Hello. My name is Jon and I’m a recovering Instagram bot addict.

Jonathan Snow

Lets just say I’ve been in the Instagram game for some time now. I‘ve been using the platform since early 2012 and was even around to experience the network before the meteoric rise of automation, spamming, fake likes, and engagement groups. With my recent work in the Instagram analytic space and my redirected focus towards improving transparency in the industry, I need to get something off my chest… I am a recovering bot user.

In fact, I had both designed and used bots heavily in the early years of social media on Twitter and Instagram. Like most great things moderation is key, a line I crossed many times in the early days. The main strategies at that time were a mix between follow/unfollow, rotating hashtags, and mass liking. I found pretty significant success in the process, growing my accounts to over 100,000 followers on each platform in less than 12 months. Brands were sending me products, my Klout score was decent (remember that thing?), and my ego was running on likes and retweets until one day I had enough. That’s it, I quit.

I gutted my Instagram account, removing nearly all of my followers and pressing pause on my curated social personality.

Enter Instagram 2017 and the environment is completely different. At the time of writing the most recent (sequential) user id number is 5,969,508,448… ALMOST 6 BILLION and is most likely more now, this figure is growing at a rate of around 5 million every day. So where the hell did all of the accounts go? If you needed some quantifiable measure of spam on the platform, there you go.

Some of my recent “tinkering” gave me the opportunity to put my personal Instagram account on the chopping block; allowing me to finally do a data driven analysis and purge of the garbage accounts that are still following me after all of these years. I had a hunch that my low relative engagement rates would lead me to find that many of my followers either have not ever engaged with me or had not in a long time. Another curiosity was whether these ghost followers were driving down my overall reach due to low relative engagement during the sample distribution that occurs shortly after posting (that was a mouthful). These are my results…

I conducted a comprehensive comparative analysis that approached the ghost problem from a few angles, but it boiled down to this higher level process. Track everyone that had ever engaged with my posts by liking or commenting, aggregate the counts for each follower, and then analyze and separate that data. I was able to sort by total engagements, so not only was I able to find the ghosts…

I could also find the most valuable members of my audience… this is SUPER powerful information (a good opportunity for brands to connect with their top contributors too 😉). More about this coming in future posts…

For additional context, I had 386 posts at the start of this process. Without overly complicating this first-pass analysis (and the scope of this article), the average number of engagement for the above 16 followers is 197. This means that, on average, they had engaged with over 51% of my posts! Some other measures in place for the second-pass will be understanding engagement opportunity, so each follower can be scored based on percentage-of-engagement based on the total possible opportunities to engage.

Welcome to 2017 folks, vanity metrics are bullshit and you’re not only lying to your advertisers and audience, you’re lying to yourself.

The next step was to design the parameters and underlying process to facilitate this purge. I decided that on the first pass I would keep the parameters simple. Grab the resulting list of followers that have never engaged with any of my posts, whitelisting everyone I follow (friends, I forgive you but I 👀 you). From my starting audience of 11,669 followers this gave me a first-pass purge list of 6,490… WOAH. This meant that almost 56% of my followers had not EVER engaged with me. I took a look at some of the usernames on the list and found that I would soon be purging, among many others, shoutouts_equal_morefollwers, troll_face247, and freeshoutoutsforyourdad, good riddance! (sorry dad, no free shoutouts)

Over the next couple of days I will finish the first-pass purge of over 6 thousand ghosts while also doubling my relative engagement. What impact would this have on my account? I will be determining the parameters of the second-pass purge and doing a deeper analysis on the remaining followers, including looking at activity recency in determining who else gets the cut. I plan to also try to get back in to Twitter and plan to complete the exact same process there. Stay tuned for more!

As always you catch me on Facebook, Twitter, and Instagram!

Until next time folks!

Jonathan Snow

Written by

Full-time Nerd. Part-time Human.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade