Interview: The Science Behind Brandwatch Search With Aykut Firat
By Phill AgnewSep 28
Combining high-quality mobile survey technology, a robust polling methodology, and expert data analysis,
our bulletins will be essential reading to get the pulse of the nation
Published January 2nd 2019
The Facebook algorithm, the set of calculations Facebook uses to decide what content you see, has a lot of sway and influence. Long gone are the days of a chronological feed, today you get what you’re given.
Publishers, from the New Yorker to the local paper got a lot of traffic from Facebook, but so did a lot of ‘fake news’. In this case, fake news refers to the intentionally untrue articles, often about politics, that aimed to inflame people’s sensibilities (or pander to them) to get the piece widely shared.
As this CNN story notes, fake news has become its very own export commodity.
Facebook became riddled with the stuff and soon became associated with it. Then came the whole Russian ad scandal which suggested the US Election was heavily influenced by the Kremlin through targeted Facebook ads.
Facebook’s reputation in the world was taking a battering.
Facebook decided to take action to curb fake news on the platform, regain some trust, and reaffirm Facebook as a platform for community and ‘meaningful social interaction‘.
Great! No more articles about Bernie Sanders ushering in an age of Satanic darkness that will last a millennia clogging up my newsfeed!
Unfortunately for all the businesses and publishers who posted their content on the site, they got caught up in this too. Facebook also served their content a lot less, favouring native posts from users, particularly if people engaged with them.
That meant a huge drop in traffic. For those relying on ad revenue they took a big financial hit. Businesses saw their referral traffic drop which made getting new leads from Facebook near impossible, while those who sold products through the site were suddenly struggling.
I mean, a cynical person might say this was a way of pushing Facebook ads, but I’m far too busy looking at my Uncle’s holiday snaps on Facebook to suggest such a thing.
So, let’s get into the nitty-gritty of what Facebook actually prefers now. Everything here is based off a webinar from Facebook, so huge thanks to Matt Navarra for publishing the slides on his Twitter. We’ve put them all the ones related to the algorithm in this Google doc here.
The aim of the webinar was to help explain the algorithm, and also explain Crowdtangle, a Facebook acquisition that helps track the performance of content. If you just want the Crowdtangle slides, you can find them here (thanks to Derek Silverman).
Facebook’s news feed and how it’s ordered and presented is based on four things: inventory, signals, predictions, and score.
This is how Facebook refers to all the available content on Facebook, whether it’s posts from your friends, family, groups you’ve joined, or pages you’ve liked.
These are what Facebook uses to help choose which content goes out. It includes a list of criteria. We’ve listed all the ones we know here, with the bolded ones being more heavily-weighted by Facebook.
As you can see Facebook is all about proper engagement from people, such as comments and shares on Messenger.
It’s also worth noting that Facebook will give more weight to conversations between people than they will to those between a page and a person.
This is where Facebook uses your profile and previous behaviours to decide what to show you. They attempt to work out how likely you are into like or interact with content, keeping stuff they think you won’t engage with out of your timeline. Although it’s worth noting a Pew Research Center study found 27% of people think the algorithm gets them wrong (make of that what you will).
This refers to a value assigned to a piece of content that will determine its ‘relevancy’ to the user. The higher the score, the more likely it will appear in the feed. Obviously this means content will get different scores for individual people.
Now that the content has a score based on on things like engagements, relevancy, and signals, Facebook uses that to order its feed for each individual user.
Obviously this score will be unique for each person based on their own habits and interests, plus how their connections interact with the content before they see it.
This means there’s a lot going on behind scenes when you see a link to an article or a status update from your uncle. In turn, as you interact with these items in your feed, Facebook will use this data in the future to continuously tailor the content you see on the platform.
A friend of yours posts an article from Wired about a new piece of technology that enables cats to understand the human voice. The add their own thoughts in their status, saying they’re so glad as they’ve always wanted to be able to talk to their cat.
It catches the interest of some of your other friends. They like the post, or leave reactions, and add their own comments, ranging from agreement to outright hostility to the idea. Debates continue in threads attached to these comments.
Finally, your Dad, who knows you’ve always had a sneaking suspicion the family cat hates you, sends you a link to the same article saying this will solve your paranoid issues related to said cat.
