The Educated Provocation Formula for LinkedIn

The posts on LinkedIn that travel furthest are almost never the most controversial. They're the ones that make a reader stop and check their own assumption. That's a different quality from outrage. Outrage is cheap — the algorithm catches up with rage-bait quickly and throttles accounts that lean on it. Richard van der Blom's 2025 Algorithm Insights Report, based on 1.8 million posts, shows top-creator content is now 31% of the LinkedIn feed, up from 15% in 2022, while pure company content has collapsed to 2%. The space is rewarding people who argue for something specific, not performers who shout.

I've been calling this "educated provocation" — strong, specific claims backed by something you actually know, aimed at a belief your industry takes for granted. Running FeedSquad's feed, the posts that consistently out-perform are the ones built this way.

The three pieces

Every post that holds up has the same three parts. Skip one and it falls apart.

A position. A clear, specific claim that contradicts something the reader probably believes by default. Not a twist for effect — a real disagreement. "Your company doesn't need a content strategy" works. "Marketing is dead" doesn't, because nobody actually believes marketing is dead, so there's no belief to dislodge.

Evidence. Something only you could say. Numbers from your own product, patterns from your customers, a reference to a study you actually read. This is what separates a provocation from a hot take. Without evidence, the post is just an opinion wearing bold type.

The implication. "So what should the reader do now?" If you can't answer that in one sentence, the position wasn't useful — it was just contrarian.

Here's a worked example from my own experience. Position: most SaaS companies under €1M ARR don't need a content calendar; they need one real customer story a week. Evidence: I've looked at the LinkedIn feeds of maybe forty SaaS founders in this range over the past year, and the ones pulling in inbound leads are all posting from the same well — their own customer calls — not from an editorial plan. Implication: throw out the calendar template; pick Thursday afternoon to write up whatever you learned that week.

That post is a disagreement someone could argue with. The argument is the point.

Why it works on LinkedIn specifically

It creates tension the reader has to resolve. When a claim contradicts something you assumed, your brain has to either refute it or update. Either way you keep reading. Clickbait manufactures the same tension and then fails to resolve it — which is why people remember they've been burned. Educated provocation resolves the tension with a genuine argument.

It signals expertise faster than credentials do. Anyone can share the consensus view. Taking a specific position against the consensus is either confidence or foolishness. If the evidence is there, readers read it correctly as confidence.

It filters for the audience you actually want. Consensus content gets polite scrolling. A strong position attracts the small subset of people who care enough to disagree, agree loudly, or add their own evidence. That subset is your actual audience — the rest were never going to convert anyway.

It gives the algorithm what it rewards. LinkedIn's ranking leans on dwell time and substantive comments, not likes. Posts that pull real replies — "I actually disagree because…" — get distributed further. AuthoredUp's analysis of LinkedIn ranking is consistent on this: long, substantive comments are among the strongest distribution signals the platform uses.

Four rules to keep it from tipping into inflammatory

The line between provocative and inflammatory is not subjective, even though people pretend it is.

Attack ideas, not people. "The traditional sales playbook is broken because buyers now self-educate" is a provocation. "Salespeople who still cold-call are lazy" is not — it's an insult, and it invites defensiveness, not discussion. The first one gets thoughtful replies. The second one gets LinkedIn's version of a flame war and tanks your next three posts' distribution.

Show the working. "SEO is dead for B2B" is dismissible. "I looked at our last 100 customers; zero came through organic search" is not. The second version takes the same position and turns the mute button off.

Acknowledge the counter. This feels counterintuitive — shouldn't you double down? No. A line like "I know this runs counter to every marketing playbook, and for B2C it doesn't apply" actually strengthens your position because it demonstrates you've thought about the obvious objection. The strongest posts preempt their strongest critic.

Scope the claim. "Marketing degrees are worthless" is inflammatory. "For early-stage startup marketing specifically, six months of building and measuring beats most four-year programs" is a defensible scoped claim. Overgeneralization is the most common failure mode.

Where to find provocations worth writing

Four sources I pull from consistently:

The industry report whose conclusions you disagree with. When you read something and think "they're misreading the data," that's a post. The strongest version names the report and links to it.

The advice that used to work but doesn't anymore. Markets change. Tactics expire. If you can articulate why a once-valid playbook is breaking, you're doing something the report cycle isn't doing.

The thing everyone says and nobody practices. "Customer first." "Fail fast." "Content is king." The gap between the slogan and the actual behavior is almost always a post.

Your own changed mind. "I used to think X. After Y experience, I think Z." This is the strongest kind of post because it's a receipt — you showed your work by doing it on yourself first.

The test before you publish

Read the post out loud. If you wouldn't actually say those sentences to a person in front of you, rewrite them. Most inflammatory posts fail this test because the author wouldn't repeat them face-to-face. Most weak, hedged posts fail it for the opposite reason — you wouldn't bother saying them at all.

Then ask: would someone who holds the opposite view read this and want to reply with their own argument, or would they want to block me? If the answer is the second one, the post is inflammatory, not provocative.

---

If structuring the position-evidence-implication loop every week is the part you keep putting off, FeedSquad's Ghost agent drafts campaigns from your actual writing — you edit, you don't prompt. Five posts free.

Sources:

• Richard van der Blom — Algorithm Insights Report 2025
• AuthoredUp — How the LinkedIn Algorithm Works in 2025
• Originality.AI — 50%+ of LinkedIn Posts Were Likely AI in 2025