OET Speaking re-sit: Band C to B playbook

There is a pattern we see almost every week on the OET Live support inbox. A nurse, often Philippines- or India-trained, took OET Speaking three months ago, got a Band C, and is now booked for a re-sit in six to eight weeks. They want to know what to do differently. The answer they expect — and almost always get from generic study advice on the open web — is "practise more".

If the first attempt failed at Band C, "practise more" is the worst thing you can do. It is the same study plan that produced the Band C, scaled up. The candidates who flip a C to a B between attempts almost never do it through volume. They do it through targeted intervention on the one criterion that actually lost them the most points.

This playbook is for those candidates. It assumes you have already taken OET Speaking once, you have your score report or at least a clear memory of which sub-scores were the lowest, and you have between four and eight weeks until the re-sit. If you have never taken OET Speaking before, the 8-week roadmap is what you want instead — different plan, different starting point.

Why most re-sits fail again: the "more practice" trap

OET Speaking is scored on nine independent criteria. A Band C means you landed somewhere between 300 and 349 on the converted score — often within 10 to 20 points of the 350 needed for Band B. That gap is small enough to be deceptive. It looks like the kind of gap that closes with general practice.

It almost never closes that way. Here is what actually produces a Band C:

• Seven criteria at B-level, two criteria at C-level. The two laggards drag the converted score down. This is the most common pattern in our user data and the one this playbook is mostly aimed at.
• Or, more rarely: eight criteria at B-level and one criterion at D-level. One criterion can sink the whole sub-test.
• Or, even more rarely: all nine criteria hovering at the B/C boundary, no single criterion clearly weak. This pattern is harder to fix and benefits less from this playbook — it tends to need a longer runway than 6 weeks.

The "practise more" approach evenly raises all nine criteria by a small amount. If your starting profile is seven Bs and two Cs, that means six weeks of evenly-distributed practice gets you to seven B-pluses and two C-pluses — still a Band C, just barely. You will hit the next test, get a C again, be confused about why, and your support inbox will look exactly like the ones we see.

The right approach is the opposite: identify the one or two criteria that scored lowest, spend 70 to 80 per cent of your practice time on those, and accept that the seven criteria you were already passing will hold their ground without dedicated drilling. The reason this works is that the B-level criteria are already overdetermined — you have practised them enough that they will not regress in six weeks of light maintenance. The C-level criterion, by contrast, has a specific structural cause that needs a specific intervention.

The diagnostic: pulling your score report apart

Step one is to figure out which criterion or criteria actually sank you. There are three sources of evidence, in order of trustworthiness:

1. Your official OET score report (if you requested it)

OET candidates can request a sub-score breakdown after the test. If you have not already done this, do it now — the breakdown is the single highest-value piece of information for re-sit planning. Without it you are guessing.

The breakdown gives you a band per criterion. Look for the criteria scored at C-level (band 4) or below. These are your targets. Ignore everything else.

2. A self-scored recording of a mock test

If you do not have the breakdown — and many candidates do not — the next-best option is to take one OET Speaking mock cold (no preparation, no special effort) and score yourself against the rubric criterion by criterion. Listen back twice: once for linguistic criteria (intelligibility, fluency, language appropriateness, grammar), once for clinical-communication criteria (empathy, patient perspective, organisation, clinical accuracy, vocabulary).

The honest version of this exercise produces a list. Something like:

• Intelligibility: B
• Fluency: B
• Language appropriateness: B
• Resources of grammar: C
• Empathy: B
• Patient perspective: C
• Organisation: B
• Clinical accuracy: B
• Vocabulary: B

If your list looks like this, "grammar resources" and "patient perspective" are your re-sit targets. Six weeks of dedicated work on those two criteria will move you from Band C to Band B. Six weeks of evenly distributed practice will not.

The harshest version of this exercise — the one that actually works — involves running the recording through an AI scoring tool that gives you a per-criterion score, and trusting it where it disagrees with your self-assessment. Most candidates over-score themselves by one band on the criteria they are weakest at. The AI will not.

