TRANSCRIPT 🔗
WEBVTT
00:00:59.339 --> 00:01:04.620
There's an enormous amount of advice about how to get attention on LinkedIn.
00:01:04.859 --> 00:01:06.379
Write a better hook.
00:01:06.700 --> 00:01:07.579
Use video.
00:01:08.219 --> 00:01:09.420
Don't use video.
00:01:10.060 --> 00:01:11.900
Tell personal stories.
00:01:12.299 --> 00:01:14.140
Put text on your images.
00:01:14.379 --> 00:01:16.620
Keep the text off your images.
00:01:17.019 --> 00:01:18.460
Write shorter posts.
00:01:18.700 --> 00:01:20.140
Write longer posts.
00:01:20.379 --> 00:01:22.299
Post at a particular time.
00:01:22.539 --> 00:01:25.340
Don't worry about the time because it doesn't matter.
00:01:25.579 --> 00:01:28.859
And of course, there's the one we hear all the time.
00:01:29.180 --> 00:01:31.019
Stop the scroll.
00:01:31.500 --> 00:01:37.340
You need something that will make people stop scrolling and pay attention to your content.
00:01:37.579 --> 00:01:42.620
But I've recently started wondering whether we're asking the wrong question.
00:01:42.939 --> 00:01:49.019
Because what if nobody actually knows what will make someone stop scrolling?
00:01:49.500 --> 00:01:53.500
Now I'm not suggesting all advice is useless, far from it.
00:01:53.659 --> 00:01:58.859
Some of it is based on good experience, some is supported by data.
00:01:59.019 --> 00:02:05.500
And there are certainly things we can do to make content more noticeable, accessible, and appealing.
00:02:05.740 --> 00:02:17.900
But there's quite a difference between saying something might increase the likelihood of attracting attention and saying we know what will make another human being stop.
00:02:18.379 --> 00:02:22.300
And I'm increasingly interested in that distinction.
00:02:23.419 --> 00:02:28.060
Part of the reason is that something fairly fundamental has changed.
00:02:28.300 --> 00:02:39.900
Towards the end of 2025, LinkedIn itself said that the volume of content being created on the platform every day had grown rapidly over the preceding year.
00:02:40.139 --> 00:02:41.979
Their words, not mine.
00:02:42.460 --> 00:02:48.620
And the consequence, according to LinkedIn, is more competition for attention.
00:02:49.259 --> 00:02:53.180
Now we need to be careful about what we conclude from that.
00:02:53.419 --> 00:03:05.580
LinkedIn didn't tell us in that particular announcement exactly how much content had increased, nor did it say generative AI was responsible for the increase.
00:03:05.819 --> 00:03:11.979
But I don't think we can ignore AI as part of the environment in which this is happening.
00:03:12.379 --> 00:03:18.379
Producing a competent LinkedIn post has become extraordinarily easy.
00:03:18.699 --> 00:03:26.859
Give an AI tool a topic and a few instructions, and within seconds you can have a perfectly serviceable post.
00:03:27.340 --> 00:03:31.579
It can have a hook, a story, three lessons and a conclusion.
00:03:31.900 --> 00:03:35.739
It can even have that obligatory question at the end.
00:03:36.379 --> 00:03:40.780
That doesn't make AI-generated content inherently bad.
00:03:41.099 --> 00:03:43.579
I use AI extensively myself.
00:03:43.900 --> 00:03:49.579
For me, the much more interesting question is what we're using it to do.
00:03:50.459 --> 00:03:58.299
Because when competent content becomes abundant, competence itself becomes less distinctive.
00:03:58.859 --> 00:04:04.620
For years we've been telling people that the answer is to create useful content consistently.
00:04:04.859 --> 00:04:08.379
Now I don't think that's wrong, but I think it's incomplete.
00:04:08.699 --> 00:04:21.579
There is now so much content, much of it perfectly competent, that producing another competent post doesn't guarantee anybody will give it their attention.
00:04:22.060 --> 00:04:23.740
So what does?
00:04:24.139 --> 00:04:31.980
Well, rather than immediately looking for another expert to tell me the answer, I decided to watch what I was doing myself.
