How to Teach IELTS Speaking Part 2 (The Long Turn)

IELTS Speaking Part 2 — the long turn — is a one-to-two-minute solo talk delivered from a cue card after exactly one minute of preparation. It is the only stretch of the speaking test where the candidate talks alone, with no examiner prompts to lean on, and it is where the fluency and coherence score is most visibly won or lost. That combination makes it the highest-value part of the test to drill in your lessons.

This piece covers what the task actually demands, how to teach the preparation minute, a four-move structure students can run on any card, five drills that build stamina, and the disasters that sink otherwise well-prepared candidates.

What actually happens in the long turn?

A Part 2 task is a cue card: a topic, three or four prompts, and a final instruction. A typical card reads: Describe a person who has taught you something important. You should say: who this person is, how you know them, what they taught you, and explain why it mattered.

The candidate gets one minute to prepare, with pencil and paper, then speaks for one to two minutes. The examiner stops the talk at two minutes — the interruption is the test working as designed, not a failure — and may ask one or two rounding-off questions afterwards. The whole segment runs three to four minutes of the eleven-to-fourteen-minute speaking test, sitting between the short answers of Part 1 and the two-way discussion of Part 3.

Teach the mechanics explicitly, because candidates carry surprising beliefs into the room. Plenty quietly believe they should talk until they run dry, or that being stopped at two minutes means they have waffled. Neither is true. Two minutes is the target, and a talk that ends before the one-minute mark is the only real timing failure.

Why the long turn is the highest-value part to drill

The long turn is the only stretch of the speaking test the candidate fully controls. Parts 1 and 3 are reactive — the examiner sets the pace and chooses every question. Part 2 is self-generated: two minutes of connected speech with nothing but a card, sixty seconds, and whatever plan the candidate wrote. Fluency under those conditions is a distinct skill, and it does not arrive automatically with good Part 1 answers.

It is also where fluency and coherence — a quarter of the total speaking mark — is most exposed. Two minutes of sustained talk is where linking phrases, signposting and smooth self-correction either appear or do not. The transfer runs one way: a candidate who can hold a structured two-minute turn brings that composure into Part 3, but drilling Part 1 short answers does almost nothing for the long turn. Put your lesson hours where the leverage is. If you are assembling the wider course around this, my notes on teaching exam prep online cover how the speaking paper sits alongside the written ones.

There is a confidence dividend too. Students who can reliably fill two minutes stop fearing the card, and that calm is visible in the rest of the test.

What should students write in the preparation minute?

The preparation minute should produce a map, not sentences. Four points of two to five words each — glanceable while speaking — beat two full sentences that will be read aloud verbatim. A candidate who spends fifteen of their sixty seconds writing "I am going to describe my grandfather, who taught me patience when I was a child" has bought something they will never use.

Give students a fixed skeleton to complete inside the first forty seconds:

Ground — the prompt that anchors the talk: who, where, when, or what.
Detail — one specific: a name, a place, an object, a number.
Feeling — one opinion, emotion or comparison.
Forward — a look ahead: "now", "next time", "if I could".

Then teach angle selection. Most cue cards allow several interpretations, and the right choice is the easiest one, not the cleverest: the specific friend rather than "a person I admire in general", the trip that actually happened rather than an invented ideal. Have students scan the prompts and pick the version with the most real detail attached. A story they remember is almost always easier to sustain than a performance, because memory supplies detail for free.

One more rule: keywords only. If a student can read their notes aloud as sentences, the notes are wrong — and the delivery will sound like reading, which is exactly what the fluency criterion punishes.

How do you structure the two-minute turn?

A long turn that fills the time runs four moves: past context, main narrative, detail and feeling, wrap-forward. At roughly thirty seconds each, the structure fills the full two minutes — and it forces a useful spread of tenses along the way.

Past context (15–20 seconds). Set the scene: when, where, who. "This takes me back to... It was a few years ago, when I was still living in..." Past simple and past continuous do the work here.
Main narrative (50–60 seconds). The story itself, walking through the card's prompts in order. Sequence markers ("after that", "in the end") plus a past perfect ("I had never seen anything like it") show range without showing off.
Detail and feeling (30–40 seconds). Zoom in on one moment: what it looked like, how it felt, what you would tell a friend about it. This is where lexical range lives — and it is the move students skip.
Wrap-forward (15 seconds). Close by looking ahead: "Looking back...", "If I could do it again, I would...". One conditional, and the talk lands on purpose instead of trailing off.

Drill the moves as moves. A student who owns "context, story, zoom, forward" can run the structure on any card, which is the entire point: the topic changes on test day, the structure does not.

What do examiners actually score?

Examiners score delivery and language, never the truth of the story. The four criteria — fluency and coherence, lexical resource, grammatical range and accuracy, and pronunciation — carry equal weight, and none of them rewards honesty about whether the cousin in the story exists. The public band descriptors on ielts.org spell all of this out, and they are worth an hour of your own reading before you teach to them.