Facebook is going to see this as ticking a lot of boxes for you. Your friends are having in-depth conversations on a post, plus a family member has even sent you the article from that post separately.
Facebook is likely to think this is something you’ll interact with, so now your friend’s post appears at the top of your feed all day, buoyed as more people comment.
Eventually you finally comment saying how glad you are as you’ll finally be able to prove everyone wrong and that the family cat does have it in for you.
A page you haven’t interacted with in five years, and which doesn’t have a profile picture or filled out details, posts a blog post announcing they’ve hired a new IT manager.
No one comments or likes the status except another page somehow related to the company.
Out of the 8k people who like the page, just 10 people click on the link.
Unsurprisingly Facebook is not going to give this a high score for you and it’s unlikely this post will ever appear in your feed.
It might be quite clear by now that this algorithm makes it far tougher to reach people. Posts made from pages will struggle to break through organically, and this has been noted across the board.
Essentially, unless you have a very engaged audience most of your posts will just be pushed off into the void, with just a smattering of you followers seeing it. Without people liking, commenting on, and sharing your posts, you won’t get far and, with less people seeing it, that just becomes harder.
In other words, Facebook doesn’t think pages are as likely to generate genuine engagement between people, which to be fair, is probably right.
It’s not the end of the world though. Facebook has offered some ways to help get your content seen.
First of all, you can just pay for some Facebook ads to promote your content. You can get really targeted and be sure of reaching the right people.
Of course, that involves extra budget, which you might not have. Plus there are ways to increaser your post’s reach without getting the wallet out.
If people start having conversations under your post then Facebook is going to give it more weight. So putting out content that encourages this is key.
In our earlier example, not that many people are going to have in-depth conversations about a new IT manager. So what does get people talking?
Prompting users can be a good start. Ask questions, whether in the copy accompanying your post, or in the post title.
Put out something unique and interesting. Covering the same ground as thousands of other pages is unlikely to spark much conversation. What new angle can you bring to it?
Videos are given more weighting by Facebook as well, so look to include video too. It engages people, is more visual, and can often get people talking.
Telling a good story can work well too. And then, there’s always controversy, but it goes without saying to be careful. Consider whether embroiling yourself in a scandal is worth the reach.
Facebook has also been pushing their group functionality. It fits in with their meaningful engagement idea as they’re often formed around specific interests and they generate a lot of conversation.
Plenty of companies have done this, looking for an interest or topic that’s related to what they do. An example we’ve mentioned on the blog before is Ahrefs, who sell an SEO tool. They’ve created a Facebook group where people can discuss SEO and ask for advice.
Have a think about whether this is something you can replicate for your company. You could even create a few different group. Then you can mix up your own and other’s content to get eyes on your work.
If you use Facebook, you’ll recognise some of the above. They were previously very effective ways of getting organic reach for your posts.
For example, tag baiting would be a post asking you to tag your friends, or encouraging you too. It might have some characters from a film and tell you to tag your mate who you think is most like them.
This obviously encouraged lots of engagement and conversations, but the new algorithm actively demotes posts like this, and will likely demote pages entirely if they engage in it a lot.
Essentially Facebook has decided this engagement is not ‘meaningful’ enough.
Occasionally this stuff still shows up in timelines, but don’t be fooled. It won’t last long so don’t be tempted to replicate it.
As noted, when you post is important too. Posting at the same time as when your followers are online will increase the chances of them seeing it. You can find this data in Facebook Insights, under ‘Posts’.
It’s also important to see how your posts perform and compare them over time. You can do this with Brandwatch Analytics, setting up a dashboard quickly and easily to see how your page has done over long periods of times.
You can then look at your top performing content, what your followers are like, and how many likes your page is getting.
This area is also a great place to see which of your posts perform best among your followers. Using the information we have so far, such as conversations helping your post’s reach to increase, you can look for what generated more interaction.
This can then inform your future strategy, combining the right time to post and the right content to post.
A good way to get started is to set some benchmarks around likes, comments, and the rest, which you can use to judge your performance going forward.
Combining high-quality mobile survey technology, a robust polling methodology, and expert data analysis, our bulletins will be essential reading to get the pulse of the nation.