00:04:32.540 --> 00:04:38.139
For a couple of days I paid a much closer attention to my own behaviour on LinkedIn.
00:04:38.379 --> 00:04:39.660
What made me stop?
00:04:39.900 --> 00:04:42.220
What did I scroll straight past?
00:04:42.620 --> 00:04:46.220
What made me start reading and then abandon something?
00:04:46.459 --> 00:04:49.900
And what made me think I'm glad I stopped for that?
00:04:50.540 --> 00:04:53.900
I expected to see some patterns, and I did.
00:04:54.139 --> 00:04:58.540
The trouble was they weren't nearly as tidy as I'd expected.
00:04:58.860 --> 00:05:08.459
One thing that stopped me was a video from Roy Kowarski, someone I know through the disruptive business leadership livestream I helped co-host.
00:05:08.860 --> 00:05:12.620
Roy had transformed himself into James Bond.
00:05:12.860 --> 00:05:17.180
And I don't mean he'd put on a dinner jacket and stood in front of a camera.
00:05:17.500 --> 00:05:23.180
His video was clever, imaginative, and extremely well executed.
00:05:23.420 --> 00:05:33.500
He'd done something similar with another famous film character the previous week, and this was sufficiently unusual that I wanted to see what he'd done this time.
00:05:33.740 --> 00:05:36.379
So perhaps novelty stops me.
00:05:36.620 --> 00:05:39.100
Except that wasn't the whole story.
00:05:39.340 --> 00:05:45.740
I noticed that I stopped for images containing very large, immediately legible text.
00:05:45.980 --> 00:05:51.100
Sometimes, in fact, I read the words on the image before I read the headline of the post.
00:05:51.900 --> 00:05:55.100
Personal stories caught my attention too.
00:05:55.420 --> 00:05:58.220
So did posts from people I knew.
00:05:58.460 --> 00:06:05.180
And I became particularly aware that I was more inclined to notice somebody if we'd interacted recently.
00:06:05.420 --> 00:06:10.060
Perhaps I'd commented on something I'd posted or would had a conversation.
00:06:11.100 --> 00:06:15.260
Traditional copywriting headlines still caught me as well.
Load full transcript
00:06:15.580 --> 00:06:20.300
Five mistakes, three myths, what I learned from.
00:06:20.939 --> 00:06:23.819
None of those is remotely new.
00:06:24.379 --> 00:06:27.500
And then there were text-only posts.
00:06:27.740 --> 00:06:32.780
No image, no video, no obvious visual scroll stopper.
00:06:32.939 --> 00:06:34.700
Some of those caught me too.
00:06:35.020 --> 00:06:41.819
At which point my nice neat theory about what stops the scroll had rather fallen apart.
00:06:42.620 --> 00:06:46.220
But that was when the exercise became much more interesting.
00:06:46.460 --> 00:06:49.900
There wasn't one thing making me stop.
00:06:50.300 --> 00:06:57.580
Sometimes it was the subject, sometimes the presentation, sometimes novelty, sometimes simply the person.
00:06:58.300 --> 00:07:09.260
And presumably there were occasions when an excellent post went straight past me because I was busy, distracted, or simply didn't feel like reading anything else right then.
00:07:09.900 --> 00:07:11.340
And that matters.
00:07:11.660 --> 00:07:17.500
Because we tend to discuss attention as though it's something entirely controlled by the creator.
00:07:18.140 --> 00:07:20.860
Write the right hook and I will stop.
00:07:21.260 --> 00:07:23.980
Use the right image and I will stop.
00:07:24.540 --> 00:07:28.060
Make the video short enough and I'll watch it.
00:07:28.379 --> 00:07:31.900
But I'm part of this transaction too.
00:07:32.460 --> 00:07:44.780
My workload matters, my interests matter, what I've already seen that day matters, my relationship with you matters, even my mood may matter.
00:07:45.020 --> 00:07:49.660
And if my own attention is this variable, yours will be too.
00:07:50.140 --> 00:07:58.700
So I think we need to be wary when anybody tells us there is one particular formula that will stop people scrolling.
