This technology would give us a much better chance of keeping up with the rate of change – and will reduce the chances of AI simply running away from us lower beings.
And it’s not just the 100 wpm typing that’s got me excited.
Because the one near-guarantee with tech advancement is that if something works, it’s likely to get better as more people invest in it.
So our ink to the machine would only be likely to improve over time – allowing us to embed ourselves further and further into our digital selves.
It’s important to remember – we are already cyborgs. Our phones have become real extensions of ourselves. Improving that connection will help us to remain relevant in an increasingly digitised world (and universe).
This might sound like hyperbole – but, as with all these things, just take a look at what has already been accomplished in the last 20 years. The world looks entirely different. That change is going to accelerate. So get ready to plug in!
Fake news has come to take on a few different meanings in the media, so it might be helpful to outline what it is before we talk about whether Facebook can beat it:
Fake news is entirely fictional, fabricated stories, made with the aim of generating clicks and, therefore, advertising revenue
The authors of this nonsense use particularly sensational headlines to attract viewers and shares to boost their traffic
They also employ ‘bots’ – or fictional Facebook profiles – to engage with their fake content to boost a story’s weight in the algorithm and therefore increase traffic
Stories can then take on a life of their own because of the echo-chamber effect Facebook’s algorithm has on the things we all see on there
Fake news is NOT when CNN reports something Donald Trump doesn’t like
Fake news is NOT the BBC
Fake news is most certainly NOT schmocial.org – here you get nothing but the truth and Rick and Morty gifs. Very little else.
It’s also worth noting that fake news is not helpful for Facebook’s business model. Facebook is so successful because it’s a network based on real life. It requires real names and real engagement. If people stop trusting what they see in news feed, the whole system could start to break down.
So how does Facebook plan to fight it?
Well in a blog post, the network outlined three areas:
Disrupting economic incentives
Building new products
Helping people make more informed decisions
To me, the second of these is the most compelling, because it’s the area that requires the innovation and the fundamental changes to the way Facebook operates.
It’s difficult to see how disrupting economic incentives will achieve much, as the fake news ad revenue is likely coming through different streams. It’s true that revenue share is available to publishers through the Facebook audience network – but this is by no means the only option for publishers looking to make click money.
Then helping people make informed decisions is a laudable aim – and involves some interesting projects, like the Facebook Journalism Project and the News Integrity Initiative. But these are broader, longer-term changes.
I will also say here that it’s encouraging to see Facebook taking news organisations seriously. For a while it looked like they’d be treated simply as more paying advertisers and would thus be strangled off of all news feed traffic.
So what new products are they building?
One of the things they are doing is making reporting of dodgy news easier.
This might sound like an easy solution. People report fake news, Facebook removes it. But it’s not that simple.
The main problem with fake news is that people don’t know it’s fake.
For example: a really fake story might have a hundred shares from credulous readers, and four or five reports from angry skeptics.
Then, a true story might have hundreds of shares simply from interested people, and four or five reports from those who, say, don’t like it politically.
How to tell the difference? This is what Facebook will be struggling with.
It’s a numbers game
Another issue Facebook is dealing with is scale.
In most programming (and I should emphasise here that I’m a basic-level programmer, certainly not an expert) it’s relatively easy to perform operations on known numbers – it’s the unknowns that become tricky.
So, one, two, three, or a hundred fake news stories are easy enough to take down one by one. But what about n fake news stories? This is what you are dealing with when they are constantly being uploaded all day every day.
And it requires systems that can understand an incredibly complex range of signals.
One of the signals they are looking at, for example, is a person’s likelihood to share a story after reading it. They reason that fake news is more likely to reveal itself post click-through, so the majority of its shares would come from people who haven’t actually read the story.
Take it from someone who has worked social at two major newspapers: sharing without actually reading an article happens all the time.
So Facebook will now be grappling with that, too. But it’s a task I expect them to do well. After all, writing smart, complex algorithms is what has got them to where they are today. They’ll have some of the best minds on the planet taking on the job.
Working with partners
Working with major news organisations to fact-check stories is another method they are using – and is similar to what Google is doing with its own fact-check roll out.
These are both great ideas and should do a lot of good. My only concern is that part of the problem is that people are coming to trust fake news sites MORE than the mainstream media. So what if they simply don’t care what Associated Press says?
