How is IELTS Speaking scored?

Speaking is the only part of IELTS where a person, not a marking key, produces your number. That changes what preparation is for. There is no bank of right answers to memorise and no raw score to push up; there is a set of published descriptions of what English at each band sounds like, and an examiner deciding which of those descriptions fits what you just said.

Four judgements, not one

The examiner is not forming a single impression and translating it into a band. They are making four separate judgements while you talk, one per criterion, and each is held against the published descriptors for that criterion alone.

That separation matters more than it sounds. Your vocabulary cannot rescue your grammar, and a strong grammatical range cannot cover for speech that keeps stalling. Because the criteria are equally weighted, the four judgements are simply averaged — so whichever one is lowest drags on the result exactly as hard as any other, no matter how strong the other three are.

Why your score is usually an average of unequal numbers

Most candidates are not level across all four. A very common shape is two criteria at one band and two at the band below, which averages to a half band in between — and that is where scores like 6.5 and 7.5 come from. It also means the fastest route up is almost never “speak better in general.” It is finding which single criterion is sitting a band behind and working on that one.

You can only do that if you know what the descriptors say. Reading the band descriptors criterion by criterion turns a vague target into four specific ones: keep going without audible searching, reach for precise words rather than approximate ones, produce complex sentences with nothing wrong in them, and stay easy to follow.

What the examiner is doing while you speak

The Speaking test is conducted face to face by a certificated examiner, and it is recorded. The examiner is running the interview and assessing you at the same time, across all three parts — so a weak opening minute is not a verdict, and a strong Part 2 does not settle anything either. They are listening for patterns across the whole eleven to fourteen minutes.

This is also why the examiner gives you nothing to read from their face. Their manner is part of delivering a standardised test, not feedback on how you are doing.

What to do with this

Record yourself answering real questions and score the recording four times, once per criterion, using the descriptor wording rather than a general sense of how it went. The criterion you find hardest to place is usually the one you understand least well — and that is the one worth attacking first.