The Five Speaking Errors We See Most
These are behaviours rather than language problems, which is why they persist in candidates whose English is otherwise strong. Each one triggers a ceiling, and each one can be removed without improving your English at all.
Why these five, and in this order
There is a category of speaking error that has nothing to do with how good your English is. It is a habit, it triggers a ceiling, and the ceiling is applied after everything else has been considered, which means eloquence cannot lift it. Candidates who work on vocabulary and fluency for weeks and see no movement in their band are almost always sitting under one of these.
They are listed here in the order worth fixing, which is roughly the order of how cheap the fix is against how much it costs you. The first three cost nothing but attention. The last two cost a little planning.
A note on how to use this. Read the five, then listen to your own recordings and mark which ones you actually do. Self-diagnosis by memory does not work here, because every one of these feels fine from the inside. Stopping at second twenty two feels like finishing. Re-describing feels like answering. Being even handed feels like being thoughtful.
One: stopping before the time does
The most common and most costly habit in the section. An answer that uses well under a third of the allowed recording time cannot score above 6 on Task Fulfilment, because the task has not been developed. On a sixty second task that threshold is around twenty seconds, which at a natural pace is about forty five words, or three sentences.
It happens because planning stops at one idea. A candidate reads the prompt, thinks of a good answer, and starts speaking. The good answer takes eighteen seconds, and then there is nothing behind it. What follows is usually either silence or thirty seconds of restating, and both of those do further damage, silence to Listenability and restating to Content.
The fix is entirely in the preparation window, and it is arithmetic rather than skill. Three anchors before you open your mouth. Not three good anchors, three anchors. A mediocre third point that fills fifteen seconds with content outscores an excellent single point that leaves forty seconds empty, because one of those answers triggers a cap and the other does not.
Two: answering the task before or after the one in front of you
This is specific to the picture pair. Task 3 asks you to describe the scene and Task 4 asks what happens next, and on the real test they use the same illustration back to back, which makes the two very easy to blur.
Re-describing the picture on Task 4 is the commonest single failure on any task in the section, and it caps Task Fulfilment at 6. From the inside it feels like answering, because you are talking accurately about the correct picture. The diagnostic is the verbs: present continuous verbs about visible things is description, modal and future verbs about invisible things is prediction.
The mirror image happens on Task 3, where candidates who run short start predicting to fill the time. That thins the description being scored and spends the material Task 4 will need in a minute. Both halves of this error come from the same root, which is arriving at a task without enough planned to fill it, so the fix is the same as the first error's.
Three: never asking for anything
Two tasks end in a request, and on both of them the request is the part most reliably dropped. Task 8's wrapper always closes with an ask, whether you should buy the thing, how to find the house, whether it can be repaired, and failing to make it caps Task Fulfilment at 8. Task 6 is a conversation whose purpose is to resolve a difficulty, and an answer that describes the situation for a full minute without proposing anything has not completed it.
The reason it is dropped is time management rather than ignorance. The description or the explanation is interesting, it expands to fill the space available, and the recording ends while the candidate is still mid-sentence on the fourth detail.
The fix is a written reminder and a glance at the clock. Write ASK at the bottom of your notes. At around second fifty on a sixty second task, look at it and land the request, even if the sentence you were in the middle of has to be abandoned. Ten seconds is more than enough. So, do you think it is worth thirty dollars, or should I leave it?
Listen to your own recordings with a stopwatch and a list of these five, and mark which you actually do. Every one of them feels fine from the inside, so memory is not a reliable way to diagnose them.
Four: surveying instead of deciding
Task 1 asks what somebody should do and Task 7 asks what you think. On both, presenting the considerations on each side and never landing on one caps Content and Coherence at 7.
This error is unusual in that it is more common in stronger candidates. Seeing both sides genuinely is a sign of comprehension, and the instinct to present the complexity is a good instinct in most contexts. It is the wrong instinct here, because the task asked a question and a survey of the possibilities does not answer it.
There is also a hidden time cost. An answer that spends forty seconds weighing has forty seconds left for content, which on a ninety second task is not enough for two developed reasons. Committing in the first sentence is both the stronger structure and the faster one: Honestly, I would tell her to take it, and now everything after that is support.
What you are allowed, and encouraged, to do is concede. Naming the strongest argument against your position and then answering it is a top band move. The difference between conceding and fence-sitting is whether you come back.
Five: ignoring who you are speaking to
Tone is a named factor of Task Fulfilment, and the ladder for that criterion is explicitly about audience: conveying, then adjusting for different audiences, then adapting to the situation, the purpose and your relationship with the listener. Tasks 1, 5, 6 and 8 all name a person, and an answer that would sound identical whoever it was addressed to has demonstrated the bottom rung.
It shows up in two directions. Business register with a relative, I would recommend that you evaluate the following options, and casual register with somebody senior, yeah, just ditch it. Both are mismatches and both are scored as such. The information is always in the first line of the prompt, which is where attention is lowest because candidates are hunting for the task.
A related version is drift. The register is right for the first twenty seconds and then dissolves as you concentrate on your content, so an answer that began as a phone call to your mother ends as a presentation. The one word fix is to write the person's name or role at the top of your notes and address them by name at least twice: once in the opening and once near the ask.
None of the five errors in this lesson requires better English to remove. That is the point of grouping them. Fix these first, and then find out what your English is actually worth.
Fix caps before you fix language. A ceiling is applied after everything else has been assessed, so a month of vocabulary work cannot lift a Task Fulfilment score that is capped at 6 because the answer ended at second twenty.
Practise it
Five questions on the behaviours that hold otherwise strong answers down.