What about the bots?
The bots – or fake profiles sharing fake news – are under constant scrutiny. Facebook looks at the names, the profile pictures, and a huge variety of ways in which they are related to real life profiles to determine which profiles are real and which ones aren’t.
But it’s an arms race. And, as I said above, it’s also a question of scale. It probably isn’t that difficult to find out whether one individual profile is fake. But determining whether an unknown number are fake when compared to an unknown number of real ones – by using a huge number of different variables – is another question entirely.
Building systems that recognise real from fake is tough. And it’s sometimes easy to forget that social media is a brave new world. Many of the systems they use will have to have been invented first.
One thing is for sure – Facebook has a huge investment in trust. And while the search for engagement sometimes pushes towards maximum engagement at the expense of all else – they’ll know that becoming known simply as bullshit merchants is very bad for business. So Zuckerberg and co be working incredibly hard to be as reliable and truthful as possible. I expect them to come out on top.
Elon Musk thinks we are all already cyborgs. I agree.
And it’s not because I’m a massive Elon Musk fanboy.
OK, it’s not just because I’m a massive Elon Musk fanboy (if you’re reading this, Elon, ilu ♥).
It’s because it’s true – we have all now come to rely heavily on our new tertiary selves – or, our digital brains. Our social media and Google-powered memories and instant communication and GPS and fitness trackers and everything else.
Here’s the man himself talking about it:
Interesting, right?!
Elon focusses mainly on the output problem – essentially that we are just using our meat-stick fingers to type things out very slowly in the digital world.
His idea is to use a neural lace to hardwire our brains into the web to speed things up. This is something that people like Ray Kurzweil have been talking about for some time now.
I can’t wait! Plug me in.
But I think his ideas also raise another couple of interesting points relevant to social media:
What role does social play in the digital self?
If Google is basically our digital memory – what does that make Facebook?
I think it’s our ego. It’s the part of ourselves that’s constantly screaming at the world about the things it likes, the things it hates, the things it wants to change, and the things it just wishes would just please stay the same ..PLEASE GOD STAY THE SAME PLEASE.
At the moment it feels like a hyper-self. Everything we believe to be true about our character, exaggerated and amplified, shouting relentlessly into the void.
It’s true that we probably communicate more, in terms of sheer volume, with our friends and colleagues on messaging apps than we do in any other way – but I still reckon we only really use the full range of our expressive communication when we’re talking to people in person.
So in many ways social is overtaking traditional person-to-person communication – but I still think there are some things it can’t do.
2. What about input?
Musk says input isn’t such an issue because we can take in lots of data through our eyes (or high-bandwidth visual interface, natch).
This has got me round to thinking about content in terms of data – ie. how much information a video can upload to the viewer’s brain in a given amount of time.
This may be the reason Facebook is now shifting back towards having sound on its videos (rather than mute by default) – because with audio and images, you can push more information per second than with visuals alone.
Which brings me neatly to speed reading. Cool films like this one:
I think stuff like this is great because it is offering you more information per second than a film with normal left-to-right text would. And certainly more than one with no text.
This is, I think, part of the reason emojis have become so popular: they allow us to communicate quite complex facial expressions in a single character. This is actually conveying a great deal of information – far more than can be expressed succinctly with words.
This trend is bound to continue as we get better at taking all the information in. Most videos on social now have some sort of text overlaid.
And every time we watch them we are practising absorbing more information, faster.
Then as we become better at taking it all in, we get increasingly bored with films that don’t satiate our voracious informational appetite. Our attention span becomes shorter.
Want proof? Simply watch a few old movies. They may have other qualities you appreciate – and you might personally enjoy the time taken over setting the scene.
But the trend between them and what we have now is clear – everything is speeding up and becoming more info-heavy.
Here are a few other examples…
LEFFE
We tried to play on this information overload with films for Leffe, whose creative platform was “Rediscover Time”.
The whole idea was that Leffe is a beer that should be savoured by slowing down and taking time to appreciate the finer things in life. And we had made a ten-minute documentary film about the lives and work of people who use time in their work (an astrophotographer being one of them).
