How the Facebook algorithm works outside of social media

March 16, 2017      Kevin Schulman, Founder, DonorVoice and DVCanvass

The DMA in UK released data yesterday that 91% of all marketers don’t think their emails are relevant to their audience all the time and 42% say that, at best, only some of their emails are relevant.

This put me in mind of one concept from social media that is worth borrowing for direct marketing, even the offline world.*

It’s from the original Facebook algorithm – Edgerank.**  The base of EdgeRank is fairly easy; it’s three factors:

  • Affinity: How close the person creating the content is to the person receiving it.
  • Weight: How much the post has been interacted with it, with deeper interactions counting more
  • Time decay: How long it has been since it has been posted.

These interactions were multiplied together and summed, roughly.

A key measure of affinity is what percentage of posts a person interacts with.  If someone likes 1% of your posts, for example, they are far less likely to see a post from you than someone who interacts with 40% of your posts.

The trick is that light, easy-to-interact-with posts garner the most interactions.  A post that has a picture of a kitten with a note that says “like this post if you like kittens” is going to get far more interactions than most posts about the impact of global warming on poverty rates in Madagascar.

Thus, there are two warring impulses: one to deliver things your audience wants to interact with and the other is to deliver things you need your audience to interact with. Or, put another way with my limited art skills:

One can think of this like a piggy bank of support where you put coins in when you have a popular post and take them out to invest in things you need.

What if you thought of your non-social-media communications the same way: that you are building goodwill with relevant messages and losing it with irrelevant ones (even with the best of intentions)?

First, to get engaging and important posts – the genius of the and — the smart social media manager makes sure all the posts about the Austin 5K are geolocated to just the greater Austin area so the interaction scores on an “ask” post are still high.  Likewise, a smart emailer makes sure that their advocacy email goes out to only people who have an interest in advocacy and/or the issue area.  In an ideal world, the person would have even opted in to advocacy or to the issue and that would be something you play back to them in the communication.

But larger than that, you can’t volume your way into being interesting.  There are some who would argue that the way to find someone’s interests is to cast a wider, broader net.  That is, if you send more emails or more mail pieces, you will be more likely to find something of relevance to each type of donor.  This algorithm and way of thinking would say the opposite – that every time you send an ill-timed or ill-considered communication, it draws from your bank of goodwill.

This latter view is backed up by the research.  Donors get irritated with the frequency of appeals.  So most revenues from a new communication (in a mature program) are cannibalized from existing communications.

Thus, volume isn’t the solution; relevance is.

The implications of the EdgeRank algorithm are also like what we see in direct marketing.  As you lose relevance, you lose audience and you lose the ability to reach your audience with the messages that matter.  The difference is that, unlike with Facebook, all variables are under our control.

So to the nine percent of organizations who have only relevant emails, kudos.  Let’s see what we can do to get those numbers up together.

 

* No, we aren’t wholesale recommending social media for nonprofits.  We’ve talked before about the value of social media to fundraising: how public pledges can decrease the likelihood of someone acting according to their pledge and your more committed donors are more likely to follow you on Facebook and Twitter (it’s not social media making them more valuable) .

** The algorithm has been altered over time significantly.  There are now significant machine learning components baked in that help with spam detection and bias toward quality content.  Additionally, now users can prioritize their News Feeds themselves.  Finally, because of the sheer amount of content available, the organic reach of an average post is single digit percentages or below, meaning that if you have 100,000 likes, maybe 2,000 people will see your average post.