How often should our agency publish on LinkedIn?

Short answer: two posts a week from the founder, one post a week from the agency page, and a fortnightly cornerstone that everything else atomises from.

Five surfaces a week, total. Less than you've probably been told. Mechanically more useful than the daily-posting advice that's everywhere.

The longer answer is about who posts, what shape the posts take, and why this cadence works for B2B service businesses where the buyer base is small and concentrated.

Why founders outperform agency pages

Founders consistently outperform agency pages on LinkedIn, by every metric that matters for a B2B service business. Reach per post is higher. Reply rate is higher. Inbound conversation - the kind that turns into a meeting - is higher by an order of magnitude.

There are mechanical reasons for this:

• LinkedIn's feed algorithm favours individual posters over company pages, and has done since the 2018 redesign. Posts from individuals get shown to roughly 5–10x more of a similar-sized network.
• Buyers want to know what a named human thinks. An agency page saying "we believe X" reads as marketing; a founder saying "I'd argue X" reads as an opinion.
• LinkedIn's engagement signals (comments, replies, saves) come from human-to-human interactions. People reply to people, not to brand handles.

So the answer to "how often should the agency page post?" is "less than the founder, and for different reasons."

The cadence I'd recommend

For a typical UK agency, founder-led, between five and forty people, B2B service mix:

Founder posts: two a week. One opinion, one case-and-craft. The opinion is a take on something happening in the industry (a category move, a piece of work, a published number, a hiring trend). The case-and-craft is something specific from the work - a problem-and-solution story, a craft note, a piece of behind-the-scenes that explains how a decision got made. Two posts is the minimum cadence that signals "this person is here regularly" without becoming the second job.

Agency page posts: one a week. Either an explicit announcement (a win, a hire, a piece of work going live) or a curated piece (a link to a founder post, a piece of trade press, a recap of a podcast). The agency page exists for the prospect who arrives via Google, clicks through and wants to see signs of life. One post a week is enough to show life. More than that and the page reads as automated.

Cornerstone: one a fortnight. The cornerstone is the high-effort, substantive piece - a long-form opinion, an answer page on a real client question, a published-in-public update. It usually goes on the agency's own site (in the journal), is announced by the founder on LinkedIn, and then gets atomised across the next two weeks of social.

That's the system: five social surfaces a week (two founder, one agency page, plus the cornerstone announcement and atomisation passes around it).

What "atomisation" looks like in practice

A single cornerstone post should produce two weeks of social material if it's been written properly. The atoms aren't summaries - they're standalone posts that each surface one specific point from the cornerstone.

For a 1,500-word cornerstone, expect:

• Three opinion posts from the founder, each pulling out one argument and writing it as a standalone post (not a teaser to the article - a fully argued standalone).
• One case-and-craft post connecting the argument to a piece of work.
• One agency page post linking to the cornerstone with a curated lede sentence.
• One newsletter inclusion, with the founder's framing.
• Optional: a short video, an infographic version of a key point, a comments-section conversation starter.

That's the leverage. The cornerstone takes two days to write. The atoms take an hour each. The output is two weeks of social that all points back at the same argument, which is the only thing the algorithm and the human reader notice.

What about reach? Engagement? Followers?

These are mostly the wrong metrics for an agency at five-to-forty people.

The right metric for B2B service businesses on LinkedIn is inbound conversation from the right people. A single message from a marketing director at a brand you'd want to work with is worth more than a thousand reach-impression increases from people you'll never speak to.

The cadence above is built around that metric, not around vanity reach. Posting twice as often doubles your reach to a network that's mostly inactive on LinkedIn anyway. It doesn't double the rate at which the right people start a conversation with you.

Common mistakes

Posting daily. Burns out the founder, dilutes the opinion, and trains the algorithm to deprioritise your posts because they're low-engagement-per-impression. The two-a-week cadence outperforms five-a-week in the metric that matters.

Founder ghost-posting only. Posts that obviously aren't written by the named founder underperform. Voice matters. Either the founder writes (with editorial support to sharpen), or the ghostwriting is invisible enough to pass the friend test. Bad ghostwriting is worse than no posting.

Posts that are only takes. A founder feed that's all opinion, no work, reads as a podcast host, not as a credible operator. Alternate. The case-and-craft post is what reminds the reader that the opinions are coming from someone who does the work, not someone who just comments on it.

Posts that are only work. The opposite mistake. A feed that's only "we won X, we made Y, we did Z" looks like a brochure. Buyers don't share brochures. They share takes.

What to do this week

If you want to start the cadence:

1. Pick the two posting days. (Tuesday and Thursday is the default for B2B; Mondays are crowded, Fridays are dead, weekends are noise.)
2. Write three founder posts in advance. One opinion, one case-and-craft, one buffer.
3. Pick the cornerstone topic. Write it. Schedule it for the fortnight you want it to land.
4. Plan the atomisation in advance - the day the cornerstone goes live, the days each atom drops, what each atom argues.
5. Track one metric: inbound DMs from people you'd want to work with. Everything else is secondary.

The cadence stops being effortful around week four. By week eight, the engine runs.