FACT: Countering Misinformation in the Media
Continuing Professional Development (CPD) points: 2
This podcast will introduce you to the fundamentals of the Rapid Response Unit’s (RRU) FACT model (Find, Assess, Create, and Target), which has been designed to help tackle misinformation.
Listen to the podcast
- Level: introductory
- Presenter: Dr. Oliver Marsh
Welcome to Government Communications Service (GCS) Accelerate essentials learning series. Accelerate is the GCS digital skills and cultural transformation programme which aims to raise the digital standards across all areas of the profession. For our first module, we will be focusing on FACT – the model used for countering misinformation in the media. If you’d like to aid your learning visually, you will find a copy of the slide deck to support this training session on the GCS website. Happy listening!
Welcome to GCS Accelerate FACT training and I’m Oliver Marsh, I’m the senior data analyst in the Rapid Response Unit and I helped create the FACT model. FACT stands for Find, Assess, Create and Target, and over the course of this recording, we’ll go over all of those in more detail.
But what you’ll find is that finding content, assessing whether it’s a problem, and creating and targeting responses is possibly something you do already in perhaps a press office or a campaigns capacity. But this model has been specifically developed for the complicated and confusing world of social media where people can still get confused as to what counts as ‘trending’ or whether something really warrants a response.
So this recording is designed to help you know about the tools, techniques and the approaches to deal with finding, creating, assessing and targeting on social media and search, and online tools more generally. I’ll be discussing some broad approaches and not specifics of tools but there is a handout on the GCS website which you can consult alongside this recording.
A final thing to say before we go into the details is that a lot of what I’m going to say might sound quite obvious, but FACT has been developed to help you in a crisis situation when sometimes you are being pressured to do something quickly or perhaps you’re not quite sure where news is coming from or you need to try new things.
So this model is designed to make sure you can stay rigorous, on top of the information and making good decisions even in those situations. But it also requires you to practice, record and monitor outside of crisis situations and I’ll be helping you develop some exercises where you can do that and providing you with examples of where we’ve used FACT successfully.
So with that in mind, let’s start with a couple of case studies that the Rapid Response Unit has used the FACT model to work through and then we’ll go through each of the stages of FACT in turn.
Before we go into Find, Assess, Create and Target in more detail, I’ll give a couple of case studies of how the Rapid Response Unit has used this process end to end and that will give us some examples we can return to throughout this recording.
The first example comes from a time we worked with DFID to rebut some claims around aid sent to India. We found an article from the Daily Mail, claiming that the UK government had given a billion pounds aid to India to help build a 330 million pound statue, was going viral on Facebook and as you may be able to guess in the headline, the article contained false information about how UK aid had been spent. We assessed that the scale of how viral this article was going on Facebook warranted a response.
We contacted DFID and they produced a GOV.UK blog rebutting the claims in the article; however, it wasn’t getting much traction. Now because the article had originally been trending on Facebook, we decided to make sure that DFID’s response was also present on Facebook, so we create an animation explaining the positive benefits of UK aid to India and how this was a mutually beneficial partnership for both countries.
As the original Daily Mail article had gone viral on Facebook, we decided this was the audience that needed to receive the information, so we shared the content on the UK Government and DFID Facebook pages. The content also included the link to DFID’s blog, but we also made sure that people casually scrolling on Facebook would still get the message.
So, in this case, we found an original article, we assessed that it warranted a response because of the false claims, we created content and targeted it appropriately to ensure that people who may have heard these claims would receive a counter-narrative.
A second example deals with the thorny subject of vaccine conspiracy theories. In February 2019 when we were monitoring Twitter, we observed an increase in references to anti-vaccine conspiracy theories, such as ideas that vaccines cause autism. We communicated these findings to colleagues in DHSC and PHE and encouraged PHE to share existing content around the availability and effectiveness of vaccines.
When we did a closer analysis, we discover that all of these tweets or all of these tweets were not that worrying: they were jokes about anti-vaccine conspiracy theories. But because we shared the content while the Twitter conversation was focusing on the question of vaccines, it meant that we were joining in a conversation that was already present and PHE’s content got more ‘eyeballs’ as a result. So people knew what vaccines were available for children and where to get them.
