We’ve been publishing experiments since late August 2017.
Like many of you I’m concerned about the validity of data on a blog post when it’s more than 12 months old.
Afterall social media changes every day it seems like!
Outdated data is a major concern for us, and is something we take rather seriously.
While we won’t re-test every experiment due to time restraints, and due to findings, we have selected a few “hot topic” experiments to run again.
The first experiment to get a re-test is one I did in September of 2017 on Instagram Carousels.
Instagram Carousels were brand new at the time and our results were not great for them:
Overall we found that Instagram Carousels in a word: SUCKED!
Maybe after nearly a year of this feature on Instagram they get better results, right?
My re-testing hypothesis: Instagram Carousels will have lower engagement than traditional single photo posts
Since this was a re-test I didn’t run a full scale test as normal.
I basically wanted to see if the data still held up — if it does then no further testing is needed, if it doesn’t a full scale test would need to be done.
I used the same 2 Instagram accounts in this re-test as I did the first test, but did post less carousels to them:
Below is an example from each account.
The data on the re-test wasn’t kind to Instagram carousels!
Here’s the data from the original test:
Space Walk Original Test Data:
Agorapulse Original Test Data:
Somewhat mixed results, but overall Instagram Carousels still under performed compared to regular, single photo posts.
The main 2 stats I look at on Instagram for success are Likes and Reach (in that order). Both were still lower for each account during the re-test.
One data set I did not include in the re-test was looking at different popular accounts, as I did in the original post.
The last time I ran this test many questioned if this was a very accurate way to pull in results and some viewed it as arbitrary.
To some extent I agree with them as I had to go back weeks, if not months, to find carousel posts, and the follower count could be drastically different depending on when the post was made.
So that data set is not included in the re-test.
And doesn’t sway my conclusion at all.
Based on the data from the re-test I can confidently say Instagram Carousels will result in a decrease in Likes and Reach.
I would recommend avoiding them, instead opting to post single photos in a post.
Bottom line: Instagram Carousels still SUCK!