IELTS Speaking Band Descriptors Explained: Fluency to Pronunciation

IELTS Speaking is scored on four criteria, each worth 25% of your Speaking band: Fluency and Coherence (can you keep going, and does it hang together), Lexical Resource (how wide and how precise your vocabulary is), Grammatical Range and Accuracy (how varied your structures are and how many sentences are error-free) and Pronunciation (how easily you are understood and whether you use stress and intonation to carry meaning). The examiner gives you a band for each criterion and the four are averaged. Understanding what each criterion rewards is the fastest way to stop practising the wrong thing.

How examiners use the descriptors

Examiners are trained to match what they hear to the published descriptions of each band; they do not count errors or tick off vocabulary items. Two consequences follow. First, the criteria are independent: a candidate can genuinely be band 8 for Pronunciation and band 6 for Grammatical Range and Accuracy, and the average lands at 7. Second, a weakness in one criterion holds the whole component back in a way that strength elsewhere cannot fully offset. You can test this for yourself with the band score calculator: raising one criterion by a band moves the Speaking average by only a quarter.

Below, each criterion is explained in plain language, with a qualitative description of bands 6, 7 and 8 and one drill.

1. Fluency and Coherence

This criterion has two halves. Fluency is about flow: how long you can speak, how often you pause, whether you hesitate, repeat yourself or self-correct. Coherence is about organisation: whether your ideas connect logically, whether you use linking words appropriately, whether you stay on topic and develop a point rather than listing several.

• Band 6: You are willing to speak at length, but coherence suffers from repetition, self-correction or hesitation. Linking words are present but not always used appropriately, and the same few ("and", "but", "so", "because") carry most of the load.
• Band 7: You speak at length without obvious effort. Hesitation is occasional and mostly about what to say, not how to say it. You use a range of discourse markers and linkers with some flexibility, though with some overuse.
• Band 8: You speak fluently with only rare, content-related hesitation. Topics are developed coherently: ideas are expanded, sequenced and connected naturally, and the listener never has to reconstruct the logic.

The most common misunderstanding is that fluency means speed. Examiners distinguish between pausing to think about an idea, which is natural at every band, and pausing to search for a word or a structure, which is what costs you. Fillers such as "um" are only a problem when they are frequent.

Drill: the two-minute expansion. Take any Part 1 question, set a two-minute timer and answer it, then keep extending in a fixed sequence: answer, reason, example, contrast, feeling, consequence. Record it, mark every pause longer than two seconds on playback, and decide whether it was an idea pause or a language pause. Repeat the same question until the language pauses disappear. The goal is not to talk for two minutes in the real test; it is to make the first thirty seconds of any answer effortless.

2. Lexical Resource

Lexical Resource measures the range of your vocabulary, its precision, your ability to paraphrase when a word escapes you, your control of collocation (the words that naturally go together), and your use of less common and idiomatic language in the right context. It is not a test of rare words; it is a test of the right word in the right place.

• Band 6: You have enough vocabulary to discuss topics at length and your meaning is clear, but word choice is sometimes inappropriate, and paraphrasing works but is visibly effortful.
• Band 7: You use vocabulary flexibly across a range of topics and include some less common and idiomatic items with awareness of style and collocation, with only occasional inaccuracies. Paraphrase is effective.
• Band 8: You draw on a wide range readily to convey precise meaning. Less common and idiomatic language is used skilfully throughout, with only rare slips.

The gap between 6 and 7 is usually collocation and precision: "make a decision" not "take a decision", "heavy traffic", "bitterly cold", "a steep learning curve". The gap between 7 and 8 is consistency: the less common language appears throughout, not in one rehearsed phrase, and almost nothing is misused. A wrongly used idiom costs more than a plain phrase would have.

Drill: topic collocation banks. Take the twelve topics that come up most often in Parts 1 and 2 (work, study, hometown, technology, environment, health, travel, food, family, free time, media, money). For each one, build a bank of ten collocations and three idiomatic expressions that you have checked in a learner's dictionary. Then record a sixty-second answer on the topic that uses five of them. Keep an expression only if you would comfortably say it to a colleague; if it sounds odd to you, it will sound odd to the examiner.