00:07:59.100 --> 00:08:02.379
There may be tactics that improve our chances.
00:08:02.860 --> 00:08:08.220
There isn't a button we can press that makes another human being pay attention.
00:08:08.460 --> 00:08:11.900
And here's where things become even more interesting.
00:08:12.300 --> 00:08:19.819
LinkedIn itself appears to recognize precisely how individual relevance is.
00:08:20.300 --> 00:08:26.620
The platform isn't trying to construct one perfect feed that everybody sees, quite the opposite.
00:08:26.939 --> 00:08:34.700
LinkedIn says it uses hundreds of signals to work out what might be relevant to each individual member.
00:08:35.259 --> 00:08:42.139
That can include the subjects we're interested in, the people we interact with, and what we've engaged with previously.
00:08:42.299 --> 00:08:44.779
It can even include what we scroll past.
00:08:45.259 --> 00:09:01.740
And in 2026, LinkedIn introduced a much more sophisticated feed system using large language models to understand what individual posts are actually about and how closely they relate to our evolving professional interests.
00:09:02.139 --> 00:09:04.220
Think about that for a moment.
00:09:04.779 --> 00:09:13.899
LinkedIn is investing enormous technical resources trying to understand relevance at an individual level.
00:09:15.259 --> 00:09:27.500
And yet we're sometimes offered content advice that suggests millions of different human beings can all be persuaded to stop by the same handful of techniques.
00:09:27.980 --> 00:09:32.620
Now I don't think those two ideas sit particularly comfortably together.
00:09:32.860 --> 00:09:35.419
It doesn't mean hooks don't matter.
00:09:35.659 --> 00:09:38.379
It doesn't mean imagery doesn't matter.
00:09:38.620 --> 00:09:47.179
It doesn't mean video, storytelling, headlines, formatting or any of the other things we talk about don't matter.
00:09:47.579 --> 00:09:51.980
Any of them might affect whether somebody notices something.
00:09:52.939 --> 00:10:05.259
What I'm questioning is the certainty, because I can't tell you which one will make your audience stop, and neither, I suspect, can anybody else.
00:10:05.659 --> 00:10:07.899
Which raises another question.
00:10:08.299 --> 00:10:14.059
Perhaps stopping the scroll isn't the most useful objective in the first place.
00:10:16.379 --> 00:10:20.779
There was something else I noticed while I was doing this exercise.
00:10:21.179 --> 00:10:28.219
Some things are extremely good at attracting attention without necessarily giving me much in return for it.
00:10:28.379 --> 00:10:29.899
We all know the sort of thing.
00:10:30.059 --> 00:10:36.379
A piece of celebrity news that has nothing obvious to do with the person's professional expertise.
00:10:36.860 --> 00:10:42.219
A viral video that's already accumulated millions of views elsewhere.
00:10:42.539 --> 00:10:49.019
A photograph that catches the eye but bears no discernible relationship to the subject of the post.
00:10:49.500 --> 00:10:57.099
An outrageous promise about how much money somebody made in an implausibly short period of time.
00:10:57.339 --> 00:10:59.259
And here's the complication.
00:10:59.500 --> 00:11:01.500
Those things might work.
00:11:01.659 --> 00:11:03.819
Some of them work spectacularly.
00:11:04.059 --> 00:11:10.299
I know people who can generate extraordinary numbers of views using precisely this kind of content.
00:11:10.539 --> 00:11:15.339
So I'm not going to argue that it doesn't work simply because I don't particularly like it.
00:11:15.659 --> 00:11:23.740
If your objective is to accumulate views, attracting attention that way may achieve exactly what you set out to achieve.
00:11:23.980 --> 00:11:27.099
But that's not the question I'm interested in.
00:11:27.899 --> 00:11:31.099
I'm interested in professional visibility.
00:11:31.579 --> 00:11:36.059
And professional visibility asks something more demanding.
00:11:36.379 --> 00:11:40.459
If I give you my attention, what are you going to do with it?
00:11:40.699 --> 00:11:45.419
What will I know after reading your post that I didn't know before?
00:11:45.740 --> 00:11:48.379
What will I understand
differently?