Ten minutes is too long for a social film, and the premise was somewhat at odds with the points I’ve outlined above. So we tried to come up with a creative solution that was direct, and challenging to the viewer (although hopefully not too challenging). It worked – the video wound up having one of the highest view-through rates they’d ever seen. We also saw a high rate of people clicking through to watch the full documentary.
LEAP SECOND
Before I joined, the clever folks at my current agency created this very short film to celebrate the leap second that occurred in 2016.
I think it’s beautiful – but I also find it interesting how it is communicating far more information than it’s possible for a human to take in in just a second. It’s taking the idea of high human-bandwidth video to its natural conclusion (i.e. the eventual failure of our own processing power).
NIKE
Another great example of speed reading being used to satirise digital culture – in an incredibly digitally-native format. It’s probably ironic or something.
Working in social is fun. You’re at the bleeding edge of the current global conversation, which can be exhilarating.
And one of the best parts of the job is talking to and advising people from different areas of business on what they can and can’t do.
These conversations can also become tricky when impossible requests come in.
I never say never, and will always try to find solutions for people’s requests, but I thought I’d share some of the most common ones. It’s just for a laugh – but I reckon most people reading this will have made or received at least one of these at some point…
1.”Can you do a viral?”
This request probably had its heyday around about 2014 and is without doubt my favourite – because it’s almost axiomatic that you can’t.
The whole point of viral content is it just takes off unpredictably when that perfect combination of elements suddenly falls into place. Creativity, influence, news, timing.
Obviously I’ve tried, with varying degrees of success, but the idea that anyone who knows anything about social media can just conjure up a viral on the spot is, sadly, wide of the mark.
2. “What do women/men aged between x and y like?”
This could apply to any different demographic, but contains the one running problem whichever – ie. they like LOADS of different stuff.
Hell, guess what? Some of them aren’t even on social media! (I know right).
And all of that is before you’ve even come on to analytics. And obviously there aren’t any analytics that have ALL THE DATA IN THE WORLD.
I mean maybe that’s what god is? Just shitloads of data on everything?
Either way. If that is what god is, she doesn’t exist yet. So we can’t run that request either. Sorry guys.
3. “How do I make Lady Gaga/Ricky Gervais/Somebody Famous/This Random Journalist follow my account on Twitter?”
I mean.
You could make friends with them and ask? But I won’t be able to do that for you.
4. “Can you get these Instagrammers to do this stuff for no money?”
If I give you this duster, will you clean my house for no money?
How about if I share one of your Facebook posts? No? Thought not.
5. “How much money do I need to spend on this Facebook advert?”
The problem is that you can literally spend any amount of money on Facebook ads. Even if it was so much money that you physically couldn’t enter the numbers in the box, I’ll bet you could phone up and pay some other way.
This one is a legit question though – the best way to handle it is to look at estimated audience size and give a range of options. But there’s definitely no one magic number.
6. “Can you just run some stats about x?”
Obviously I know exactly what stats you want. Shall I just check how many times they’ve mentioned the word ‘chicken’ on Google Plus? Would that be useful? If not I’m going to need more info pls.
7. “We want to give this cat/dog/cartoon bear an Instagram account. It’ll be hilarious!”
Algorithm changes mean longer videos now have a better chance
Was anyone else starting to notice Facebook videos were all getting a bit samey?
For a long time now, best practice for creating Facebook videos has been to make them work with no sound, and at a length of about 15 seconds.
The reason for this was that Facebook was measuring the success of a video by ‘percent completion’ – or the proportion of it being watched by most people.
A video watched to 75% of its length was deemed ‘better’ than one watched to 25%. And the better ones were, of course, being shown in news feed more often.
This sounds like a reasonable thing to do, until you realise that the most successful videos were almost always the shortest ones. Because it’s much easier to get your audience to watch most of a 15 second video than most of a 15 minute video. It’s a simple question of time. Facebook has a very transient audience who tend to flick through content quickly.
Thirty seconds is a long time on Facebook newsfeed. Fifteen minutes is a lifetime.
But now that’s all changed
Facebook now say in a blog post that ‘percent completion’ will be more important on longer videos. So now if your audience watches 75% of that 15 minute video, it will be deemed better than watching 75% of the 15 second one.
Here’s what they said:
“…we know that completing a longer video is a bigger commitment than completing a shorter one.
