Link to Experiment
RICHARD: Today, will text on your Instagram images affect your performance? Turns out the answer is yes. But in what way? Let’s find out on today’s social media lab
RICHARD: Welcome back, Scott! Happy 2020!
SCOTT: Thank you, you too! Yeah, we were a little slow to get rolling this year but we’re finally back in the swing of things, I think.
RICHARD: Yeah, experiments like ours take time, sometimes 6 to 8 weeks to complete but we’re rolling now
SCOTT: Yeah we’re rolling now
RICHARD: So! What’s our first experiment of the new year?
SCOTT: Of the new decade! We’re starting out this decade with an Instagram experiment that was suggested by viewers of our Social Media Lab LIVE show. (Never miss an episode by going to socialmedialab.live btw)
RICHARD: That’s totally awesome! Love what you and Owen Video are doing on the Live show, you gotta have me on some time!!
SCOTT: Done!! Hahaha. Some of our viewers suggested adding text onto their Instagram photos was creating more reach and engagement. So we wanted to test this theory in the Social Media Lab.
RICHARD: Interesting. So simply adding some text onto an Instagram photo before posting might be helping your posts?
SCOTT: That’s what some claimed, However, I couldn’t find any data from other sites that either confirmed or refuted this claim.
RICHARD: Yeah, just lots of blog posts suggesting Instagram users do it and how to make them look all pretty!
SCOTT: Exactly… So we gotta test it.
RICHARD: And how are we gonna test it?
SCOTT: Well, we’re gonna use 7 Instagram profiles from different industries with varying audience sizes. Each profile will publish 7 posts with text on the photos and 7 posts without text on the photos giving us 49 posts to analyze across multiple industries.
Each of these profiles will publish content that is in line with their typical content strategy, so our fitness profile will continue to post fitness photos like they normally would, our tourism profile will continue to publish tourism content and so on and so on.
RICHARD: Yeah, and as usual it looks like we’re set to prevent anomalies as best we can-
No hashtags will be used
No users or locations will be tagged
These posts will be published on the Instagram feed only, so no posting across networks or to stories
None of these posts will be used as ads (I can’t believe we have to say that one, but…)
SCOTT: But believe it or not, a lot of the social media studies we read about do this, they analyze posts that were spread out in multiple places or used as ads.
RICHARD: And that just skews the data too damn much so we don’t do it. Ummm, each profile is sticking to its regular publishing schedule and all posts will be scheduled and published using Agorapulse.
SCOTT: Yeah, and you can see examples of the posts we’ll use and the actual publishing schedule at agorapulselab.com
RICHARD: Ok, sounds like a tight experiment. Scott, what did we find?
Do we need to worry about coming up with witty text to place on our Instagram photos?
SCOTT: Well, the short answer is no. Instagram photos without text outperformed in terms of both reach and engagement. Now, typically, we focus solely on reach when it comes to organic content but for obvious reasons engagement is possibly more important on this one.
RICHARD: Understood so let’s break it down. Instagram photos with no text received:
14% higher reach
40% more likes… WOW!
And between 2 and 3 % more comments.
Damn, now again, these results are from multiple profiles from multiple industries with varying audience sizes and the performance convincingly indicates Instagram photos without text perform better. Much better.
SCOTT: Yeah, and just to take it a step further, we ran our results through a statistical significance calculator and got 100% certainty that these results will continue to be true in the future.
RICHARD: 100% statistically significant. Well, what if one of our listeners out there thinks their Instagram profiles may perform differently? How can they run an experiment of their own?
SCOTT: It’s simple.
First – isolate your experiment. Don’t make the mistake so many others make by just tallying up past data from previous posts. Set aside time to alternate photos with text against photos without keeping with your typical content strategy and publishing schedule, then wait 7 days after the last post in your experiment to draw conclusive data and compare.
RICHARD: How many posts do you think they’ll need to run for their experiment?
SCOTT: Um, we ran 7 per profile, you could double that up and run 14. Just be sure to run your results through a statsig calculator to make sure the results of your A/B test are significant. There are plenty of them out there and they’re easy to use. I use Neil Patel’s statsig calculator for our experiments and it will work fine for our listeners as well.
RICHARD: Alright, well, it’s great to be back and I know what I won’t be doing this year – striving to create witty copy to overlay on our Instagram photos.
SCOTT: No, not only is it a time suck it could actually be hurting your performance. According to our statistically significant results, Instagram photos without text will outperform every time which could lead to even higher reach and engagement for your brand in the future.
RICHARD: HEY! Is there an experiment you’d like the social media lab to run for you?