Part 2 is the only part of the test where you speak alone, uninterrupted, for two minutes. It is also the part candidates prepare for most and misunderstand most.
Two beliefs cause most of the damage: that you must cover every bullet point, and that the preparation minute is for writing sentences.
Both are wrong, and both cost marks.
The structure, in short
| Preparation | 1 minute, with paper and pencil |
| Speaking | 1–2 minutes, uninterrupted |
| Under 1 minute | Penalised under Fluency and Coherence |
| Stopped at 2 minutes | Good — it means you sustained it |
| Shape | Introduction 5–10s · narrative 70–80s · evaluation 20–30s |
| Notes | Keywords only, never sentences |
The bullet points are not compulsory
This surprises people, so it is worth being precise.
Task Response is not a marking criterion in IELTS Speaking. It exists in Writing; it does not exist here. You are marked on fluency and coherence, vocabulary, grammar and pronunciation — and none of those measure whether you addressed bullet three.
So the bullets are a suggested structure, not a checklist. A talk that covers three of four bullets richly scores better than one that ticks all four in a rush.
Use them as a spine, because a chronological order is easier to speak from than no order at all. But if you lose one, keep going. Stopping to backtrack costs you far more than the missed bullet ever would.
The last bullet is where band 7 lives
The final prompt is almost always explain why — why it mattered, why you remember it, how you felt.
That is the only bullet that demands evaluation rather than description, which means it is the only one that naturally produces conditionals, hypotheticals and abstract vocabulary. It is also the one candidates rush, because it comes last and they are running out of time.
Give it 20 to 30 seconds. Deliberately.
What to write in the preparation minute
Not sentences. This is the most common preventable error in Part 2.
You can write three or four full sentences in a minute at best. Having written them, you will read them out — and reading aloud is instantly audible. It flattens intonation, kills eye contact, and produces the unnatural delivery that Fluency and Coherence marks down.
Write a keyword outline instead: single words, a collocation or two, and one grammar structure you want to hit. Notes are visual triggers, not a script.
A workable minute:
- 0–15s. Choose your subject and commit. The first genuine thing that comes to mind is almost always better than the clever thing you find at 40 seconds.
- 15–40s. Note two or three specifics — a place, a person, a moment.
- 40–60s. Note the language you want to use, and your closing reflection.
How do you fill two minutes?
Three phases:
- Introduction, 5–10 seconds. One or two sentences naming what you are going to describe.
- Narrative, 70–80 seconds. The descriptive bullets — who, where, what — in order.
- Evaluation, 20–30 seconds. Why it mattered. The last bullet.
An alternative that works when the topic is an object or a habit is past, present, future: how it came into your life, what it is like now, what you expect of it. That shape forces three tense groups on you, which showcases grammatical range without you having to think about grammar.
If you dry up early
Go to evaluation and speculation. Description runs out; opinion does not.
Why it mattered. Whether you would do it again. What you would change. Whether other people would feel the same.
Those four questions can carry thirty seconds on any topic on the card.
Model answer
Describe a skill you learned that took you a long time to master.
You should say:
- what the skill is
- how you learned it
- how long it took you
- and explain why it took so long to master.
I’d like to talk about learning to swim properly, which I only managed as an adult and which took me an embarrassingly long time.
I grew up a long way from the coast and nobody in my family swam, so I reached my late twenties having never really learned. I could stay afloat, more or less, but I couldn’t swim a length without stopping. What finally pushed me into it was a friend signing up for a triathlon and suggesting, not very seriously, that I join her.
I started with adult beginner lessons at a local pool, one evening a week. The instructor was patient, although I suspect I was the least promising person in the group. After a few months I moved on to swimming on my own, and then to longer sessions once I could actually complete a length without panicking.
Altogether it took me the best part of two years to feel genuinely comfortable in deep water, which is far longer than most people would need.
As for why it took so long, I think the physical side was never really the problem. The difficulty was that I’d built up a fear of water over about twenty-five years, and no amount of technique instruction addresses that. Every time I got out of my depth, my breathing went and the technique fell apart with it. If I’d learned as a child, I doubt any of that would have happened — children seem to acquire it without the fear, whereas adults have to unlearn something first. That’s the part nobody warns you about.
264 words — almost exactly two minutes at a natural speaking pace of around 130 words per minute.
What that answer is doing
The introduction is one sentence. It names the subject and gets out of the way.
The bullets are the spine, not a checklist. Skill, how, how long, why — followed in order, but nothing is announced. There is no “Firstly, I will tell you what the skill is.”
The evaluation bullet gets real time. The last paragraph is a third of the talk, and it is where the complex grammar sits: If I’d learned as a child, I doubt any of that would have happened. That is a third conditional produced naturally, which is worth more than any vocabulary list.
Tenses range without being forced. Past simple, past perfect, past continuous, a conditional and a present generalisation, because the story genuinely requires them.
The detail is specific and small. A friend’s triathlon, one evening a week, the least promising person in the group. None of it is impressive; all of it is concrete.
It ends on a reflection, not a summary. That’s the part nobody warns you about is a natural close, and it leaves the examiner with a complete thought rather than a trailing sentence.
Phrases worth having
Opening
- I’d like to talk about…
- The thing that immediately comes to mind is…
Moving through the talk
- What finally pushed me into it was…
- After a while I moved on to…
- Altogether it took me…
Reaching the evaluation
- As for why it mattered…
- Looking back on it now…
- If I’d known then what I know now…
Recovering when you lose the thread
- Sorry, where was I — the point I was making was…
- Anyway, the reason it stuck with me is…
That last pair matters. Losing your place and recovering in natural English is a band 7 behaviour. Losing your place and stopping is not.
What the examiner is marking
Part 2 leans hardest on Fluency and Coherence, because it is the only part testing sustained speech.
- Band 5 — “Usually maintains flow of speech but uses repetition, self correction and/or slow speech to keep going.”
- Band 6 — “Is willing to speak at length, though may lose coherence at times due to occasional repetition, self-correction or hesitation. Uses a range of connectives and discourse markers but not always appropriately.”
- Band 7 — “Speaks at length without noticeable effort or loss of coherence.”
The line that decides most Part 2 performances is willing to speak at length against speaks at length without noticeable effort. Both describe someone who fills the two minutes. Only one of them sounds comfortable doing it.
The hesitation that separates 6 from 7
Band 6 pauses are language-related — hunting mid-sentence for a word or a tense. Band 7 pauses are content-related — thinking of an idea, then delivering it fluently.
The examiner can hear the difference, because language hesitation lands in the middle of clauses and content hesitation lands between them.
What caps Part 2 at band 6
Stopping before one minute. A direct Fluency and Coherence penalty, and the easiest one to avoid.
Reading your notes. Audible, and it costs exactly the criterion the part exists to test.
Announcing the structure. Firstly I will describe… secondly I will explain… turns a story into a form.
Rushing the last bullet. Skipping the evaluation removes the only section that naturally produces band 7 grammar.
Memorised talks. Examiners are trained to spot them and will ignore rehearsed sections entirely when scoring — and the rounding-off question afterwards is unscripted, which is where prepared candidates come apart.
Practise it
Time yourself. Most people who think they are speaking for two minutes are speaking for seventy seconds, and the only way to find that out is a stopwatch.
Then practise the prep minute separately: read a cue card, write keywords only, and stop at sixty seconds whether you feel ready or not. Feeling unready is the normal condition of Part 2.
You can practise IELTS Speaking with AI band feedback on IELTS Studio, which marks all four criteria separately so you can see which one is capping your band.
Related: Speaking Part 1 covers the short interview questions, Speaking Part 3 covers the discussion that follows this topic, and how IELTS Speaking is scored explains the four criteria.
Frequently asked questions
- How long should you speak in IELTS Speaking Part 2?
- Between one and two minutes. Stopping before one minute is penalised under Fluency and Coherence. Aim to still be speaking when the examiner stops you at two minutes.
- Do you have to cover all the bullet points on the cue card?
- No. Task Response is not a marking criterion in IELTS Speaking, so the bullets are a suggested structure rather than a checklist. Use them to organise your talk, but do not panic if you miss one.
- What should you write during the one minute preparation?
- Keywords only — single words, collocations and one or two grammar structures you want to use. Writing full sentences means you will read them aloud, which sounds robotic and lowers your fluency score.
- What happens if the examiner stops you in IELTS Speaking Part 2?
- Nothing bad. Being interrupted at exactly two minutes means you sustained speech for the full long turn, which is what the part is testing. It is a good sign, not a penalty.
- What if I run out of things to say before two minutes?
- Move to the last bullet and explain why the topic matters to you, then speculate about the future. Evaluation and hypotheticals are open-ended in a way that description is not.
- Can I talk about something invented in Part 2?
- Yes. You are marked on English, not honesty, and no one verifies your story. A memorised answer is the real risk, because examiners are trained to spot rehearsed speech and ignore it when scoring.