Three consequences follow. First, fluency is measured as flow, not speed: a smooth self-correction ("my cousin — sorry, my cousin's wife — picked us up") reads as fluency, while five silent seconds hunting for the perfect word does not. Second, examiners are trained to recognise memorised answers — flat intonation, a suspiciously polished opening, then a hard stop the moment the rounding-off question leaves the script. Third, invented content is never penalised for being invented, so candidates should build their talk on whatever story they can tell most fluently, bending the card to fit.

Which drills build long-turn stamina?

Stamina comes from short, repeated exposure, not from marathons. Five drills that earn their place in a lesson plan:

Sixty-second topics on rotation. Any noun on the board — "a meal", "a journey", "a decision" — sixty seconds, timer visible. Rotate the topics every few days so students practise starting cold rather than polishing one talk to death.
Record and self-review. Record the turn, then have the student count their own fillers and mark the second where the talk ran dry. Self-diagnosis on the second listen sticks far better than your correction, and it trains the habit of reviewing their own speech.
Buy-time phrases on repeat. A small set, drilled until automatic: "That's an interesting one — let me think...", "I'd never really considered it, but...". These are scored as natural discourse, unlike a bare "err", and they buy the two seconds a stalling brain needs.
Skeleton in sixty seconds. New card, sixty seconds on the clock, four points of four words maximum. Prep speed is a separate skill from delivery, and it needs its own practice.
The expansion re-run. The student delivers a 45-second version, then re-runs it with one forced extra layer per point — a feeling, a comparison, a mini-story. This is the drill that teaches students where extra seconds actually come from.

Keep a log of every attempt: topic, seconds sustained, filler count, one language note. A visible trajectory does more for a nervous student's confidence than any pep talk, and it shows you exactly which move of the structure is collapsing first. If you keep records online, a shared progress tracker makes the trend visible to the student as well as to you — and if you record the takes inside your teaching platform, the classroom tools keep every recording and note with the lesson they belong to.

One skeleton per cue-card type

Cue cards cluster into four types, and each type rewards a slightly different plan. Hand students the table and drill one type per lesson.

Cue-card type	Skeleton plan	High-value language
Person	Who + how you met → what they are like → a specific memory → why they matter	"I've known her for...", "What makes him remarkable is...", "If I saw her tomorrow, I'd..."
Place	Where + when you went → what it looks like → what you did there → why you would return	"It's the kind of place where...", "I was wandering around when...", "I'd go back in a heartbeat if..."
Object	What it is + who gave it → what it looks like → how you use it → why it matters	"It was given to me by...", "What makes it special is...", "I'd be lost without it if..."
Event	When + where → what happened → who was there → why it is memorable	"It had already started by the time...", "In the end...", "If it hadn't been for X, I would never have..."

Every row ends in a conditional or a forward look, deliberately: that is the move candidates almost never produce unprompted, so it goes in the skeleton where they cannot avoid it.

Which disasters sink prepared candidates — and what fixes them?

Five failures account for most collapsed long turns, and each has a short, teachable fix.

Disaster	Root cause	Fix
Finishing at 45 seconds	Plan exhausted; no expansion layer	Drill the zoom-in: one feeling and one comparison per point
Freezing mid-turn	Brain racing ahead of the mouth	Buy-time phrases; treat restarting a sentence as fluency, not failure
Sounding memorised	Rehearsed script with flat delivery	Ban sentence notes; rotate same-type cards so no script fits twice
Reading notes aloud	Prep sheet written in sentences	Enforce four-to-six words per point during the prep minute
Drifting off the prompts	Long turn treated as free-form	Use the prompts as rails: touch each one, in order

The forty-five-second collapse is worth demonstrating live. Ask a student for their standard talk — it hits the four points and stops — then re-run it with one added layer per point: a feeling, a comparison, a half-sentence anecdote. The difference is usually forty seconds of speech the student already owned. Once they see where the time was hiding, the fear of "not having enough to say" mostly dissolves.

The memorised-answer problem is the serious one, because it can cap an otherwise strong performance. A script that fits every card sounds flat, races at unnatural speed, and collapses the moment the examiner's follow-up leaves the prepared track. The honest fix is variety: one memorised talk cannot survive twenty cue cards of the same type, but eight flexible skeletons can cover most of the deck.

Frequently asked questions

How long should a student speak in IELTS Speaking Part 2?

The examiner asks for one to two minutes and stops the talk at the two-minute mark. Students should aim to fill close to the full two minutes; a talk that ends before one minute makes strong fluency scores very difficult.

Can a student choose their own topic in Part 2?

No — the cue card is chosen by the examiner. What students control is the angle: the prompts allow loose interpretation, so the safest strategy is the true story they can tell with the most detail.

Is it acceptable to invent a story in the long turn?

Yes. Examiners score language and delivery, not the truth of the content, and an invented story carries no penalty provided it does not cause hesitation. In practice, most students speak more fluently about a real memory stretched to fit the card than about a wholesale invention.