Emilia's unfair advantage is that she teaches one distinction over and over
Almost every post makes the same move: here is a thing you think you understand, and here is the part you are missing.
Emilia Möller works on growth at Searchable. Across the 97 posts we analyzed, published between February and July 2026, she posted 4.3 times a week and averaged 577 reactions. 96 of those 97 posts are images, almost all of them carousels. Her median post is 397 words, by far the longest in this series. Not one post landed on a Saturday or a Sunday.
The content is narrower than the volume suggests. Distinction teaching is when you build an audience by repeatedly separating two things your reader believes are the same, so every post delivers a correction rather than an opinion. SEO is not AEO. AEO is not GEO. SEO is not rankings. Modern marketing is not posting online. Her top post, at 2,052 reactions, is titled 'SEO, AEO and GEO are NOT the same thing.'
'5 SEO tips for 2026.' Useful, forgettable, and indistinguishable from a hundred others.
'Common misconception: SEO = rankings.' The reader now has something to be wrong about.
“SEO might seem simple. But simple doesn't mean easy.”
— Opening a post that earned 2,039 reactions
Five findings that repeated across 97 posts
- It is a carousel account, entirely. 96 of 97 posts are images. The single video in the sample earned 44 reactions, her weakest result.
- She writes long. A 397-word median caption, on top of a carousel, which is roughly four times the median of the other accounts in this series.
- The audience forwards her. She averages 89 reposts per post with a median of 56, high for an account this size.
- She never posts at the weekend. Zero posts on Saturday and zero on Sunday, across five months.
- She reruns posts, and it does not work. 18 of her 97 posts are republished versions of an earlier one, and 7 of the 8 rerun sets performed worse the second time.
The numbers behind the account
Consistent, tightly clustered, and carrying one finding that contradicts the other teardowns in this series.
Emilia's median post earns 393 reactions and her mean is 577. 16 of 97 cleared 1,000, 2 cleared 2,000, and only 4 finished under 100. The comment numbers are unusually stable: a mean of 125 against a median of 119, which means almost every post lands in the same band rather than a few outliers carrying the average. Her comment-to-reaction ratio of 21.7% is roughly three and a half times the norm.
The long caption is the interesting choice. Conventional advice says a carousel should carry the content and the caption should get out of the way, which is exactly what the Corporate Natalie teardown found for video. She does the opposite and it works, because her carousels are reference material and the caption is the searchable version of it. For more on how format affects distribution, see our guide to how the LinkedIn algorithm works.
When she posts
Where the engagement comes from
The rerun experiment, run 8 times
| Post | First run | Later run | Result |
|---|---|---|---|
| Getting cited by AI isn't random | 1,558 | 239 | Down |
| Stop blaming AI for terrible results | 1,145 | 338 | Down |
| AEO/GEO is not just SEO with a few tweaks | 1,000 | 524 | Down |
| 99% of people are learning AI the wrong way | 893 | 387 | Down |
| SEO, AEO and GEO are NOT the same thing | 2,052 | 1,555 | Down |
| This funnel powers multimillion-dollar businesses | 962 | 1,589 | Up |
The top posts
| # | Post | Reactions | Comments | Reposts |
|---|---|---|---|---|
| 1 | SEO, AEO and GEO are NOT the same thing | 2,052 | 254 | 313 |
| 2 | SEO might seem simple | 2,039 | 273 | 258 |
| 3 | This funnel powers multimillion-dollar businesses | 1,589 | 283 | 348 |
| 4 | Common misconception: SEO = rankings | 1,561 | 178 | 339 |
| 5 | Getting cited by AI isn't random | 1,558 | 220 | 194 |
| 6 | The modern marketing stack has changed | 1,523 | 232 | 238 |
Want to see whether your own carousels are being forwarded? Run your profile through our free LinkedIn analyzer.
The six content pillars
Six shapes, one job: move the reader from thinking they understand search to knowing they do not.
Two things the reader believes are the same, separated and defined. Her signature move.
'Common misconception: X.' Then the corrected model, in numbered layers.
A complete funnel or system, handed over free, with every stage explained.
What actually causes an outcome, reduced to three or five named signals.
What changed this year, framed as a playbook that has gone out of date.
Numbered workflows and prompts marketers can run today.
Pillar 1: The distinction (the reach engine)
Why it works: The capitalised NOT is doing the work. A distinction post only lands if the reader currently holds the wrong model, so the hook has to assert the error rather than tease the answer. 313 reposts, because people forward the thing that corrected them.