“As we continue to understand how our community consumes video, we’ve realized that we should therefore weight percent completion more heavily the longer a video is, to avoid penalizing longer videos.”
Anything else?
You may also have noticed, as I have, that Facebook is making the sound on/off button more prominent on videos.
Look at that big button
It’s been assumed for a while now that nobody will ever listen to sound on video content – this is obviously their play to try to change that.
It would make sense – sound adds a whole different dimension to videos, and bringing it back would certainly make Facebook vids more interesting.
So what does that mean?
It means longer videos will start to perform well on Facebook.
And I think these changes are a good thing. The whole 15-seconds-no-sound paradigm was definitely starting to get tired.
Whilst it certainly presented an interesting creative challenge for a while, newsfeed video content felt like it was getting dumber.
Allowing content creators to make longer videos – which is essentially what these changes are doing – means our newsfeeds are likely to be filled with more interesting and in-depth stuff.
And those of us that actually make the videos have more time to play with. I can certainly think of a few colleagues who will be delighted.
But…
Don’t go thinking that just because videos can be longer that they should be. Videos will still need to work hard to hold people’s attention – which will arguably be harder the longer it goes on for.
It’s no good making a great 10 minute long video if the first 5 seconds are slow or dull – people will still drop off before they get to the good stuff and the content will still perform badly.
Videos on Facebook have a 1.5 second audition time – and that certainly won’t be changing.
A way to guarantee a certain number of people will see your ad
It has a minimum spend – so not possible with small campaigns
It won’t guarantee any other KPIs like clicks or engagement (although you can still set these as targets)
You don’t always get the best price
Reach and Frequency is a type of campaign Facebook rolled out in 2014 to try to make their network feel more familiar to traditional advertisers – and has some big benefits.
It essentially allows you to guarantee hitting a given number of people with a piece of content – meaning that if you’ve set reach as your KPI for a campaign, you can crack open a cold one and put your feet up. Results are in the bag.
So how does it work?
First, let’s clarify terms ICYMI:
Reach = the number of unique Facebook users your post was seen by. A single user can see a post more than once. (We call the total number of times a post was loaded – by the same person or not – impressions)
Frequency = the number of times a given person has seen a post. So a frequency of one means any one person will see your post only once – two is twice etc.
Setting a good frequency is essentially a balance between making sure a user has actually noticed your post, and bombarding that user with the same post they weren’t interested in first time around. This just annoys people, and setting a high frequency is one of the top things my clients say they don’t want to do (who wants to annoy people, right?).
Set up
The set up of the ads is similar to auction advertising – you can simply choose the reach and frequency option in ad manager right at the start:
From there, setting up interest targeting is exactly the same as auction campaigns, with one key difference.
Once you have defined your audience on a reach and frequency campaign, you get a much better visualisation of your target audience:
I’ve found this to be incredibly useful for helping clients understand how big their potential audience is – and giving them different options in terms of how much to invest.
Facebook can sometimes feel like firing messages out into the void for some clients, and it can be difficult to get to grips with who the audience is and how many of them we can talk to.
This tool can help lay all of the target information out in a much easier-to-understand way. The buying then gives you the added bonus of guaranteeing to hit any percentage of it.
This is probably a good time to mention that there is a minimum spend on this type of campaign, which is based on reaching at least 200,000 people. The actual spend will vary, but it would rarely come in under around £500.
The frequency will also affect how your spend translates into reach – so on a frequency of one you’ll reach more people than you would on a frequency of two. This is because the impressions, which your CPM is based on, will stay the same (or very similar) but lots of people will see the ad twice (due to the higher frequency).
So costs should factor into the balancing act too. A general rule of thumb is that a frequency cap of two will work best for most clients. One ad impression can get missed – but if it’s appeared twice it’s likely to have been noticed. Any more than two times means you start getting into annoying ad territory, and you start getting far less bang for your buck.
If you just want to hit as many people as possible – a frequency of one will do the trick.
Secondary consideration
It’s possible to give a secondary consideration to your campaign, so that all the people you are reaching are encouraged to do something. This might be click-throughs to website or video views or post engagement.
Facebook will use this to serve your ad to people most likely to do the thing you want them to. But it’s important to realise that this isn’t what you’re being charged for.
