Most Band 5 candidates who plateau there don’t have a talent problem — they have a strategy problem. They keep doing the same kind of practice (chatting in English, running through cue cards, hoping fluency arrives on its own) and wonder why the band never moves. The jump from 5 to 7 isn’t a vague “get better at English” project. It’s a specific, checkable set of changes across four scored criteria — Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation — and it responds far better to a system than to volume. Below is that system, in the order it actually needs to run.
Try the free 2-minute AI speaking diagnostic →
Step 1: Diagnose your real weak criterion before you touch a single drill
The single biggest reason candidates stall at Band 5 or 6 is that they train the wrong thing. Someone who is actually held back by pronunciation spends months memorizing vocabulary lists. Someone whose grammar is fine but whose fluency collapses under pressure keeps grinding grammar exercises that were never the bottleneck.
The IELTS Speaking band is an average across four criteria weighted equally, so your overall band is really hiding four separate sub-scores. A 5 overall could mean “5 across the board” or it could mean “7 on grammar, 3 on pronunciation” — and those two candidates need completely different training plans. Before doing anything else:
- Record yourself answering a full Part 1, Part 2, and Part 3 set under timed conditions.
- Listen back (or get scored feedback) against each of the four criteria separately, not just for an overall impression.
- Identify which one criterion is genuinely dragging your average down the most — not which one feels most uncomfortable, which one the descriptors show is actually weakest.
This diagnosis step is the difference between training and guessing. Skip it and you can practice for months without your band moving, because you’re reinforcing skills you already have while the actual bottleneck sits untouched.
Step 2: Drill the specific sub-skill that criterion needs, not the criterion in general
Once you know your weak criterion, the mistake is to attack it with generic advice like “improve your vocabulary” or “work on fluency.” Those instructions are too broad to train directly. Each band gate from 5 to 7 has a small number of named, isolated sub-skills behind it, and targeting those directly is what moves the needle.
A few real examples, criterion by criterion:
- Fluency and Coherence (5→7): learning to fill thinking time with real language (“well, let me think about that…”) instead of silence or fillers, extending every answer to three-plus sentences by default, and building spoken discourse markers (“the thing is,” “having said that”) rather than relying on written-style linkers that sound stiff out loud.
- Lexical Resource (5→7): building active topic-specific word banks you can recall under time pressure, learning to paraphrase around a gap instead of freezing, and developing precision between near-synonyms rather than reaching for the most “impressive” word.
- Grammatical Range and Accuracy (5→7): moving beyond simple sentences into one reliable complex structure at a time (relative clauses first, then conditionals), and specifically drilling hypothetical language (“if I had more time, I would…”) — examiners listen for this structure directly at this band gate.
- Pronunciation (5→7): correcting word stress on multisyllabic words, since misplaced stress causes more comprehension breakdown than most individual sound errors, then building sentence stress and natural chunking (pausing at clause boundaries instead of randomly).
The point isn’t to drill everything at once. Pick the two or three sub-skills tied to your actual weak criterion and train those in isolation — short, repeatable drills, not open-ended conversation practice — until they hold up consistently, not just once when you’re paying close attention.
Step 3: Rebuild it under real exam conditions
Isolated drilling teaches a skill in a controlled setting. It does not prove the skill survives exam pressure — the 60-second prep clock, the stranger across the table, the unpredictable follow-up questions. That’s a different test, and skipping it is why some candidates drill for weeks and still underperform on test day.
Once a sub-skill feels solid in isolation, force it back into timed, full-format practice:
- Full Part 1 (4–5 minutes), Part 2 with the actual one-minute prep and two-minute talk, and Part 3 with unscripted follow-ups — run back to back, not separately.
- No pausing, no restarting, no editing your answer after the fact.
- Record the whole thing, exactly as you would sit the real test.
This step is where you find out whether a skill is real or just rehearsed. A discourse marker you can use cleanly in a slow drill often disappears the moment you’re improvising an answer to a Part 3 question you didn’t expect.
Start step one of the system today — ActuallySpeak’s free 2-minute AI speaking diagnostic tells you exactly which criterion to target first.
Step 4: Get specific feedback and iterate on the same criterion
The loop only closes if you measure whether the score on your target criterion actually moved — and moved consistently, not just in one lucky recording. Generic feedback like “sounds better” or “good job” is not useful here; you need feedback that names the specific sub-skill, tells you whether it showed up under exam conditions, and flags what’s still missing.
After each timed mock, go back to the same criterion you diagnosed in Step 1 and ask three questions: did the target sub-skill appear naturally, without you consciously forcing it? Did it hold up through all three parts, including the more abstract Part 3 questions? And is there a new, smaller gap now visible underneath the one you just closed? IELTS band gates layer — closing one issue often reveals the next one sitting just behind it (say, fluency improves, and now a grammar pattern you never noticed because you were hesitating too much becomes the visible ceiling). Re-diagnose, re-target, re-test. That repeating cycle — not open-ended practice — is what actually carries a candidate from Band 5 to Band 7.
Watch the full lesson
FAQ
How long does it realistically take to go from Band 5 to Band 7 in speaking?
There’s no fixed timeline that applies to everyone, because it depends on how many criteria need work, how much time you can dedicate to focused practice, and how consistently you follow a diagnose-drill-test-iterate cycle rather than unstructured conversation practice. Candidates who isolate their weakest criterion and drill it deliberately tend to see steadier progress than those who simply “speak more,” but treat any specific week or month count you see elsewhere with caution.
Can I improve pronunciation without a native-speaking environment?
Yes. Pronunciation at the Band 5–7 level is mostly about control of specific features — word stress, sentence stress, and chunking — which can be trained through recorded shadowing and targeted minimal-pair drills against your own accent patterns, without needing to live somewhere English is spoken. What matters more than exposure is deliberate, corrected repetition of the exact sounds and stress patterns that are currently causing strain for a listener.
Is grammar or vocabulary more important for reaching Band 7?
Neither is inherently more important — each is 25% of your speaking score, so the right answer depends on which one is actually holding your average down, which is why diagnosis comes first. A candidate with strong vocabulary but simple, repetitive sentence structures needs grammar work; a candidate with varied grammar but narrow, repeated word choice needs lexical work. Assuming you know which one it is without checking is exactly the guessing this system is built to avoid.
Do I need a tutor to run this system, or can I do it alone?
You can run the diagnostic, drilling, and timed-mock stages independently if you’re disciplined about recording yourself and checking against the actual band descriptors rather than your own impression. Where solo practice tends to break down is the feedback stage — it’s hard to hear your own recurring errors objectively, and generic self-assessment (“that felt okay”) isn’t the same as criterion-specific feedback. A tutor, an examiner-trained rubric, or a scoring tool built around the same four criteria can fill that gap.
What’s the single biggest mistake candidates make trying to jump from 5 to 7?
Practicing broadly instead of targeting the specific criterion and sub-skill actually holding their band down. Talking to a language partner for an hour a day feels productive, but if your real bottleneck is word stress or hypothetical grammar, general conversation may never force you to correct it. The fix isn’t more speaking — it’s a diagnosis that tells you exactly what to speak about differently.