Pillar 2: The misconception fix (the most reposted)
Why it works: The structure is name the wrong belief, deny it in two words, then give the corrected model a number. 'Five layers' is what makes it saveable: a reader can hold five, and can check whether they cover all five. 339 reposts is her second-highest.
Pillar 3: The framework giveaway (the most saved)
Why it works: Her highest repost count at 348, and the only post that improved across every rerun: 962, then 1,309, then 1,589. Framing each funnel stage as a question the reader can answer about their own business is what turns a diagram into a self-assessment.
Pillar 4: The signal breakdown (the mechanism)
Why it works: 'Isn't random' is a promise that a system exists, and 'most brands are only strong in one or two' tells the reader they are probably failing before the list arrives. Worth noting this is the post whose reruns fell hardest, from 1,558 to 239.
Pillar 5: The stack update (the urgency)
Why it works: Her second-biggest post, and the shortest sentences she writes anywhere. Naming the cost of the misconception, 'visibility, customers, and revenue', is what converts a technical correction into something a marketing lead has to act on.
Pillar 6: The AI workflow list (the utility)
Why it works: The utility tier is her lowest-reaching pillar and her most practical: numbered workflows and prompts a marketer can run the same afternoon. Note the single-line caption here, which is rare for her. When the carousel is a checklist, the caption has less work to do.
The hooks that earned the save
Every one of them tells the reader they are currently wrong about something specific.
'SEO, AEO and GEO are NOT the same thing.'
'Common misconception: SEO = rankings.'
'Getting cited by AI isn't random. It comes down to 3 core signals:'
'This funnel powers multimillion-dollar businesses. And now I'm giving it to you completely free:'
'Most brands are still operating with a 2018 marketing playbook.'
'99% of businesses optimise for traditional SEO.'
The unifying move is that each hook installs a problem the reader did not know they had. For the mechanics of openers, our guide to writing LinkedIn hooks goes deeper, and you can pressure-test your own first line in the free hook generator.
Her line versus the tips version
- 'SEO, AEO and GEO are NOT the same thing.'
- 'Common misconception: SEO = rankings.'
- 'Getting cited by AI isn't random.'
- 'Most brands are still operating with a 2018 marketing playbook.'
- 'A quick guide to SEO, AEO and GEO.'
- '5 things to know about modern SEO.'
- 'How to get cited by AI: some thoughts.'
- 'Marketing has changed a lot recently.'
A voice that reads like documentation with a pulse
Short declarative lines, heavy vertical spacing, glyph bullets, and no jokes anywhere.
- Writes one sentence per line, with a blank line between almost every one, so a 400-word post scans like a list.
- Uses checkbox and arrow glyphs for structure, which keeps long posts readable past the 'see more' cut.
- Capitalises for emphasis rather than bolding. NOT, and numbered layer counts, do the shouting.
- Names the cost of being wrong. Misconceptions are framed as costing 'visibility, customers, and revenue'.
- Never posts at the weekend. Zero across five months, on an account otherwise publishing every weekday.
- Publishes essentially one format. 96 of 97 posts are images, and the single video earned 44 reactions.
The structure is consistent enough to lift wholesale, and it transfers to any field where your audience holds a slightly outdated model.
[The denial: 'Common misconception: X = Y.' or 'X and Y are NOT the same thing.'] [The correction, with a number: 'It isn't. X now spans 5 different layers:'] [Why the old model persists, in two short lines] [What it costs you to keep believing it] 1️⃣ [Layer one, with 3 to 5 arrow bullets] 2️⃣ [Layer two, same shape] [The carousel carrying the same structure visually]
What she does, and doesn't, do
- Assert the reader's error in line one
- Attach a number to the corrected model
- Give whole frameworks away free
- Write long, in one-sentence lines
- Keep a strict five-day week
- Tips lists and neutral guides
- Humour and personal material
- Video, which underperforms badly for her
- Weekend posting, entirely
- Waiting long enough between reruns
Producing a carousel plus a 400-word caption, four times a week, on one narrow topic, is a genuine production load, and it is why most education accounts thin out after a quarter and start recycling. CaptureFlow is an AI content agent that turns your expertise into weeks of on-brand content for every platform. You capture one idea in 5 minutes, a voice note explaining the distinction you keep having to make on client calls, and CaptureFlow, trained on your voice and your past posts, reshapes it into a LinkedIn post, a carousel, an infographic, and a short video. See how the AI content agent works.
The system underneath the posts
One narrow subject, one repeated move, and a company cluster running behind it.
