Delayed engagement experiment

The relentlessly helpful blog

8 March 2021

LinkedIn delayed commenting

My February 2021 experiment tested whether delaying replies to comments for the first 3 hours after publishing my LinkedIn posts would have a negative impact on the visibility and engagement on those posts.

Introduction.

It’s often said that you need to be around to reply to early engagement on your LinkedIn posts if you want those posts to have maximum visibility in the feed.

People have previously mentioned there being a “golden hour” directly after a post’s publication during which time comments should be left – both by your followers and by you in replying.

Recent speculation about changes to the LinkedIn algorithm seems to suggest that this window of opportunity – the golden hour – has been extended to be 2 or 3 hours.

This is good news if true because a widening of the engagement window would help to dilute the potency of engagement pods, which I rail against in my LinkedIn course.

I wanted to test whether early responses to engagement on posts really mattered, so I decided to do an experiment in February 2021. This would involve deliberately avoiding replying to comments for the first 3 hours after each of my posts were published.

As I stated in the January post to announce my plan to do this experiment:

Because I’m a pathetic dopamine addict, I’m going to find this tough. Digital hugs matter to me more than I probably want to admit. Real hugs still matter more, though.

True enough, it was a tough one to commit to.

A quick summary of the results, taken from a chat with LinkedIn trainer Lynnaire Johnston:

The experiment.

I made a note of the time of publication of each post during February 2021 so that I could check the engagement stats after exactly 3 hours.

I could then compare this with the “final” stats that I gathered a full 7 days later, to look for any interesting correlations.

Here are the posts made during February 2021, showing both the 3-hour stats and the 7-day stats for each piece of content.

To save space, I’m not including data about shares, as the numbers are always low and never particularly interesting. See my article that explains why sharing posts on LinkedIn isn’t effective.

3-hour stats.

Title Type Views Reactions Comments
01. Check your connections’ languages Image 592 9 9
02. Why I didn’t become a biochemist Video 280 19 8
03. Use a secret word in your About statement Image 730 19 17
04. War and Peace in DMs Text 639 11 4
05. Friday Shout Vietek Ladislaav Image 746 15 14
06. B2C can and does work on LinkedIn Text 1007 34 8
07. How I handle some LinkedIn invitations Video 252 20 9
08. Reshared LinkedIn post text in bold Image 714 11 8
09. LinkedIn banner dimensions Text 620 20 3
10. Friday Shout Anthony English Image 697 16 23
11. Empty Cache and Hard Reload Image 929 20 5
12. LinkedIn scheduling experiment results for Jan 2021 Text 515 15 12
13. LinkedIn data analysis chat on Clubhouse Image 728 16 5
14. Your LinkedIn posts are mini blog posts Image 1290 31 14
15. Say my name (say my name) Video 670 58 37
16. Friday Shout Karen Tisdell Image 720 27 10
17. Audio pronunciation examples needed for my blog Text 723 11 5
18. LinkedIn audio pronunciation how-to & 25+ examples Text 644 13 18
19. Should you add captions to every LinkedIn video? Text 1021 22 20
20. How are you so productive? Text 845 21 10
21. Good-looking personal branding coaches Image 964 23 9
22. Friday Shout Ed Han Image 1159 35 13
23. My personal posts are outperforming my work ones – now what? Text 1427 33 19
24. TextExpander – save time with your commonly used text snippets Video 297 28 15

7-day stats.

Title Type Views Reactions Comments
01. Check your connections’ languages Image 3021 45 63
02. Why I didn’t become a biochemist Video 895 40 25
03. Use a secret word in your About statement Image 4117 86 93
04. War and Peace in DMs Text 2488 32 35
05. Friday Shout Vietek Ladislaav Image 3450 61 79
06. B2C can and does work on LinkedIn Text 5811 135 72
07. How I handle some LinkedIn invitations Video 1909 97 103
08. Reshared LinkedIn post text in bold Image 3182 47 66
09. LinkedIn banner dimensions Text 3103 82 33
10. Friday Shout Anthony English Image 3572 61 80
11. Empty Cache and Hard Reload Image 3026 47 27
12. LinkedIn scheduling experiment results for Jan 2021 Text 3737 106 120
13. LinkedIn data analysis chat on Clubhouse Image 2153 43 24
14. Your LinkedIn posts are mini blog posts Image 6317 137 84
15. Say my name (say my name) Video 3709 177 168
16. Friday Shout Karen Tisdell Image 3171 89 105
17. Audio pronunciation examples needed for my blog Text 4434 37 64
18. LinkedIn audio pronunciation how-to & 25+ examples Text 3816 55 146
19. Should you add captions to every LinkedIn video? Text 4425 57 95
20. How are you so productive? Text 7778 109 128
21. Good-looking personal branding coaches Image 4145 75 58
22. Friday Shout Ed Han Image 5533 109 87
23. My personal posts are outperforming my work ones – now what? Text 7829 106 88
24. TextExpander – save time with your commonly used text snippets Video 1205 80 67

Here’s how the 3-hour stats (orange) compare with the 7-day stats (green).

Views.

Views

Reactions.

Reactions

Comments.

Comments

The 3-hour stats for each post are a pretty good indicator of the 7-day performance.

There are some exceptions:

  • Post 20 had more views than expected.
  • Posts 3, 12 and 20 had more reactions than expected.
  • Posts 6, 7, 12, 18 and 20 had more comments than expected.

I can’t see anything obvious to explain what happened in these cases.

It’s hard to conclude anything from a small dataset of only 24 posts, but the view count measures over 3 hours and 7 days seemed to correlate well: in other words, 3-hour views were a good predictor of 7-day views.

During this experiment, my 7-day view counts were on average 5.1× higher than the 3-hour view counts, and the variance was low (16 of the 24 posts were in the 4× and 6× range).

Here’s how my engagement stats for the February 2021 experiment compared with my averages for the previous 6 months.

Month Views Reactions Comments
Aug 20 9499 173 103
Sep 20 7080 126 85
Oct 20 5866 110 89
Nov 20 10146 223 136
Dec 20 4774 96 93
Jan 21 8076 124 120
Average 7574 142 104

Compare the above 6-month average with the average during the February experiment and you’ll see that there was a significant drop in all engagement stats:

Month Views Reactions Comments
Feb 21 3868 80 80
Change –49% –44% –23%

Let’s wrap up.

The stats suggest that there was a negative impact associated with not replying to comments for 3 hours after my posts were published.

So, when you post content on LinkedIn, make sure you’re around to do at least a little bit of “babysitting” to help the post achieve the full visibility it’s capable of.

   

Be part of Espresso+

The community for freelancers & small business owners.

114 recommendations
for John

John Espirian

I’m the relentlessly helpful®️ LinkedIn nerd and author of Content DNA

I teach business owners how to be noticed, remembered and preferred.

Espresso+ is a safe space to learn how to ethically promote your business online and get better results on LinkedIn.

Follow me on
LinkedIn

Share on
social media