2026 Guide: How to Organize PTE Retell Lecture Notes and Order to Get a Higher Score - 优秀PTE博客

When many people do the PTE Retell Lecture task, it’s not that they truly don’t understand it at all.

Instead, it's much more common: you nod while listening, thinking, "Yeah, I know what it's about." But the moment you speak, only a few scattered words come out—climate, research, education—like you're trying to dig through a junk drawer.

And you know it yourself. "Game over. I’m going to talk in fragments again."

I recently reviewed Pearson's official description of the Retell Lecture task, and I feel like its intent is actually quite direct. This task isn't asking you to repeat every sentence you hear back verbatim, nor is it asking you to use a generic "one-size-fits-all" template to fill 40 seconds. It is more about checking if you can grasp the main thread of the lecture and relay a few key points in the correct order.

So, this article isn't about that mystical "perfect oral phrasing" for a full score. I’m only going to talk about something more practical: why you know everything but speak chaotically, and exactly how to take notes and arrange the order.

Retell Lecture relies more on "Sense of Main Thread" than on "Vocabulary Stock"

Pearson's current official task description is quite clear. Retell Lecture is listening to a 90-second clip, getting 10 seconds to prepare after listening, and then speaking for 40 seconds.

This rhythm suggests one thing: you don't have time to translate the content completely into your head before outputting it. Anyone trying to do that will usually get stuck.

Moreover, the official scoring isn't just looking for how many keywords you can spit out. It checks if you can sequence the main content, actions, relationships, and development logic of the lecture. In other words, keywords are certainly important, but they are just the bricks, not the house.

Many students lose out right here.

While listening, they keep picking up bricks. Picking them excitedly.
But when it comes time to actually speak, they realize: Oh, the house hasn't been built.

Writing only words without their relationships makes the second half of your speech scatter

The official RL suggestion repeatedly mentions one point: record keywords and phrases, and mark the relationships between the information. This is crucial.

The most annoying part of lecture audio isn't that the words are hard; it's that it keeps making turns.

The beginning talks about a phenomenon.
The middle starts explaining the cause.
Then it adds an example or a result later.

If your notes only leave three nouns, when you open your mouth, you easily turn into:

The lecture talks about climate, research, students, and also data...

It sounds like you didn't completely go off-topic, but it's scattered and very empty.

A more stable way of note-taking isn't "writing more," but "writing less but showing relationships." For example:

topic -> bird migration
cause -> weather change
example -> northern areas
result -> travel timing changes

These notes might look like only a few lines, but when you speak, you know exactly what comes first and what follows. This is really important.

Grabbing the first layer of skeleton first is worth far more than frantically copying details

Pearson's official test technique PDF mentioned starting to write notes as soon as the audio starts, but don't just stare at single words—write them as phrases so you can understand what you wrote. This advice sounds simple, but it has already pointed out the pitfalls.

Many people get excited when they hear numbers, years, or names and furiously start writing. As a result, they write furiously for the first 20 seconds, but once the content changes, their brain doesn't catch up, and their hand is still rushing to complete previous words.

At this point, you aren't doing Retell Lecture; you are chasing a bus—or you’ve already lost it.

I personally suggest grabbing these four layers of skeleton first:

What the Theme actually is
Which core point the speaker mentioned first
Is it explanation, an example, or a result afterwards?
Is there a conclusion-heavy wrap-up at the end?

If you get this skeleton, even if a few details slip away, you can still speak.
But if you only cling to a string of words, the moment your mouth opens, they will slip away on the ground.

The 10-second preparation time is for ordering, not for making sentences on the fly

Many people think the 10s preparation time is huge, thinking they can secretly organize a mini-essay. It's actually not enough.

The most important things to do in these 10 seconds are not:

Finishing notes
Thinking about fancy expressions
Reciting a template

Just do these three things faithfully:

Glance at the theme word
Decide which two big points to cover first
Think about how to end the speech

That's it.

If you're still thinking about starting with something like This lecture is very interesting and informative at this moment, you're wasting time or trapping yourself in a template pit. Pearson's 2024 official article explicitly warned: Don't use memorized templates. If the system detects a large amount of pre-backed content, it will directly give 0 points, and this doesn't count towards fluency or pronunciation.

This isn't meant to scare you. It really is a huge loss.

Speaking order is more useful than fancy sentences

I found that many students have a habit when doing RL: they desperately look for "pretty sentences" the moment they open their mouth. But the first thing to preserve in this task isn't how fancy the sentences are, but that the order doesn't crumble.

