: identifying the specific 3-4 mechanical errors that recur in your transcripts, and drilling those, alone, until they are gone.
The four-question framework
Before any 21-day plan can work, you need to know what to drill. Pull out your last two scored attempts — full transcripts of your Speaking, both Writing tasks — and answer these, honestly, in writing:
1. Which module is dragging me down?
Look at your most recent overall band. Now look at the four module scores. The lowest two are where 70% of your prep should go for the next three weeks. If your Reading is already 8.0, do not spend another minute on Reading. Even if reading feels productive. Even if you enjoy it.
2. What's the recurring mistake in my weakest module?
Open the transcript. Read it slowly. Underline the first error of each type you see — grammar, vocabulary, coherence, pronunciation. Stop at five. Almost certainly, you'll find the same two or three categories repeat across both attempts. That's your drill list. Not the words. Not the topics. The categories.
3. Am I making errors I know are errors, or errors I can't yet see?
This is the diagnostic split. If your errors are "I knew the answer was 'fewer' but said 'less' in the rush" — those are speed/anxiety errors and disappear with timed drilling. If your errors are "I didn't know the rule" — like comma splices, present-perfect-versus-past, articles before abstract nouns — those need 20 minutes of explicit study, then 20 minutes of practice each, for about a week.
If you're using an AI scoring tool (ours or another), it will tell you what you got wrong but not which category it belongs to. The category step is on you. Without it, you'll keep drilling the wrong layer.
4. What does my best attempt look like — and what was different about that day?
Find your best score from the last 90 days. Read it. Notice the structure, the connectors, the way you opened paragraphs. That's your achievable ceiling on a good day. The goal of the next 21 days isn't to invent a new way of writing or speaking — it's to make your good days more frequent. That means sleep, exam-time alignment, and reducing decision-fatigue on test day.
The 21-day plan
Once you've answered the four questions, the plan writes itself. Here's the structure 80% of our students follow, with a 90-minute daily commitment. You can run this without a tutor. You will need one good practice platform, a quiet 30-minute slot for Speaking, and the discipline to read your own transcripts the next morning.
21 days will move a real Band 6.5 student to around 7.0–7.5. It will not move a 5.5 to a 7.0 — that's a 4–6 month commitment. If you're below 6.5 on a real mock, build a longer base before running this plan.
Speaking — the daily drill that actually works
Most 6.5 Speaking attempts fail on two specific axes: fluency under pressure (long pauses, fillers, restarting sentences) and lexical resource (over-using "very good", "really nice", "a lot"). Both are fixable in two weeks of structured practice.
The drill is this. Each day, pick one Part 2 topic at random. Speak for two minutes — recorded, on a phone, no preparation. Listen back the next morning. Count three things:
- Filler count. Every "um", "uh", "like", "you know". Goal: under 6 in two minutes by day 14.
- Generic adjectives. Every "very", "really", "a lot", "good", "bad", "nice". Goal: replace at least three with module-2 vocabulary the next day.
- Restart count. Every time you start a sentence and abandon it. Goal: down to one restart per 2 minutes by day 14.
That's it. Three numbers, tracked on a notes app, daily. By day 14, the change is visible in the recordings. By day 21, it's in your score.
Writing — the rubric trick most students miss
IELTS Writing is scored on a public rubric — four criteria, each on a 9-band scale. Most students study examples of high-band writing, then try to imitate the style. This is backwards. The rubric is the actual scoring document. Read that.
Pull up the public IELTS Writing Task 2 rubric (the official one, not a third-party summary). Read the Band 7 descriptors for each of the four criteria. Then read the Band 8 descriptors. The differences are very specific — and they are the entire game.
Example. The difference between Band 7 and Band 8 Coherence and Cohesion is essentially: "logically organised" (7) vs "uses cohesion in a way that attracts no attention" (8). Band 7 students are visibly using connectors ("Firstly... Secondly... In conclusion..."). Band 8 students hide them ("This pattern emerges from a different cause...").
Pick one criterion. Read its descriptors. Write three Task 2 essays focused only on closing that one Band 7 → Band 8 gap. Repeat for the other three criteria. That's four weeks of deliberate Writing practice that beats sixteen weeks of "write more essays".
Five mistakes to skip
- Doing back-to-back mocks. A full mock without a 24-hour transcript review is fitness training, not skill training. Both have a place — but you've already built the fitness by Band 6.5.
- New vocabulary lists. By 6.5 you have enough vocabulary. Use what you have, more precisely. Adding 200 new words in 21 days is a distraction; using your existing 4,000 words 20% more accurately is the win.
- Memorising templates. Examiners can spot a memorised opening by paragraph two. Templates lower your score above Band 6.5.
- Practising the wrong test. If your university accepts PTE and PTE plays to your strengths (faster, AI-only), switch. Don't waste 21 days hammering the same exam you're stuck on. Run our score-compare tool first.
- Skipping sleep the week before the test. Sleep loss in the 7 days before the exam costs ~0.5 bands across Speaking and Listening. Treat your sleep schedule as part of the plan.
A reality check
21 days is a real timeline, but it is not a magic timeline. If you have a test in 7 days and you're at 6.5, this plan won't get you to 7.5 — it will get you to 6.5–7.0 if you execute it well. If you have 60 days, run this plan twice, with a 7-day rest between them. If you have 6 months, ignore this plan and build a base instead.
The best signal that the plan is working isn't the score on day 21. It's whether you can read your own transcript on day 14 and tell, without grading, what would have lost you marks. When that skill arrives, the band score follows within 2–3 attempts.
If you want a second pair of eyes on your transcript, we offer human re-grades through a small network of CELTA-certified examiners — details on the prep page. The free 30-minute counselling call includes one transcript review.