00:11:48.860 --> 00:11:58.299
Perhaps you've taught me something, perhaps you've shown me something, maybe you've made an unexpected connection or challenged something I assumed was true.
00:11:58.620 --> 00:12:02.860
Perhaps you've given me access to an experience I haven't had.
00:12:03.099 --> 00:12:09.019
Or perhaps you've simply entertained me through genuinely exceptional creativity.
00:12:09.899 --> 00:12:15.099
There are lots of perfectly legitimate ways to reward someone's attention.
00:12:15.339 --> 00:12:18.779
But there's another question underneath all of those.
00:12:19.179 --> 00:12:23.740
What, if anything, do I now understand about you?
00:12:24.939 --> 00:12:31.980
Because stopping somebody's scrolling is not the same thing as becoming known for something.
00:12:32.219 --> 00:12:35.179
And that's the distinction I think matters.
00:12:35.819 --> 00:12:40.539
This is where the conversation about AI becomes much more interesting to me.
00:12:40.779 --> 00:12:46.539
I'm not particularly interested in trying to divide LinkedIn to AI content and human content.
00:12:46.699 --> 00:12:50.860
I think that's becoming an increasingly meaningless distinction.
00:12:51.259 --> 00:13:02.139
AI can help us research, it can organize ideas, challenge our thinking, improve our writing, create imagery, and help us express things more clearly.
00:13:02.459 --> 00:13:06.299
The question is, what are we putting into that process?
00:13:06.699 --> 00:13:09.899
What experience do you have that I don't?
00:13:10.139 --> 00:13:12.139
What evidence have you gathered?
00:13:12.379 --> 00:13:13.980
What have you observed?
00:13:14.459 --> 00:13:19.099
What judgment have you developed after years of working in your field?
00:13:19.659 --> 00:13:21.819
What access do you have?
00:13:22.059 --> 00:13:25.500
What relationships have informed your thinking?
00:13:26.139 --> 00:13:36.939
What interpretation can you offer that I won't get simply by asking an AI tool to produce another post on the same subject?
00:13:37.259 --> 00:13:40.379
That's the scarce ingredient.
00:13:40.699 --> 00:13:43.819
And I don't necessarily mean originality.
00:13:44.539 --> 00:13:50.860
Something happened while I was discussing this subject on LinkedIn recently that changed my thinking about it.
00:13:51.259 --> 00:14:02.779
I've been talking about the importance of producing something that wasn't easily replicated, and one of my connections left a comment that made me reconsider it.
00:14:03.419 --> 00:14:11.419
His point, in essence, was that something can be completely original and still not deserve his attention.
00:14:11.740 --> 00:14:15.099
It first has to matter to him.
00:14:15.819 --> 00:14:19.500
And he's right, being different isn't enough.
00:14:19.740 --> 00:14:21.899
Being original isn't enough.
00:14:22.219 --> 00:14:23.980
It has to be relevant.
00:14:24.219 --> 00:14:27.740
And then you have to contribute something worth receiving.
00:14:28.059 --> 00:14:31.899
So perhaps contribution is a better word than originality.
00:14:32.860 --> 00:14:36.059
What are you bringing to the conversation?
00:14:36.539 --> 00:14:47.740
Because if competent content is becoming easier and easier to produce, the difficult to replicate part may increasingly be the person behind it.
00:14:47.980 --> 00:14:50.459
And that leads us somewhere else.
00:14:50.860 --> 00:14:53.419
Suppose your post does work.
00:14:53.980 --> 00:15:01.099
Somebody stops, they read it, and something you've said makes them think, that's interesting, who is this person?
00:15:01.899 --> 00:15:03.659
What happens next?
00:15:03.980 --> 00:15:09.500
Because increasingly LinkedIn is making it easier for them to find out.
00:15:09.899 --> 00:15:12.459
They may click through to your profile.
00:15:12.860 --> 00:15:20.219
They may encounter information about you through some of the newer profile previews LinkedIn's been introducing.
00:15:20.620 --> 00:15:27.899
They may see your headline, your experience or other signals that tell them who you are and what you know.