A very usable order is actually like this:

First say the theme of the lecture
Then say the two or three main points the speaker covered
Seamlessly connect the causes, examples, and results
Finally, close it with an overall conclusion

For example, you can speak very simply:

The lecture is about ...

The speaker first explains ...

Then the lecture gives an example of ...

Finally, it shows that ...

Really, no need to act.

RL is a bit like moving boxes. It's more important for you to load the boxes onto the truck in order than to study what label is on the box.

Images and keywords are just aids; don't let them drive you crazy

The official task page also mentions that Retell Lecture sometimes comes with a picture. The picture is useful, but it’s not the main character.

Some students, seeing a picture, start talking around it, and the real focus of the audio ends up not being mentioned. For instance, if the picture is a map but the audio mainly talks about the cause and impact of migration, if you only grab the things on the picture, you sound like you are guessing the answer.

A safer approach is:

Use the picture only to confirm the topic
The actual content order still follows the audio
Prioritize writing down when you hear transition words and cause-and-effect relationships

Don't invert the main and the secondary.
The picture is the handrail.
It is not the steering wheel.

Speaking for 40 seconds doesn't mean stuffing everything in

Pearson's official article suggests trying to speak as close to 40 seconds as possible because a more complete narrative usually has a higher chance of hitting the correct content. This is true, but it easily leads to the misunderstanding that "I must cram everything in."

That’s not what it means.

Pearson also stated that it's impossible to fit every point of the lecture into just 40 seconds, so you must choose the most important content. This is a realistic balance:

Too short, content is insufficient
Too full, sentences start to get messy
Too greedy, repetition and stumbling appear easily

So a reasonable goal isn't "saying everything," but "don't lose the main thread, don't break the key points, and speak through to the end."

It sounds less "hot-blooded," but the score is usually more honest.

Heavy template flavor can cause rescuable scores to slip away too

Right now many RL templates online look very aggressive: one sentence for the intro, one for the transition, one for the conclusion, looking like they can be applied to any lecture.

But that’s the problem.

It’s not that they aren't "advanced," but that they are too tidy.

Tidy to the point of being fake.
It sounds less like you just heard a lecture, and more like you rehearsed that segment three times last night.

Pearson has already written down this risk, so I really don't recommend you turn RL practice into a template reading exercise. Especially those long, identical openings for every question: the benefit is low, but the risk isn't small.

What you should practice are these small but useful things:

Taking notes in phrases
Using arrows for relationships
Using 10 seconds to order your thoughts
Once you start speaking, don't stop because you can't remember a single word

These actions don't look cool, and are a bit "rustic." However, in the actual exam, these rustic methods often save you.

It is best to fix a feedback environment for RL practice

RL also has a annoying point: after you speak yourself, you often misjudge. You might feel like it went smoothly today, but when you listen to the recording, it's just going in circles. Or you feel you spoke "okay," but actually, the main thread is much steadier than a few days ago.

So I always feel that it's best to fix a platform with relatively continuous feedback for practicing this question. Otherwise, you brush a question here today and there tomorrow; the grading standards get mixed up, and you become more and more anxious.

If you want to practice speaking tasks like RA (Read Aloud), RS (Repeat Sentence), DI (Describe Image), and RL together, I would suggest you fix your practice on Youshow PTE. You can download it from the Apple App Store or visit their homepage directly: https://pte.youshowedu.com/en.

I like one aspect of it in particular: you can record constantly, view AI scores, and take mock exams on the go. This way, you are more likely to realize whether you can't note down what you hear after listening, or if you can note it down but can't speak through it clearly. Once the problem is clarified, you won't always be all over the place panicking.

Improving RL scores usually isn't "suddenly enlightened" but "stopping the fragments"

If your current real status doing PTE Retell Lecture is: listening is fine, but speaking is a mess of here and there, then first don't set a goal like "I want to be as fluent as a podcast host."

Lower the requirements first, and you are statistically more likely to see a score increase:

While listening, grab the topic and two or three main points first
Definitely leave relationship words in your notes
10s of prep is only for ordering
Once you start speaking, keep moving forward; don't stop for a small vocabulary

First achieve these.
Really first achieve these.

RL is often not that you don't know English, but that the output is too chaotic. Once you practice the order, the feeling of your score going up usually comes earlier than you think.