The ad will be delivered on reach alone – which is, as I said, guaranteed. But if you wanted to generate click-throughs at a specific price per click, say, then this is not the type of ad you should be using.
(That would be an auction campaign)
Downsides
The relationship between set price Reach and Frequency campaigns and the auction bidding campaigns is the same as any set price vs auction. You don’t always get the best price – you simply get a predictable one.
You know how much you’re going to pay and what you’ll get for it. That means poor quality work will still reach lots of people, but great quality work might not get great pick-up for little money. You’re unlikely to get a surprise hit like you can with auction buying – so it kinda takes the fun away a bit.
It’s also worth noting that costs start to increase exponentially once you’ve hit around 50-60% of the audience. This is because some people use Facebook more than others – and once you get closer to total coverage of an audience, you start targeting people who hardly ever log in, who therefore cost more to access. You should never hope to hit 100% of an audience.
Thoughts
These campaigns are super useful – particularly for big ad campaigns where saturating a market is the tactic
They are also really good for talking through targeting with clients who have little experience with Facebook marketing. You can plug in different audiences and give loads of options for accessing them
They work better bigger – you can’t hope to spend any less than about £500
These are just two of many well-reported examples of Trump saying and doing incredible things – and getting away with it. All of which begs the question…
How is he doing it?
The reasons are probably myriad, and include the simple fact that he just hogs the news agenda more and more with every passing outburst. But social media could be playing a role, too.
The point I made was that when we engage with a post, Facebook logs it and remembers to show us more like it in future. This means we tend to close, rather than broaden, our horizons in terms of the things we look at online.
And it could well be that this effect is allowing Donald Trump to get away with far more than he would’ve been able to without it.
What I’m suggesting is that people following, commenting on and engaging with Trump may be indulging in a kind of mutual reassurance that some of his ridiculous ideas actually aren’t all that ridiculous.
And because they tend to see more and more of the same things, they may be exposed to fewer sources of information contradicting it.
It’s certainly true that The Donald has a far greater social media presence than his competitors for the Republican nomination or the Democrats he could face in the Presidential run. His Facebook following is about double his nearest rival, Bernie Sanders, who has about 3.3m followers to Trump’s 6.1m*.
And his engagement is through the roof.
Here’s a content analysis of Trump over the last year (via Buzzsumo)
That’s 23,733 average shares.
And here’s the same for Sanders:
That’s 11,447 average shares on content. Less than half.
So not only are Trump’s supporters seeing more and more of Trump’s own content – the stories about him elsewhere on the web are getting viewed more, because more people are sharing it more often.
But Facebook isn’t just about reading – it’s about talking. Those who are regularly engaging with Trump content online will tend to just be talking to each other.
That’s because people who support, say, Sanders, will be engaging with Sanders’ content and will be having a whole different conversation on his page.
All of this just fuels a situation where people of opposing political views just don’t talk to each other. And without proper debate, people can get away with saying stupid stuff unchecked.
Of course, it’s also true that the rise of social media has seen a reversal of the ‘third way’ politics typified by Blair and Bill Clinton, in favour of more extreme politicians. And this may also be due to Facebook’s echo chamber.
The rise of Jeremy Corbyn here in the UK, for example, was very much driven by an incredibly vocal and active social media community. One which is still campaigning now under the name Momentum.
Welcome to the Momentum Facebook page. Something wonderful happened this summer. After years of being told that there…
The idea that Facebook can win elections would have sounded crazy just a few years ago. Now it seems like a pretty reasonable – if arguable – assertion.
And while we don’t have access to the kind of money the candidates in the US election are spending, it’s a fair bet that it’s not nothing.
Any Facebook marketer will tell you that anyone who can combine sensational content with a big ad spend is onto something huge. With his deep pockets and big mouth, Trump seems to have found the golden combo.
*It should be noted that Sanders actually has two pages of about 3m each, but it seems fair to assume a large proportion of his supporters will be following both, so adding them together wouldn’t be a fair representation.
With Canvas, Facebook has gone one better. They’ve basically created a platform that allows brands to build a microsite within the network itself. It looks a bit like Snapchat, and utilises devices’ gyroscopes to let users to roll through different product images and videos.