00:15:28.379 --> 00:15:32.139
So your content isn't operating in isolation.
00:15:33.339 --> 00:15:40.939
If the post creates curiosity, the rest of your LinkedIn presence has to substantiate it.
00:15:41.259 --> 00:15:43.339
Can I understand what you do?
00:15:43.579 --> 00:15:46.539
Can I see evidence of your expertise?
00:15:46.860 --> 00:15:51.339
Does your experience support the authority with which you're speaking?
00:15:51.659 --> 00:15:56.379
Is there enough substance there for me to decide you're somebody worth remembering?
00:15:56.620 --> 00:15:59.419
Perhaps even somebody worth talking to.
00:16:22.459 --> 00:16:29.579
Because professional visibility isn't created by one brilliant piece of content, it's accumulated.
00:16:29.980 --> 00:16:33.819
One encounter gives me a reason to notice you.
00:16:34.139 --> 00:16:37.659
Another gives me a reason to remember you.
00:16:38.139 --> 00:16:48.219
Over time I begin to associate you with a subject, an expertise, a perspective, or perhaps simply a way of thinking that I find valuable.
00:16:48.539 --> 00:16:53.339
And the next time you appear in my feed, perhaps I'm more likely to stop.
00:16:53.899 --> 00:16:59.980
Not because you've discovered the perfect hook, but because now I know who you are.
00:17:01.339 --> 00:17:07.420
Normally this would be the point where I'd pull everything together and give you five things to do next.
00:17:07.660 --> 00:17:09.579
But I'm not going to do that.
00:17:10.299 --> 00:17:13.660
Because I don't think we know enough to do that.
00:17:13.980 --> 00:17:17.420
And I'm increasingly comfortable saying so.
00:17:18.059 --> 00:17:21.099
Instead, I want to suggest an experiment.
00:17:21.579 --> 00:17:26.059
For the next day or two, watch yourself using LinkedIn.
00:17:26.220 --> 00:17:28.539
Don't change anything, just notice.
00:17:28.779 --> 00:17:30.220
What makes you stop?
00:17:30.779 --> 00:17:33.339
What do you scroll straight past?
00:17:34.140 --> 00:17:37.980
When does the identity of the author make a difference?
00:17:38.299 --> 00:17:41.740
What catches your eye but then disappoints you?
00:17:42.140 --> 00:17:46.860
What looks completely unremarkable but turns out to be worth reading?
00:17:47.099 --> 00:17:52.620
And when you do scroll past something that on another day you might have read.
00:17:52.940 --> 00:17:59.900
Your answers won't tell you what everybody else wants, and that's precisely the point.
00:18:00.220 --> 00:18:07.339
They'll show you how complicated the attention of just one human being can be.
00:18:07.980 --> 00:18:10.299
And then turn the exercise around.
00:18:10.539 --> 00:18:12.460
Look at your own content.
00:18:13.259 --> 00:18:17.099
Not with the question, how do I stop the scroll?
00:18:17.339 --> 00:18:18.700
Try different ones.
00:18:19.019 --> 00:18:20.619
Like why this?
00:18:21.019 --> 00:18:21.899
Why you?
00:18:22.299 --> 00:18:23.099
Why now?
00:18:23.500 --> 00:18:31.659
And perhaps the most important question of all, if somebody gives you their attention, what are you giving them in return?
00:18:32.779 --> 00:18:38.859
I don't think anybody has discovered the universal formula for earning attention on LinkedIn.
00:18:39.019 --> 00:18:40.859
I'm not convinced there is one.
00:18:41.099 --> 00:18:44.700
But perhaps that's the wrong thing to be searching for anyway.
00:18:45.180 --> 00:18:47.899
We can't guarantee attention.
00:18:48.220 --> 00:18:51.099
We can't dictate relevance.
00:18:51.339 --> 00:19:00.700
And we certainly can't control the state of mind of the person whose thumb happens to reach our post in the feed.
00:19:01.099 --> 00:19:06.299
What we can control is what we've put there for them to find.
00:19:06.619 --> 00:19:15.099
So perhaps the better ambition isn't to become better at stopping people, it's to become more worth stopping for.