LinkedIn - Likes versus Comments

The results are in from our recent 'likes' versus 'comments' test on LinkedIn.

A quick reminder about the test

We issued a post where we switched off the opportunity to comment.

We issued another post where we asked people not to 'like' it. One person did.

The results

The "like" post has so far got 88 likes and 1,799 views (below is the live link):


The "comment" post so far has 136 comments and 11,910 impressions (see live version below):


What does this mean for our LinkedIn strategy?

Lesson one: do not switch off comments.

The first thing to note is that switching off comments kills a post's total reach. If we wanted to get reach of 1,800 views, it'd normally take around 34 likes or comments combined. We got 1,800 views but it took 88 likes. If we reverse that, we'd expect around 4,500 impressions for that number of likes. However, it's only 41% as powerful with the comments switched off. 

Lesson two: comments considerably outperform likes (as I suspected they would) in terms of total impressions.

The "comments" post got around 87 impressions per comment, the "likes post got 20. Now, this is not a pure comparison because we can't switch off "likes" and so a couple of bots did, in fact, like the content. But everyone else respected the rules and again, based on our prior model of the algorithm, we'd estimate that this post has over-performed by over 5,000 impressions (so far).

Lesson three: use the right content strategy for the outcome you're seeking.

If we want more impressions (which is a comms strategy) then we need to change our content to reflect that and draft posts which invite comment.

If we want higher click-through or contact, then we need a business development strategy (which might well involve fewer impressions).


On LinkedIn, we: analyse firms' and leadership teams' performance and benchmark them against each other; and we train people how to write better posts and achieve more impressions.

Please drop us a line if either or both of those services are of interest: