Think Tank: Instagram Likes

October 30, 2014

instagram, statistics, think tank, writing

As part of my duties as student assistant extraordinaire, I help run Cal Poly Pomona’s Instagram. On my way into the office at 1 p.m., I snapped a photo of the CLA Building, edited it a little bit, and posted it to the account. By 5 p.m., it had nearly 400 likes.

While I was monitoring the photo this afternoon, I started to think about what makes a photo on social media so successful. There have been a couple of photos we’ve posted that barely graze 200 likes. I know I have a horrible habit of looking at a post and either making a comment to myself or scrolling past it. So why do people like the posts they do? It it because they’re familiar with the subject, like the colors or just think its good photography?

Of course, I went to Google.

I didn’t find any substantial data, since a lot of the studies I found gave me no information about sample sizes and methodologies. (Ever since I read this book, I can’t trust a statistic until I know exactly how the study was conducted.)

Via a Time article, I did find a research project conducted by some MIT researchers. They gathered date from a little over 2 million photos on Flicker and came up with an algorithm. The formula supposedly takes the actual image into account, and the aesthetics too: color, texture and gradient. The algorithm then predicts how popular the photo will be, based on the subject and the context of the photo. You can test your own photos on their website.

However, I’m not completely sold. The algorithm said that the same photo would only get about 21 likes per day, It didn’t ask me upfront for any information about my audience (2,500 people), whether or not I had visibly edited the photo (yes) and if I had used hashtags (Only one: #calpolypomona.) Their research paper, which is interesting to skim through, did discuss those factors a little bit. But they only made one solid deduction that I thought was kind of obvious: for Flickr at least, better-looking photos equal more followers.

It seems like cold, hard science hasn’t encroached on the realm of Instagram research yet. So I’m opening up this Think Tank to you too. Why do you like a photo on Instagram? What compels you to leave a comment, notice the photo editing or read the caption? You guys are better than science, anyways.

Leave your comments below.

2 responses to “Think Tank: Instagram Likes”

Michael J

November 2, 2014 at 10:56 pm

I like pictures that interest me; particularly, pictures that leave me feeling like I gained something from viewing it, even if that something is a bad taste in my mouth. With that said, I am absolutely flabbergasted by the relatively high number of likes the picture in question has generated, as it seems woefully ordinary by comparison to some of the other pictures posted this past month. Maybe a nearby planet was reflecting a disproportionate amount of Alpha waves, or perhaps there was some secret gentleman’s agreement that blue and green were the desirable colors for the week. At this point, anything seems plausible.

As for commenting, I think the reasons are a little less esoteric. Based on those that I’ve read over the past few years, I’d venture a guess that most people who do either want a.) the poster to know that they saw the picture (and in some cases, acknowledge them) b.) others viewers to think they are funny/clever/insightful or c.) just feel like making their opinions/feelings seen. Since I don’t personally care about any of those things, I never comment on photos (like, ever), but, then again, I’ve also been described as a “slightly-less-comically-serious Ron Swanson,” so perhaps my use of Occam’s razor here is skewed or inappropriate.

Anyways, this was an interesting read, and I’m curious as to what you discover about these social science mysteries going forward.

Sidenote: I still can’t believe an entire algorithm was written to predict human behavior without taking into account something as basic as audience demographics. MIT, I am disappoint.

Reply
1. zoelance
  
  November 3, 2014 at 8:34 am
  
  Thanks for the reply, Michael. I’m just as disappointed in MIT as you are.
  
  Reply