3. The "memory" diagnostic

If you have no recording and no breakdown, the third option is to write down everything you remember about your first attempt. Where did the role-play feel awkward? What did the patient say that you struggled to respond to? Did you understand every question? Did you find yourself reaching for words mid-sentence?

This is the weakest evidence but it still narrows the field. If you remember the patient asking emotional questions and you remember responding clinically without acknowledging the feeling, your weak criterion is probably patient perspective (failing to address what the patient is actually worried about) and possibly empathy. If you remember getting stuck on phrasing and using fillers a lot, the weak criterion is probably fluency or grammar resources.

Write the list. Pick the two most likely targets. Move on. The rest of the playbook depends on having a target.

The 6-week intervention

This plan assumes 5 to 7 hours of practice per week. If you have less time, see "When to delay" at the end. The structure is the same as the 8-week roadmap, compressed and re-weighted toward your one or two target criteria.

Week 1: diagnose + commit to the target

Time: 3 hours. Goal: produce a written diagnostic and commit to the criterion you will be working on for the next five weeks.

Day 1: take the mock if you have not already, listen back, self-score against the rubric. Two hours.

Day 2: write a 200-word diagnostic. The format we recommend:

My target criterion: <name>
Why I think it failed me last time:
  1. <specific example from your recording>
  2. <specific pattern>
Evidence I will be tracking weekly:
  - <objective metric — fluency = words per minute,
     grammar = number of self-corrections per role-play, etc.>
What "passing this criterion" looks like in week 6:
  <describe a role-play snippet that would clearly hit Band B
   on this criterion>

Putting it in writing matters. The most common reason a re-sit fails is the candidate silently revised their target half-way through the plan — typically because the work was hard or boring and they pivoted to practising something they were already good at. The written diagnostic is the thing you re-read in week 4 when you start drifting.

Week 2 and 3: structural intervention on the target

Time: 5–7 hours per week. Goal: change the underlying behaviour, not paper over it.

This is the hard part. The intervention depends entirely on which criterion you targeted.

If your target is intelligibility or pronunciation: drill the specific sound substitutions that caused the most listener confusion. For Filipino-trained candidates, this is usually the /θ/ → /t/ substitution in words like "thirty / thirteen". For Mandarin-trained candidates, it is usually consonant clusters at word endings. For Arabic-trained candidates, it is /p/ versus /b/. See our pronunciation guide for Tagalog speakers for the pattern; substitute your own L1 if different.

The intervention is minimal-pair drilling — 10 minutes a day, 6 days a week, on a list of 30 to 50 word pairs that contain your target contrast. Two-week pronunciation drilling on the right pairs moves intelligibility from a 3 to a 5 reliably. General "speak more English" practice barely moves it at all.

If your target is fluency: the intervention is slow speech with controlled pauses. Speak at 90 to 110 words per minute, deliberately. Pauses go at clause boundaries, not in the middle of clauses. This sounds counterintuitive but the rubric does not reward fast speech; it penalises pauses that disrupt meaning. A candidate speaking at 110 wpm with clean pauses sounds more fluent than a candidate speaking at 160 wpm with mid-clause hesitations.

Drill: record yourself doing five role-plays at the slower pace. You will hate the first three. By the fifth, the rhythm will feel natural.

If your target is grammar / resources of grammar and expression: the intervention is phrase memorisation, not grammar study. Memorise 30 high-frequency clinical phrases — "I can see this has been difficult for you", "It is completely understandable to feel that way", "Let me explain what we found", "There are a few things we can try together". Use them. The rubric is not testing whether you can parse a relative clause; it is testing whether your speech contains varied, grammatically clean clinical phrases.

A candidate who memorises 30 of these and works them into every role-play scores B on this criterion within three weeks. A candidate who studies grammar textbooks scores C still.

If your target is empathy: the intervention is stating the patient's emotion before answering their question. Every time the patient expresses worry, fear, or frustration, your first sentence reflects it back — "That sounds really frightening" — before you proceed to the clinical content. This is a single, learnable behaviour that moves empathy scores by one full band in two weeks.

