I'm going to repeat myself more on LinkedIn

I’m going to start repeating myself more on LinkedIn.

Not by copying and pasting old posts. Not because I’ve run out of things to say. Almost the opposite.

I downloaded a year of my LinkedIn analytics this weekend and put the data through the same kind of process I’ve been building into Editorial Intelligence: preserve the evidence, look for patterns, separate the signal from the tempting but misleading number, and decide what should change as a result.

The export covered 10 August 2025 to 9 August 2026.

Across that period, my account generated:

• 62,209 impressions
• 19,325 members reached
• 758 engagements
• 3,195 followers at the end of the period

Those aren’t influencer numbers. That’s useful, because I’m not trying to turn this into a post about how to become a LinkedIn creator.

The interesting part was what happened when I stopped treating the biggest number as the most important one.

---

The highest-reach post was not the most useful signal

My highest-reach post in the export generated 5,778 impressions and 10 engagements.

Another post generated only 815 impressions but 33 engagements.

The same pattern appears at monthly level. July 2026 produced 10,712 impressions, the biggest month in the dataset, with 71 engagements and 22 new followers recorded in the export. November 2025 produced only 2,858 impressions, but 72 engagements and 39 new followers.

There are too many variables here to pretend this proves a causal relationship between a particular kind of post and follower growth. It doesn’t.

But it does prove something simpler that is easy to forget when looking at social analytics:

Reach and response are different outcomes.

A post can travel widely without creating much visible reaction. A smaller post can produce a much denser response. Neither is automatically better. They are doing different jobs.

That immediately changes the question I want to ask about Editorial Intelligence on LinkedIn.

Not:

How do I maximise impressions?

But:

What kind of publishing helps the right people discover, understand and respond to the ideas?

For Editorial Intelligence, qualified resonance matters more than raw reach.

I am not starting from zero

The audience data was useful for another reason.

The network is already weighted towards the kinds of people I want to learn from and write for. The export shows 35% of the audience at senior level, 19% at director level, 7% owners and 6% CXOs. Founder is the largest listed job title at 7%.

The leading industries include software development, advertising, PR and communications, IT consulting, technology, marketing and business consulting.

There is also an interesting split by company size: 19% of the audience works in organisations with more than 10,000 employees, while 16% works in organisations of two to ten people and another 14% in organisations of 11 to 50.

That is unusually useful for Editorial Intelligence.

The enterprise version of the problem is often: we have enormous amounts of knowledge and evidence, but struggle to connect and activate it.

The founder or small-company version can be: I have expertise and ideas, but haven’t turned them into a recognisable narrative or repeatable system.

Different scale. Similar underlying problem.

So I do not need to build an entirely new Editorial Intelligence audience from scratch.

I need to help an existing audience understand what Editorial Intelligence is, one useful idea at a time.

And that is where repetition becomes important.

An idea is not exhausted because I posted it once

Content calendars can create a strange pressure towards novelty.

You publish an idea on Tuesday and mentally mark it as used. Wednesday requires another idea. The result is a production system that rewards constant invention even when the audience may never have seen Tuesday’s post in the first place.

That makes little sense for a publication.

Publications return to important arguments constantly. They revisit them when there is new evidence, when a different example makes the idea clearer, when events change the context or when the audience has changed.

Editorial Intelligence now has a growing bank of frameworks, essays, field notes, research questions, examples and observations. The constraint is no longer finding enough things to say.

The constraint is deciding which ideas deserve another useful expression.

A strong idea can become:

• a short observation
• a personal story
• a response to a current event
• a diagram
• a deeper article
• a counterargument
• an application for founders, marketers or leadership teams
• a research question
• a later post with better evidence

That is not repurposing in the sense of squeezing five assets out of one piece of content.

It is developing the same understanding from different directions.

The unit of reuse is the idea, not the wording.

LinkedIn can become part of the Evidence Engine

This is the part I find most interesting.

I have mostly thought of LinkedIn as a distribution channel: publish something, hope people see it, perhaps send them to the website.

That is still part of the job. But it is incomplete.

If Editorial Intelligence is about capturing evidence, shaping narratives and learning from what happens next, publishing itself can generate evidence.

The loop becomes:

Field note → idea → post → audience signal → learning → stronger idea

A post is therefore not necessarily the end product of the thinking.

It can be a probe.

Did people understand the distinction? Did a practitioner disagree with the premise? Did the personal example create a way into an otherwise abstract idea? Did a founder respond differently from someone working inside a large organisation? Did a post get enormous reach but almost no meaningful response? Did somebody send a message that revealed a problem I had not considered?

Those signals should not dictate the work. Algorithms are very good at rewarding things that have little to do with building a useful body of knowledge.

But ignoring the response would be equally odd.

If I argue that organisations should learn from the evidence their work creates, Editorial Intelligence should be prepared to do the same thing to itself.

AI is what makes the loop operationally viable

There is another reason I can work this way now.

I wrote recently that AI didn’t create Editorial Intelligence; it made it economically viable.

This is a small example of what I mean.

A year of LinkedIn data can now be exported, cleaned, compared and interrogated quickly. The findings can be checked against an existing content library and publishing strategy. The useful evidence can be preserved in the same knowledge system that contains the ideas themselves. Then the next piece can be developed from that evidence without starting again from a blank page.

Before generative AI, all of those steps were possible. A technically capable person could certainly analyse an Excel export. A developer could build a database. A content strategist could maintain a calendar. A writer could revisit old ideas.

The problem was the accumulated friction of moving between all of those modes of work.

AI reduces the cost of crossing those boundaries.

That matters more to me than asking it to write another social post.

The judgement still sits elsewhere: what result is meaningful, what is noise, which idea is worth repeating, when a high-performing topic would pull the work in the wrong direction, and what Editorial Intelligence should become.

Analytics are evidence for editorial judgement, not a replacement for it.

What changes from here

I am turning the last year into a baseline rather than a formula.

That distinction matters because the historical dataset is not an Editorial Intelligence dataset. It includes work posts, personal posts, football, politics, travel, technology and content-industry commentary. It tells me about the audience and the behaviour of the account, not exactly how Editorial Intelligence performs.

From August 2026, I can start building that evidence deliberately.

The working rules are simple.

Publish frequently, without requiring daily invention. There is already enough material in the knowledge bank to support a regular publishing rhythm. Some days should develop something new. Others should surface something worth returning to.

Repeat ideas, not posts. If an argument matters, it should survive more than one expression.

Use the world as a way into the discipline. A post about immigration, climate, AI, football, leadership or creator culture can still be an Editorial Intelligence post if it genuinely reveals something about evidence, language, narrative or sense-making.

Review performance in batches. One post is noisy. Ten or twenty posts start to create something worth comparing.

Separate different kinds of success. Reach, response, relationships, website discovery and strategic value are not the same metric.

And most importantly:

Return the learning to the system.

If a month of publishing teaches me something but nothing changes in the knowledge base, the loop is still broken.

The first subject for Editorial Intelligence is Editorial Intelligence

I’ve spent the past few months arguing that organisations discard too much of the evidence they create.

A research project becomes a report. A customer interview becomes a quote. An event becomes a recap. A campaign ends and everybody starts again.

Social publishing can suffer from exactly the same problem.

A post gets published. It performs well or badly. Someone glances at the number. Then tomorrow needs another post.

I don’t want to work that way.

The more interesting model is cumulative:

knowledge bank → select → adapt → publish → observe → learn → return the learning to the bank

That means I will probably post more often.

It also means I will deliberately repeat myself more.

Because if an idea is worth building a discipline around, saying it once was probably never enough.