Since this was a very specific audience, parents, we also paid for Facebook advertising to this audience for the PHE content, again maximising the number of ‘eyeballs’. So here, we found an opportunity to join a conversation online that was relevant to the government – a health conversation – we assessed that both Twitter and Facebook were both good ways of reaching people who were talking about vaccines, we didn’t create content because PHE already had some, but we did target it at the relevant audience. So, even though there wasn’t necessarily much to worry about, it was good communications in general – increasing the number of eyeballs on relevant government content.
A final example deals specifically with search and it relates to D-Day. We were monitoring social media and search discussion in the lead-up to the 2019 D-Day commemorations, partly due to the visit of the President of the United States. We found that the question, ‘What is D-Day?’ was a trending UK Google Search.
We worked with the Ministry of Defence press office to produce and disseminate an explainer video that went right back to the basics of what D-Day is to respond to this trend. So this was very straightforward, just noticing a Google trend, found the resources to address it and very quickly creating relevant content, and targeting it widely as plenty of people were searching for this question.
So these are three quite different examples involving rebuttal and non-rebuttal, but they have the same ultimate aim of finding ways of getting government messages out to a wider audience based on good rigorous analysis of the online environment.
And now we’ll use the FACT model to show you how you can do that analysis and use it to improve your comms.
So we’ll begin with how you find topics and trends that you might want to respond to. We define ‘Find’ as “constantly monitor news sources and publicly available social media posts to identify themes, discussions and stories of potential concern to government”.
If you work in comms, you’re probably used to finding stories and sources whether through newspapers or focus groups or polls, but the online environment throws up some new challenges. It can be quite noisy, crosses national boundaries, there are plenty of unexpected sources that you may not be aware of. You need to be aware of the different platforms, the different sources of data and the different tools available to you.
You also need to understand how people use the platforms: how is Facebook different to Twitter and different to Google, for the sorts of subjects you look at. Now we won’t go into detail across all of these because it would be a very long recording but there is a handout on the GS website that will help you. But a few broad things you need to think about.
Firstly, you need to think beyond what is often called ‘The Westminster Bubble’. So, for instance, if you are trying to find stories about knife crime, don’t go to Twitter and search for ‘serious violence strategy’. You will need to know the terms the people you’re interested in might want to use. And on that note, you’ll want to know something called ‘Boolean’.
This is a kind of language that helps you search for specific terms and phrases on social media. So, in the vaccine case study I mentioned earlier, we couldn’t just search for references to vaccines because that would bring in all the content from America. We needed to make sure that people were talking about vaccines specifically in relation to the UK or the NHS or something else relevant to our government.
There’s plenty of resources on how to do Boolean well, in fact, we’ve created some in the Rapid Response Unit that you can access, but it’s important to note that you can only learn through practice and working with team members to really refine your skills and your searches. Make sure you’re practising ‘Boolean’ as early as you can, don’t wait for a crisis to get good at it.
Another broad thing you’ll need to consider is when you’re searching social media or search data, there’s always a trade-off. Again, if you’re looking at something like vaccines, we’d have looked at something very specifically at vaccines in relation to the Health Secretary, for instance. But then we’d missed lots of people talking about vaccines who weren’t also talking about the Health Secretary which is most people. But then if you make your search less specific, you get lots of noise to wade through. There is no right answer, you just need to think about what is best for you and your objectives.
And the final bit of advice is always collaborate. For instance, collaborate with the media monitoring unit, they produce social headlines which you can subscribe to if your department is an MMU subscriber which will give you some really useful read-outs of the online environment that you can use, though you will need to bring your own knowledge of the subject area to really do it properly.
You can also collaborate within your team to help each other make better searches, write better boolean, understand how Facebook and Twitter work in your subject. And – and here we’re starting to move into assessing – you’ll also need to make sure that you’re watching the online environment a lot, not just when you think you’ve found something so you can pick up anything unexpected and get a sense of what is normal. But we’ll move on to that now.
Once you’ve found your content or trend you need to decide whether to respond to it. After all, you can’t respond to every bad tweet about the government. In fact, we define Assess as “determine the scale of engagement with the risk identified and establish if it is appropriate to respond to the content. Flag to relevant press offices and special advisors, with a recommended approach to response”. In short, do you need to do something about what you found and if so, who needs to know about it?
Now again, assessment is something you’re probably already doing as a press officer or a campaign specialist. There are things you’ll be familiar with: how severe are the claims? Are they false and if they’re false, is it deliberate disinformation? Who is making the claim, for example?