If your target is patient perspective: the intervention is asking one question that surfaces the patient's actual concern in the first 30 seconds of every role-play. Not "do you have any questions" — that is empty. Something more specific: "what is worrying you most about this?", "what would you like to understand first?", "is there something specific you came in hoping I would address?". This shows the assessor you are taking patient perspective into account from the start, not as an afterthought.

If your target is organisation of talk: the intervention is signposting. Every role-play opens with a 10-second outline of what you are going to cover — "I would like to talk through three things today: first, what the test results showed; second, what that means for you; and third, what we should do next" — and then explicitly transitions between sections. Two weeks of this and your organisation score jumps by a band.

If your target is clinical accuracy or vocabulary: the intervention is building a personal glossary of the specific clinical terms in your professional area. Refresh it every day for the first two weeks. By week 3 you will reach for the right word automatically.

Pick the intervention for your criterion. Spend weeks 2 and 3 on it. Resist the urge to "also work on" the other criteria. They will hold.

Week 4: integrate, do not regress

Time: 5 hours. Goal: combine the targeted intervention with full role-plays, without losing what you have built.

Do three full role-plays this week. Listen back to each one. Score it on the target criterion only. The first one will probably regress slightly — you are juggling more cognitive load. The second one will hold. The third one will be at the new level, but now under role-play conditions.

This is also the week to re-read your written diagnostic from week 1. If you have drifted, restart the intervention. There is no shame in repeating week 2 — better than going into the re-sit with no real change in the criterion.

Week 5: full mocks, half-pace

Time: 6 hours. Goal: build re-sit endurance and stress-test the intervention.

Three full OET Speaking mocks (two role-plays each), with self-scoring after each. The intervention you trained for two weeks should now be holding under stress. If it is not — if the target criterion regresses under mock conditions — the answer is to repeat week 4, not to do more mocks.

A common failure mode at this stage: candidates take five mocks in a week and score Cs on all of them, conclude they "are not ready", and panic. Five mocks in a week is not productive. Three with deep listening back is.

Week 6: final polish + exam-day rehearsal

Time: 4 hours. Goal: arrive at the test rested, with a clear pre-test routine.

Day 1–3: one mock under full exam conditions — including the actual reading-time you will have on the day, no pauses, no second takes. Listen back. Make notes.

Day 4–5: rest. Light reading only. No new practice. The plateau is real and you do not improve by drilling on the last 48 hours.

Day 6 (test day): pre-test routine. Light breakfast, no caffeine if you are sensitive to it, 10 minutes of warm-up speaking out loud in English — read a news article aloud — to get your articulators loose before the test.

Three "stuck at C+" archetypes

Across hundreds of re-sit candidates, we see three recurring profiles. Identify yours and the intervention is faster.

Archetype 1: the clinically excellent, communicatively understated nurse

Profile: experienced clinician, often 8+ years bedside, scores B or A on clinical accuracy and vocabulary, scores C on patient perspective and empathy. The pattern: the candidate gives correct clinical information but does not address what the patient is actually worried about.

Why this happens: in busy clinical practice, you learn to deliver information efficiently. The OET role-play rewards a slower, more patient-centred style than real shifts allow.

Fix: the patient perspective intervention from week 2 above. Two weeks of asking the patient one specific concern-surfacing question at the start of every role-play moves both empathy and patient perspective up a band. Total time investment: 4 hours.

This archetype passes the re-sit reliably with a targeted intervention. We see it at probably 30 per cent of our re-sit users.

Archetype 2: the fluent talker with grammar leaks

Profile: high fluency, good vocabulary, scores B on most criteria, scores C on grammar resources. The pattern: under role-play stress, the candidate produces sentences with verb-tense slips, article omissions, or preposition errors that pull the grammar score down.

Why this happens: the candidate is good enough at English that they have stopped consciously monitoring their grammar. Stress + speed compounds — by the second role-play their accuracy drops further.

Fix: the phrase memorisation intervention from week 2. Memorise 30 high-frequency clinical phrases with clean grammar. Under stress, the candidate reaches for memorised phrases instead of constructing sentences from scratch. The errors disappear because the phrases are pre-built.

