The next AI: Algorithms and influencers

Algorithms make our lives radically easier. But they also work against us. Whether algorithms give you excitement or anxiety, they are top of mind in agency world: we’re all working with Facebook’s new tweaks to its algorithms, which are giving greater weight to engagement.

This is a significant shift for brands, not only on their own channels, but in their influencer marketing strategies as well. Likes, shares, comments—comments more than anything else—are now top-of-mind for content that wins. We need to change how we evaluate content performance too (more on that in a second.)

But first, a telling tale: here’s what I mean when I say algorithms can work against us. A so-called Facebook engagement “hack” is making the rounds, entering the realm of cautionary tale meets urban legend. The first time I read about it, I wasn’t even sure it was true. I’m still not, but here’s the gist…

A woman posted a video to her Facebook feed that was annoying to her friends. People saw it and commented on it. Because it was receiving comments it was appearing on the top of her friends’ feeds for days and days because comments kept pouring in. These were generally about the fact that they disliked it and wanted it to leave their feeds. All of the chatter about wanting the content gone ironically kept the content alive. The vicious circle of engagement continued. The hypothesis is that this is a result of the new Facebook algorithm providing more weight to engagement especially as it pertains to comments. Read the whole piece here. It’s short.

Never Miss a Story

Get the latest news and insights once a month from GCI Canada.

A couple of things. First, the algorithms by nature are heuristic so the odds of this continuing are slim. Over time they will be refined. Second, let’s circle back to where I misleadingly said I would take a second to get back to thoughts on measurement of engagement.

Engagement as a macro-evaluation tool isn’t good enough. Depending on the platform, generally “engagement” means likes, comments and shares. We need to stop rolling them together. Comments are king now—more important than ever. Emphasis on likes is inflating engagement disproportionately.

So, how do we evaluate our content’s performance with this definition of engagement in mind? Here are some good places to start:

Likes to Reach | Shares to Reach | Comments to Reach

When we look at influencer marketing, it’s time to measure their performance based on investment. Theoretically this could also be undertaken from a brand perspective as well, but determining total cost might be a rabbit hole in accounting.

Dollars per like | Dollars per share | Dollars per comment

These numbers become very interesting campaign by campaign, influencer by influencer and post by post.  Don’t get me started on the sentiment surrounding comments. I’ll save that for another time.

We’re at a stage that we need to give more precision to measurement around engagement. The changes to the algorithms demand it and a macro rollup of engagement buries the important details surrounding engagement.