3. Grammatical Range and Accuracy

This criterion looks at the variety of structures you use (tenses, conditionals, passives, relative clauses, comparatives, modals, reported speech), how often you build complex sentences, and what proportion of your sentences are free of error. It also considers whether errors cause any misunderstanding.

• Band 6: You mix simple and complex structures but with limited flexibility, and errors are frequent in the complex ones, although they rarely cause misunderstanding.
• Band 7: You use a range of complex structures with some flexibility, produce frequent error-free sentences, and have a few persistent errors.
• Band 8: You use a wide range of structures flexibly, the majority of sentences are error-free, and the remaining mistakes are occasional slips rather than systematic problems.

Band 7 needs both levers, range and accuracy. Candidates stuck at 6.5 usually either attempt complex structures and get most wrong, or play safe with simple sentences and get everything right; the target is to attempt the complex structures and get the majority right. Persistent small errors (articles, third-person -s, subject-verb agreement, countable and uncountable nouns) are the first to fix because they recur in every sentence.

Drill: structure targets with transcription. Record a Part 2 answer and transcribe it word for word. It takes ten minutes and is the most revealing exercise in speaking practice. Tick every error-free sentence and work out the proportion. Then note which structures appeared. On the next attempt at the same cue card, aim for one of each: a past perfect, a conditional, a relative clause, a passive and a comparative, while holding the error-free percentage steady. Track the two numbers week by week.

4. Pronunciation

Pronunciation is not accent. The examiner is listening for three things: intelligibility (can you be understood without effort), control of individual sounds (vowels, consonants, consonant clusters, word endings), and the features of connected speech: word stress, sentence stress, rhythm, chunking and intonation. Candidates who speak clearly but flatly are typically held at band 6 on this criterion, because the descriptors reward using those features, not just avoiding errors.

• Band 6: You use a range of pronunciation features with mixed control. There is some effective use of stress and intonation, but it is not sustained, and individual words or sounds occasionally reduce clarity.
• Band 7: Typically described as showing all the positive features of band 6 consistently plus some of band 8: stress and chunking are mostly sustained, with lapses.
• Band 8: You use a wide range of features flexibly and sustain them throughout, you are easy to understand from start to finish, and your first-language accent has minimal effect on intelligibility.

Trying to "lose" your accent is wasted effort. Work instead on chunking speech into sense groups, stressing the words that carry meaning, and letting intonation fall at the end of statements and rise before a contrast.

Drill: shadowing with chunking. Choose a sixty-second clip of a clear speaker. Transcribe it, then mark a slash wherever the speaker pauses and underline the stressed word in each chunk. Play one chunk, pause, and repeat it matching the rhythm and stress exactly. When you can do every chunk, play the whole clip and speak over it simultaneously. Ten minutes a day produces an audible change within a couple of weeks.

Putting the four criteria together

Criterion	What moves you from 6 to 7	What moves you from 7 to 8
Fluency and Coherence	Fewer language pauses; linkers used naturally, not mechanically	Developed ideas; hesitation only ever about content
Lexical Resource	Accurate collocation; some less common vocabulary	Precise, idiomatic language sustained across topics
Grammatical Range and Accuracy	Frequent error-free sentences while using complex forms	Majority error-free; wide, flexible range
Pronunciation	Consistent stress and chunking; clear sounds	Features sustained throughout; effortless to follow

The four criteria are averaged to give your Speaking band, which is averaged with the other three skills for your overall score. Only the component band is reported, never the four criterion scores, so your results cannot tell you which criterion is holding you back. That is why self-study plateaus: you need a trained ear to say whether the problem is grammar or pronunciation before you can fix it.

Getting a criterion-by-criterion diagnosis

Fleydo's one-to-one IELTS course is taught by native English-speaking teachers who are Band 8.5+ certified IELTS specialists and who mark your speaking against all four criteria after every mock. The free trial lesson includes an honest band estimate for each criterion, so your first study plan is built around the one that is actually costing you points.