The correction loop
- 1A client holds an outdated modelSEO means rankings, AEO is just SEO with tweaks.
- 2The correction becomes a carouselThe wrong belief, the denial, the numbered replacement.
- 3The reader recognises themselves21.7% comment ratio, more than three times the norm.
- 4They forward it to their team89 reposts on average, peaking at 348.
- 5The next misconception surfaces in the repliesAnd becomes the following week's post.
The Searchable cluster
38 of her 97 posts mention Searchable, and this is the second Searchable account we have taken apart. Chris Donnelly, the founder, runs a broad founder-brand account across business, marketing, and personal branding. Emilia runs the narrow technical beat. Same company, two very different jobs, and almost no overlap in subject matter, which is the same split we found across the three Lovable accounts.
Read this alongside the Chris Donnelly playbook for the company pairing, and the Lily Ray playbook for the same subject handled evidence-first rather than education-first. For marketing teams deciding who should post what, the contrast between those two is the most useful thing in this series.
Your 30-day challenge
Sixteen posts, weekdays only, each one correcting something your clients actually get wrong.
- Days 1-2: List every wrong assumption you have corrected on a client call this quarter
- Days 3-4: Turn the most common one into 'Common misconception: X.'
- Days 5-7: Give the corrected model a number, and build the carousel around it
- Days 8-10: Write the distinction post: two things your audience conflates
- Days 11-12: Name what it costs them to keep conflating those things
- Days 13-14: Give one complete framework away with no gate
- Days 15-17: Break one outcome into three named signals
- Days 18-19: Answer the misconceptions that appeared in your own comments
- Days 20-21: Check your repost rate; it is the signal that a post became reference material
- Days 22-24: Note which posts you are tempted to republish, and leave them alone
- Days 25-27: Write a fresh angle on the same idea instead of reposting it
- Days 28-30: If you must rerun, leave at least six months, not two
The stop-doing list
| Stop doing | Do this instead |
|---|---|
| Neutral tips lists | Naming the misconception the reader holds |
| Unnumbered explanations | A counted model: 3 signals, 5 layers |
| Gating your best framework | Giving it away and earning the reposts |
| Republishing every two months | Waiting, or writing a fresh angle instead |
| Posting video because it is recommended | Checking your own format numbers first |
| Weekend posting out of habit | A strict weekday schedule, as she runs |
The metrics to track weekly
| Metric | Benchmark to aim for |
|---|---|
| Reposts per post | 50+ (her median is 56) |
| Comment-to-reaction ratio | 15%+ (she sits at 21.7%) |
| Share of posts that correct a belief | 50%+ |
| Gap between a post and its rerun | 6 months minimum |
| Rerun performance vs original | Track it; 7 of her 8 declined |
| Posts per week | 4 to 5, weekdays only |
The takeaways
- 01Correct a belief, do not offer a tip. Emilia Möller's top post, at 2,052 reactions, opens by denying that SEO, AEO and GEO are the same thing.
- 02Give the corrected model a number. Five layers, three signals, five funnel stages. A counted model is what makes a post saveable.
- 03Write long if the post is reference material. Her 397-word median is the longest in this series, and her repost median is 56.
- 04Give whole frameworks away. Her most-forwarded post, at 348 reposts, hands over a complete funnel for free.
- 05Do not rerun too soon. 7 of her 8 rerun sets performed worse, with one falling from 1,558 reactions to 239.
- 06Go narrower than your founder. 38 of her posts mention Searchable, and she covers the technical beat its founder does not.
Frequently asked questions
- How did Emilia Möller grow her LinkedIn following?
- By publishing carousels that correct a specific misconception about search. Across 97 recent posts she averaged 577 reactions and 89 reposts, posting 4.3 times a week, with 96 of 97 posts as images and a 397-word median caption.
- What kind of post performs best for Emilia Möller?
- Distinction posts. Her top post, arguing that SEO, AEO and GEO are not the same thing, earned 2,052 reactions and 313 reposts. Her framework giveaway earned her highest repost count at 348.
- Does republishing old LinkedIn posts work?
- Her data says usually not. 18 of her 97 posts are reruns, and 7 of the 8 rerun sets performed worse the second time, one dropping from 1,558 reactions to 239. Only her funnel post improved, rising across three runs. The gaps that failed were two to four months.
- How do you sustain a carousel cadence on one narrow topic?
- Capture the corrections as they come up on client calls, so the backlog refills faster than it drains. CaptureFlow turns one 5-minute capture into a week of native posts, carousels, and infographics across platforms.
