Gaming the Facebook Algorithm... Lessons from our fundraiser

Disclaimer: The “Behind the scenes” blog is not entertaining at all. This is “journal of how we are creating Mirth and Dignity”, written as we go. It’s a purely educational resource (and a place for me to record what I’m thinking… A kind of diary). I’d have killed for this resource five years ago, but unless you’re my particularly narrow kind of weird, this blog is not recommended reading.

Our fundraiser absolutely tanked this year.

Usually, there’s a huge rush in the first day, big rush second day, then a trickle for a while and then a rush on the last day. So, when the first day brought in only a couple of thousand, I knew I was in trouble. Which was really sickening, because it’s not personal trouble. It’s an inability to pay other people for their work. An inability to keep the whole thing functioning. I don’t actually have a plan b. The fundraising has to work.

So I took a deep breath, walked off my anxiety until I could think clearly, and sat down to figure out why on earth this could be tanking so badly. I had a compelling “why”. I had a great story about who we are and what we create. I had fabulous testimonials and lovely graphics. I even tried—at the adamant urging of my communications specialist niece—experimenting with being brief and to the point. Nothing was working.

There is a crucial difference between social media fundraising, and, say, mailing list fundraising or in person fundraising. In social media fundraising, the main task of the post isn’t to be effective at convincing a person to donate. The main task is to get the post in front of people.

On the fourth day, we got a huge bump in donations. Facebook was notifying me of comments as “x number of people answered your question”. This confirmed it. Facebook is (so far as I can tell) pushing questions, right now, which gave the post a bit of a bump. The huge difference in donation numbers solidified that the problem wasn’t how good our ask was. The problem was that people weren't seeing the ask.

With a smaller group, this is easy to solve. You get a small, dedicated group of volunteers, and you coordinate them to all like and share and comment on the posts (I always start a “Friends of x campaign” group a month ahead). This bumps the post up in the Facebook algorithm. I’ve used that trick quite a few very successful fundraisers—but the group had outgrown it. In a group of 140,000, there’s a networking problem. The people I recruit for the “Friends of” group would all be friends with each other. The vast majority of the 140,000 are not my friends. Or even friends of my friends. It’s not enough people, and it’s not a diverse enough group of people (they’re all interconnected). That would only help me reach people I’d probably already reach anyway. The only way to get broader sharing was to ask people directly.

But I knew that would only work once, so I had to make a post that would nail it on the first try.

There are a few things that I know will get engagement. Starting with a question (at least right now). Having an image to go with the words. Outrage gets engagement—that’s universal, and is true in person and online. Righteous outrage is a particular favourite. Also, resentment towards Facebook, and a desire to screw with it. Being able to do something important. I knew I needed to hit a lot of key points for it to go viral.

The post began with the question “Wanna help avenge our mods and screw with Facebook’s algorithm?” The avenge our mods part was both intended to spark curiosity about what the controversy is (everyone loves a controversy) and a desire to get in on the shenanigans. And, to be honest, to spark a bit of outrage. Normally I avoid leaning on outrage, which I see as selling my soul a bit…. But I was in some amount of righteous indignation at the time. I felt like “%(&*ing Facebook makes a lot of money on the back of the work of our mods and was blocking me from getting them any compensation at all”.

I put the hook, then the story, then the request—donate, comment, share etc. And the specific request to comment “first” if this was the first you were seeing of the post. This was motivated primarily by a genuine desire for data, but also an awareness that comments produce algorithm buzz.

For good measure, I added a graphic of my kitten attacking a foot, with the words “Okay Facebook, you have tangled with the wrong humour group!”. To nurture a good old mob mentality and some more curiosity. And of course, a kitten photo is always a good idea.

By my estimation, that post brought in nearly twenty thousand dollars. I did a follow up (“We are vanquishing our robot overlords!” with my cat sitting on the Roomba), that was mostly a thank you. But the vast majority of the 27,000 or so (26,000 or so after fees) was brought in by that single post.

What are the lessons? Well, there are the specific lessons about how to game the algorithm, above. And there is the overarching, gnawing lesson about relying on other peoples’ platforms. About the perils of that. In an overarching sense, we are picking a very precarious middle ground. If we were using Facebook’s inbuilt tools (we can’t because we’re a not for profit, not a charity), we would be able to reach people much more reliably. Facebook is currently really pushing fundraisers and will likely continue to do so, because they offer such valuable data. But there’s a deal-with-the-devil there… That approach intensifies our dependence on Facebook.

In the long term, we will need to have a two pronged approach—making use of the Facebook group to reach people who are very lightly appreciative of what we’re doing and hard to reach otherwise…. And also having a smaller group of people who really believe in what we’re doing that we can access directly through email.

I think. Like everything from this blog, it’s a work in progress.