Does putting a link in a LinkedIn post really cut your reach?

LinkedIn has never published any statement that a post is demoted for containing a link. Independent measurements disagree on both the size and the direction of the effect: studies covering hundreds of thousands of posts have reached opposite conclusions on the same question. BrandLoop’s analytics tab runs that comparison on your own account data, not on published averages.

“Put a link in your post and your reach dies — put it in the first comment instead.” Repeated often enough that it now travels as fact. Open the sources one by one and a different picture appears.

LinkedIn has never said this

Help pages, the engineering blog, the corporate blog, on-record employee statements — there is no LinkedIn sentence saying distribution is reduced because a post contains a link. The help page language is about “low-quality and unsafe content”; links are not a named category.

LinkedIn’s February 2026 ranking paper does not list link placement among its objectives either. Absence from a paper is not proof of absence in the system, but the most technical primary document available does not support the folklore.

The closest on-record comment came from a LinkedIn product director in August 2025, advising that if you use an external link, the post should stand on its own without it. That is advice about post quality, not a statement that links are penalised.

The 60% figure has no source

It appears across dozens of articles, always uncited. An independent audit that tried to trace the “60% external link penalty” concluded it leads back to no identifiable source.

The most-cited experiment used seven posts

The claim that linkless posts get six times the reach traces to a 2021 Hootsuite experiment. The sample: 184 posts from a single company page over two months, of which only seven contained no link. The limitation is not ours — Hootsuite printed it, noting the linked posts had 177 chances to go viral while the linkless ones had 7.

Seven posts, one account, five years ago. That number cannot carry the weight it is asked to carry in 2026.

Large datasets contradict each other

Two things stand out. The account-type split is inverted between two large studies: one finds company pages penalised, the other finds them performing above average. And in the 900k dataset, link posts got less reach but a 22% higher engagement rate — fewer people, more response from them.

The variable may be the preview card, not the link

The 567k-post measurement puts almost the entire deficit in one place: the attached link preview card. Posts with a card had a median of 414 impressions, while posts with a plain URL in the body slightly outperformed posts with no link at all. The study’s own caveat supports this — card posts are overwhelmingly the article format, which underperforms independently.

That reframes the whole argument. What is being measured may be the publishing format, not the presence of a URL. Any study that lumps the two together is measuring the wrong thing.

Does link-in-first-comment work?

The honest answer is that nobody knows. Of the two datasets large enough to answer, one states plainly that it did not measure it, and the other does not disclose the subset size and describes its own finding as a signal to monitor rather than a playbook. LinkedIn has never commented. Confident claims in either direction are unsourced.

What to actually do

• Write the post so it makes sense without the link. This is the advice closest to evidence, and it comes from LinkedIn’s own employee.
• Try a plain URL in the body instead of an attached preview card. The measured deficit sits mostly with the card.
• Measure on your own account. Two large studies invert on account type, which suggests the answer may be account-specific.
• Distrust confident percentages. Anyone quoting a precise figure in 2026 has not read the sources.

The analytics tab exists for exactly this kind of comparison on your own posts. What the feed ranks for is covered separately in our piece on feed ranking.

The larger lesson here

The link question is really an example of a bigger problem: most social media advice rests on things that look like measurements and are not. The same pattern shows up elsewhere — hashtag counts, emoji use, editing a post, posting at particular hours. Each has the same chain: an agency post citing an agency post, with no measurement at the end.

Four questions are enough to test any claim: how many posts, over what date range, using what method, measured by whom. Without answers you have a habit, not a finding. And as the table above shows, even studies that can answer all four can contradict each other — but at least the contradiction is visible.

Measuring on your own account is easier than it sounds: publish the same subject in two different formats two weeks apart and put the impression counts side by side. One comparison proves nothing; ten show a tendency. The one thing published averages can never tell you is what your own audience does.