Canvas also loads almost instantly because the user stays within Facebook while they’re engaging with it. This is to minimise the drop offs you get with long load times on mobile website click-throughs.
Using it seems pretty straight forward, as it’s accessed through the advert section of Power Editor (not advert set or campaign):
Once you’ve hit ‘+’ to create a new Canvas, you’re faced with options for the main components – a title, a header and a photo.
So far, so simple. But the real fun is in the rest of it. If you hit ‘add component’, you get another list of options, which includes carousels, buttons & photos:
So Canvas ads will likely take much longer, and require much more planning, than any other type of Facebook ad. You’re basically building a mini shop within Facebook, so you need to really think about how it looks and functions.
It seems to me that the best option would really be to sketch out roughly what the journey will be before even starting to build the canvas.
Another thing to note is that, for the first time, you need to preview the ads on your mobile device rather than on desktop. The feature itself is mobile-only, so it makes sense that the preview should be too.
It’s incredibly simple to do. Just hit ‘Preview on Device’, and Facebook sends you a notification to view the ad.
Who will be happy abut Facebook Canvas?
Well, brands, agencies and social media managers should all be pretty excited. If the success of carousel ads is anything to go by, Canvas could really drive up engagement and sales for all sorts of different businesses.
One of the case studies on their blog – ASUS – saw a 42% increase in clicks on the Canvas, and 70% of those who did went on to click through to the website.
Canvas should also be great for creative agencies as digital teams have so many features to play with. The planning and execution of these sorts of campaigns will probably become a crucial part of any content strategy.
Who won’t be happy about Facebook Canvas?
Legacy media, Twitter and old style digital advertisers. This will take yet more money away from standard online banner advertising as it’s quicker, slicker and will likely deliver far better results.
Facebook are getting closer and closer to that mythical, impossible legend – the online advert people actually want to look at.
This means we’ll all soon have the option to report friends we’re worried about. The app, which was developed with Samaritans, will then send that friend a message with support options, like this:
Feedback seems to have been generally positive from what I’ve seen. But it does still raise legitimate concerns about just how far we’re willing to let social networks into our lives.
How will the network be able to use this data, for example? We already know Facebook records everything you ever do, and the vast majority of this data is used in ad targeting. Will we get to a stage where a person’s mental health is fed into the picture Facebook builds of them for advertisers?
I think it’s unlikely. I honestly don’t think Facebook has any ulterior motive in developing this tool. But what if there was a data leak? If you were suicidal, imagine the kind of damage it could cause your mental state knowing that the whole world knew about it.
So there are legitimate concerns. But none, as far as I can tell, are really outside of the broader issues about Facebook becoming more and more integrated in our lives and storing huge amounts of data on each of us.
And anyway, it could actually save lives. And I knew one specific person it genuinely could’ve helped.
In a past life I used to play drums in a rock band. One of the quirkier things about being in a group with a presence on social media is that you start to pick up fans from all over the world. One of ours was a guy we’ll call John (not his real name) from San Francisco.
John found us on MySpace and seemed like a nice guy. He’d regularly get in touch about new demos we’d posted, and all of us would chat to him about our music on Facebook.
He started adding our friends, and other UK-based fans, and catching up with them to ask about how our gigs had been. Then conversations began to get more in-depth. He’d tell us more personal details about his life, and talk to us all as if we’d been lifelong friends.
We actually used to wonder whether it was a wind up. We thought it could’ve been one of our friends pretending to be an international super fan. But it became clear he was just a pretty lonely guy, reaching out for company.
Anyway, John was clearly in a bad way. He’d post strange things, and we started to receive messages from people he knew who were worried about him, asking us about what he’d said to us.
The whole thing was surreal. We liked John, and worried about him, but were completely powerless to do anything to help. We were a bunch of guys from London who didn’t spend all that much time on Facebook anyway, and he was a person we’d never even met who lived in San Francisco. The connection was tenuous to say the least.
Then, tragically, John killed himself.
At least, we think he did. The first any of us in London knew about this troubled man we barely knew taking his own life were the messages of condolence that started to appear on his wall. We didn’t know how to react, but I’m certain we all felt guilty for not doing more to help.
So, yeah, for my money, the button is definitely a good idea. We certainly would have reported some of the updates and messages John posted, and he might still have been around today.