We analyzed 69,420 LinkedIn posts in Q1 2026. Here's what actually works now

That's a lot of posts haha. Except now everyone is doing that exact same thing because thats what AI tells them when they ask. Which I'm not sure if there's a workaround to that or not

Thats is the reason I see bunch of infographics in Linkedin recently.

The 69,420 post dataset is impressive, but the insight quality depends entirely on how the data was cleaned before analysis — and a few of these findings are worth pressure-testing.

The "links in post vs links in comments" finding especially has confounding variables: account size, industry, post age, and whether the posts were by high-follower accounts could all explain the performance difference without LinkedIn's algorithm actually changing behavior. That's not a knock on the finding — it's directionally useful — just something to validate against your own account cohort before making it a permanent strategy.

The 6x/week frequency multiplier and the 1,250–3,000 character sweet spot are more robust because they hold across variance. Those I'd operationalize right away.

From a data infrastructure perspective: the most valuable analytics you could add to LiFast isn't benchmarking against industry averages — it's helping each user understand their own performance trend over time. Their cohort data matters more than population norms.

Interesting data work here — for anyone building analytics on top of large datasets, I've put together a free SQL Server interview Q&A that covers the data reasoning patterns behind this kind of analysis → https://growthwithshehroz.gumroad.com/l/vgiex

Interesting result.

One question I’d be curious about: how well does this work for low-signal LinkedIn accounts?

For example, founders who are just starting out, have a small network, very few past posts, and not much engagement history yet.

Does the system still find a useful content direction from the product/service page alone, or does it work much better once the account already has enough audience and content data?

The "links in post vs comments" finding is the one that matters most because it contradicts what 95% of LinkedIn coaches still teach. If that holds up at scale, it changes a lot of playbooks overnight.

One pushback on the 6x/week claim though. Total views going up 3.1x is interesting but reach-per-post almost always drops as frequency climbs. A founder posting twice a week with sharp content often outperforms someone posting daily with filler on the metric that matters, which is qualified replies per post.

Curious how you controlled for account quality (follower count, niche, age) across the 69,420 sample. Anything in the methodology that surprised you?

BEST RECOVERY HACKER CRYPTOCURRENCY /BANK ACCT RECOVERY/ MARK WIZARD HACKERS RECOVERY

I became a victim of a romance scam that led me to invest $654,000 in Tether USDT on a fake site. After realising I had been scammed, I felt utterly hopeless and desperate for help. I began searching for a hacker online and came across Wizard Hackers Recovery. His professionalism and expertise gave me hope. Wizard Hackers Recovery worked to recover my lost funds, and I’m thrilled to say he succeeded! I am incredibly grateful for their assistance and highly recommend their services to anyone in a similar situation or any issues you can connect with them Thank you,

Email:       :Markwizard 23 @ gmailcom

WHATSAPP    +4915218465599

TELEGRAM :   @ wizardhackersrecovery

The niche beats broad point hits hard. Saw the same thing on Twitter building Trakly, replies to hyper-specific personal finance threads (first paycheck, broke after rent, etc.) consistently outperformed anything I posted broadly. The more specific the pain point I was speaking to, the better the engagement.

The links in post finding is interesting too. The 'hide it in comments' meta has been around forever, good to see actual data challenging it. Will be testing that.

Curious about the Claude vs GPT finding, do you think that's an audience sophistication thing or just novelty at this point?

Saw the number 69,420 and thought this post might be a joke. But, this looks insightful. Thanks.

I believe the consistency really works. Even if it is a random post, if it is consistently posted, then it gathers engagement.

Thanks! Always struggling to find the right way.

Interesting read, especially as someone who just started building on LinkedIn properly (and it's me). One thing I'd add from the past couple of weeks: when a post raises a real pain point, comments do most of the reach work for you. Two of my best-performing posts so far came from genuine frustrations I'd been sitting on. The connection requests that followed were the kind I actually wanted to accept, which doesn't always happen with high-volume content.

The niche-over-broad point feels especially true right now.

The posts that perform best don’t just teach — they make a specific audience feel understood within the first few lines.

We’ve been seeing something similar while building AI learning experiences. Users engage much more deeply with focused, guided understanding than with generic “AI for everything” positioning.

Feels like specificity is becoming a competitive advantage across both content and products.

The 6x/week consistency point hits close to home — it's another reason why I built PostShuttle. The consistency problem is real and most people underestimate how much the algorithm rewards it.

What was the engagement difference between accounts posting 3x vs 6x/week? Curious whether the drop-off is linear or whether there's a meaningful cliff somewhere in between.

Wow, this has been really useful. Great work keep it up!

Great insights. The links-in-post shift and the rise of infographics are especially interesting, definitely challenges a lot of outdated LinkedIn advice.

Useful data drop. The links in post vs comments thing is the one I would not have guessed since that was the standard advice for so long. Curious if you have a sense of when LinkedIn actually changed that or if it just gradually shifted. Also surprised by the Claude versus GPT engagement gap

The 6x/week finding lines up with what we see across SocialPost.ai users, but there is a caveat worth flagging that this report does not address: post velocity matters more than raw weekly count.

What we see at scale across founder accounts:

Posting 6x in 6 days outperforms 6x in 2 days. Clusters get penalized. The algo rewards steady cadence, not "I batch wrote 12 posts on Sunday and queued them all for Monday afternoon."

The carousel decline is real, but the shift to single image infographics seems linked to dwell time, not format preference. LinkedIn started weighting average time on post in late Q4 2025. Dense text infographics hold attention longer than 12 slide carousels people scroll past. The lesson is optimize for dwell time, not just pick infographics.

On links in post: links to your OWN domain seem to be punished less than links to external articles. The algorithm is discouraging traffic OUT, not traffic to verified content from posters who already drive engagement on platform.

The Claude vs GPT freebie shift is the most interesting data point in your post and almost nobody is talking about it. That is a B2B buyer signal showing up in social engagement data before it shows up anywhere else.

Question for your dataset: is the Tuesday/Wednesday/Thursday window still dominant, or are Sunday posts getting reach now that creators are flooding weekdays?

The 3,000 character sweet spot is surprising. Most people think shorter is better. Any chance you looked at the difference between list-style posts vs narrative posts at that length?

Solid data set. The "links in post vs comments" finding is the one I'd be most cautious about treating as universal though. We've watched the same pattern flip twice in 18 months at Hivemind across different LinkedIn-heavy verticals. SaaS/B2B founders are seeing links-in-post outperform right now. Creator-economy and personal brand accounts are still seeing links-in-comments win. The algorithm seems to be category-specific, not platform-wide.

Same caution on 6x/week. It works for accounts with a clear niche and a content engine behind them. For accounts still finding their angle, posting 6x/week of underbaked content actively trains the algo to suppress you. Frequency only compounds after positioning is locked.

The Claude > GPT freebie shift is the one I'd bet on hardest. That's not a posting tactic, that's a buyer-preference signal showing up in LinkedIn data first.

Love it!

Interesting to see niche content winning again. I’ve noticed the same — the more specific the audience and problem, the better the engagement. Also surprised the “links in comments” tactic is fading now.