Facebook’s latest algorithm changes came into view last Friday. Posted late in the evening UK time, the social network said it was going to using data from a tiny sample of users via surveys to help determine what everyone else saw in their feed.
The idea, claims Facebook, is to make the news feed ‘fundamentally human.’ Basing big decisions on the feedback of around 1,000 users via a survey when Facebook has such a vast volume of user data at its disposal seems a very odd decision, and Facebook runs the risk of making the mistake so many publishers have made over the years.
We’ve all worked in newsrooms where audience research has concluded that people don’t want to read about crime, or that football dominates the paper too much. Yet newspaper sales managers will tell you newspaper sales spikes are pretty guaranteed thanks to those two factors. More recently, live audience data from online platforms also tells you that spikes in readership are guaranteed around crime and football.
The conclusion? That what readers say they want isn’t always backed up by their actual reading habits. Online data has helped newspaper newsrooms adapt and change, but also confirmed that readers often ask for what they think they should ask for, rather than what they actually consume.
A degree of nuance is needed when balancing data from surveys with real-world data. Whether Facebook has got that balance right based on this latest change is certainly open to question.
On one hand, it has built its feed in the past around signals based on things like how much people interact with posts – if you liked a post from a person or brand frequently, it was more likely to appear near the top – and how long you spent on the pages you visited from links posted on Facebook. In short, it has the sort of insight into user engagement many brands would give their right arms for.
Putting the views of 1,000 users – out of the 1.59bn users it is reported to have globally – on a par with that level of real user data is certainly to have a lot of faith in those 1,000 users.
But what does it mean for publishers?
Facebook’s advice to publishers is typically vague. It says:
The impact of these changes on a story’s distribution will vary depending on the composition of your audience and your posting activity. In general this update should not impact reach or referral traffic meaningfully for the majority of Pages; however, some Pages may see some increases in referral traffic, and some Pages may see some declines in referral traffic. Pages might see some declines in referral traffic if the rate at which their stories are clicked on does not match how much people report wanting to see those stories near the top of their News Feed. This update helps rebalance those two factors, so people are seeing relevant stories to them.
Impossible to fathom, and it’s ironic that Facebook makes a point of clamping down on people who try to ‘game’ the feed – think those brands which used to say ‘click like if you agree, or comment if you don’t which were outlawed on Facebook last year – yet creates a culture of second-guessing by being as clear as custard with its announcements.
So we can only go with what we’ve seen so far. And to be fair to Facebook, data I’ve seen shows some brands seeing a lift in Facebook referrals, and some seeing declines. But why?
Based on a representative sample of one – me – links shared by friends are more likely to appear near the top of my feed than before. Posts on brand pages I’ve liked – such as newspaper fan pages – are highly unlikely to appear near the top of the feed, but page posts do appear to bunch up midway down the feed.
But at certain times of day, especially on mobile, page posts appear to be far more visible on the feed. And posts made in Facebook groups – a near-forgotten feature of Facebook for many – are suddenly getting a lot of attention.
It would be a giant leap to second guess the changes Facebook has made based on what I’ve seen, but I’ll give it a go anyway. People have told Facebook they value what their friends share more than what brand pages share. If this is true, it brings into question Facebook’s role and responsibilities around helping ‘important’ news get going on its feeds.
You could argue Facebook has an ethical responsibility to ensure that news is given the chance to breathe, grow and reach an audience on its network, given how much traffic it sends to news platforms. That’s certainly a recurring theme at Online News Association conferences I have attended. But Facebook will ultimately argue its job is to keep shareholders happy by keeping users on its platform, and making money off them as a result.
So the challenge to get stories seen by readers on Facebook sits firmly with journalists. The latest algorithm change has confirmed what many have known for a long time: Facebook pages for news brands are just one part of a successful social media strategy.
To be successful on Facebook, journalists need to be good at getting people to share links themselves. That means getting readers on to stories and thinking to themselves: “I want to share this with my friends.”
I’m still seeing lots of content from brands I work with and brands I follow – it’s just the links are coming from friends, or in groups I’m a member of.
Stories which prompt a reaction are the ones people want to share. That can be a positive or negative reaction – shock at the news of a death, or a smile at a listicle about life in the 1990s – but Facebook’s 1,000 panellists have effectively laid down this challenge to journalists: If you want us to see your stories, make sure our friends are sharing them.
Social media should never be seen as a broadcast outlet at the end of the story process. Social media needs to sit at the heart of every content decision. Paying attention to audience reactions through audience data has just become even more important to newsrooms everywhere.