This archetype also passes reliably with the right intervention.

Archetype 3: the diligent over-preparer

Profile: studied for months, can recite the rubric, scores B on linguistic criteria, scores C on patient perspective or empathy. The pattern: the role-play sounds rehearsed. The candidate hits all the rubric points mechanically — "I acknowledge your concern" — but the assessor scores it as performative rather than empathetic.

Why this happens: over-preparation produces a script. Scripts feel inauthentic to assessors, who are trained to detect mechanical empathy.

Fix: this is the hardest archetype to fix in 6 weeks. The intervention is to stop drilling rubric phrases and instead spend two weeks practising in unscripted conditions — peer practice, AI patient practice, free conversation in English with a friend playing patient. The goal is to recover natural conversational rhythm. By week 4 the rehearsed quality fades.

About 15 per cent of our re-sit users fall into this archetype. Most pass on the second attempt; some need a third.

What to do differently on exam day vs. the first attempt

Even with a perfect 6-week intervention, the second test attempt has its own failure modes. Three things to do differently:

Slower start to the first role-play. Your first attempt probably opened with high anxiety and a slightly rushed first 30 seconds. Deliberately slow the first role-play opening by 20 to 30 per cent. Take longer reading the role card. Open the role-play with 5 to 8 seconds of greeting and patient-name use before getting into the content. This single change adds clarity and gives you time to settle.

Re-read the role card for the patient's emotional state, not just the clinical scenario. Most first-time candidates read the role card for the medical facts. Re-sit candidates should also note any line about how the patient is feeling — "anxious", "frustrated", "worried about returning to work". These cues are what your patient-perspective intervention is targeting.

Active recovery from a stumble. If you make a mistake or misspeak in the first role-play, you cannot rewind, but you can recover. Have a one-sentence reset memorised — "Let me put that more clearly" — and use it once if you need to. Two times is too many; once shows composure.

When to delay the re-sit

The 6-week plan above works for candidates whose first attempt scored Band C with one or two specific weak criteria. It does not work as well if:

• Your first attempt was Band D or below. That signals a broader gap than this playbook is designed to fix. The 8-week full plan — or its 12-week extended variant — is the right starting point.
• You have less than 4 hours per week to practise. The targeted intervention works because it is concentrated. With under 4 hours, it dilutes too much. Delay 4 to 8 weeks and find more time.
• Your weak criteria are intelligibility AND organisation. Both are slow-moving, and fixing both at once divides attention. Pick the one with more impact (intelligibility for Band C candidates who are losing points across multiple criteria; organisation for candidates whose linguistic criteria are otherwise solid).
• You have not received your sub-score breakdown and cannot self-score reliably. Without a target, the plan loses its core advantage. Request the breakdown, wait the 4 weeks it takes to come through, then start the plan.

Delaying a re-sit is not failure. Re-sitting and getting another C is much more discouraging than waiting 4 weeks and getting a B.

What this plan is not

A few things to be clear about, since these come up in our support inbox:

• This is not a substitute for the rest of OET preparation. Listening, Reading, Writing scores stand on their own — if any of those also failed, this plan does not address them.
• This is not a guarantee. About 70 per cent of our users who follow a targeted-intervention re-sit plan flip C to B on the second attempt; the rest need a third. The discriminator is usually the third archetype above (the over-prepared script).
• This is not an excuse to skip the diagnostic. Candidates who skip week 1 and dive into the intervention without confirming the target criterion have the worst outcomes of any group we track.

The next step

Step one is the diagnostic. Pull up your score report — or your recording, or your memory — and write down the one or two criteria you will be working on. Write the 200-word diagnostic from week 1. Then start the intervention.

If you want to score a fresh role-play against the 9-criterion rubric to confirm the target before you commit six weeks of practice to it, the OET Live app does that. Run a sample role-play, get a per-criterion breakdown, and use it as the baseline for your written diagnostic. The intervention works; the diagnosis is the part that decides whether you get a B or another C.