DD - Our Development Process
Hey everyone! Today I’m going to change it up a bit and not talk about our development process; because this week’s episode of In the Loupe is on social strategy, today we’ll be talking about one of the most discussed (but often not understood) social topics – “the algorithm.”
Chances are that if you’ve spent any time on Facebook or Instagram, you’ve heard of it. Both Facebook and Instagram stopped showing posts chronologically years ago, and now use a machine learning algorithm to try to show you the posts that it thinks you’ll enjoy the most. This means that social media experts or marketing experts like our own Katie Kinlaw can use this to their advantage and maximize the effectiveness of social strategies. But first, we have to know – how does this mysterious algorithm even work?
In the most basic terms, machine learning algorithms are just complex pattern finders. They work over massively large sets of data (millions or even billions of data points). Developers writing the algorithm will specify several factors that they think will contribute to predicting the outcome that they want, and the computer will do some crazy math to try and find patterns that it can use to predict future results. In the case of Facebook and Instagram, this essentially means that they use your previous interaction to show you new posts that they think you’ll enjoy.
There are a lot of different machine learning algorithms that you can use to try and make these predictions (if you’ve heard of neural networks, those are the most popular right now), but what they all have in common is that the developers pick which criteria they want to focus on, and how heavily they want to weight these criteria. How developers select what is important to them can make a shockingly big difference in how the algorithm performs. For instance, Facebook used to prioritize posts from friends and family, as well as content that it deemed to be informative or entertaining. However, in 2018, they announced that they would be focusing more heavily on posts that get high levels of engagement. As a result, a study in 2019 showed that engagement on Facebook had gone up a whole 50% because of the change, but it also found that the best performing content was controversial, which led to an increase in divisiveness on the platform.
What this means for you is that if you want your social strategy to be effective, it’s important to pay attention to what factors contribute to being ranked highly by their own algorithm. It’s also important to remember that the criteria for being successful are changing all the time, so you’ll need to adjust accordingly!