But there are also elements of social media that you might be less familiar with: how much engagement has this content got? Has it come from some dubious online source that might have quite a large reach in dubious spheres? You need to make sure you understand all of these, for instance, the difference between likes and comments, or search volumes, Twitter retweets and similar. And the only way to do that, like finding, is to practice and to be on top of the news and social media environment.
So the strongest piece of advice I can give you is make sure you’re monitoring social media and search around your subject area constantly, not just in crisis time. This will give you a sense of what is normal and whether you should respond. So, for example, with the DFID India aid story, the Daily Mail regularly publishes articles that UK aid money is being spent badly. But having seen such articles before, we knew that the one we flagged to DFID was doing exceptionally well on Facebook, really very well even by the standards of all sorts of government-related articles.
And we only knew that because we monitor government news on social media and record it on three daily reports. You might not want to do three reports a day, you might not have the capacity to but make sure you are recording so that you can help teammates to develop that instinct for what is normal and also, have some rigorous benchmarks to assess against.
Secondly, make sure you really understand the data and what the tools are telling you. Some of the tools are very easy to use, even the free ones. So, in Tweet Deck, you can just search for tweets mentioning certain words and then filter by all sorts of measures. In Google Trends, you can plug in a word and see how volumes of searches for that word have changed over time.
But the tools can often be misleading and you need to dig into them. So Tweet Deck claims it can pinpoint tweets to particular locations, but that’s not actually as accurate as they would claim. On Google Trends, it’s easy to make it look like some niche topic has suddenly had a massive spike in search until you compare it to a word like ‘Brexit’ and then you suddenly realise that spike wasn’t actually that big.
I’ve given more details in the handout, but it’s impossible to cover all the bases so it’s important that you try out the tools, dig into the FAQs, and discuss with your colleagues to make sure you fully understand what they’re showing you and don’t inadvertently misunderstand or mislead.
And finally, remember that when you’re assessing the scale of engagement or type of engagement with some content, this can actually help you find more things. So for instance, if you’re assessing whether a tweet is particularly problematic or worrying, maybe it’s worth seeing if people are using the hashtag that that tweet uses which you might not have searched for before. So you can make your boolean bigger and perhaps discover that the topic is going a lot more viral than you originally thought.
So in sum, you need to make sure that you’re always monitoring, recording and practising with the tools and digging into what they show and make sure that you do so in a time where you’ve got some headspace and can work with colleagues so that in a crisis, you’re not accidentally assessing something as not a problem when actually there’s something you’ve missed or finding something and flagging it as a problem when actually, it really isn’t that big.
Once you’ve made that assessment and contacted the relevant press office or special advisor or other communications professional, then it’s time to think about how you leverage that assessment to create and target content appropriately.
Finally, we’re going to talk about Create and Target together as communicators, you know it’s really hard to create something if you don’t know who you’re targeting at, it’s hard to come up with a message when you don’t know the audience.
But to look just briefly at Create on its own, in fact, we define Create as “Create appropriate content or share existing content with the aim of rebalancing the narrative and promoting official information. This may be a press office line, a social media post, or a more complex asset (such as a video)”.
Again, this may some very familiar. Press officers and other government communicators have always had a role of getting the government out where appropriate. But social media widens the range of possible options of how you reach people from different platforms – Facebook, Twitter, Instagram, Reddit, LinkedIn. But it can also make it more difficult as you can reach lots of different audiences, but you need to cut through all the information overload that’s there online, all the other people who are trying to reach the same audiences as you. And that’s why you need to target very well.
Target, in the FACT model, is defined as “Promote content across digital media to ensure information is accessible and highly visible to the public. This may involve paid-for and social or search advertising targeted at those most at risk”. Before we move to talk about paid-for advertising, it’s worth noting that if you use FACT well, you can actually increase the reach of your message for free.
By jumping into a conversation that is already happening online, you can boost the number of people who are engaging with it, sharing the content with their friends, actually commenting on what you’re saying and as a result, boosting it up your Facebook or Twitter feeds.
Since we’ve started using FACT in the Rapid Response Unit, we’ve discovered that we’ve increased the engagement and reach of government Facebook post for free, just by posting at the right time. Knowing that we’re putting the posts out at a time when people are actually talking about the topic. But you have to always remember, of course, that just using a government Facebook page means you may be limited to people who are already engaging with you, who have ever liked or followed your page, or perhaps who friends with people who have.
So in a real crisis situation or when you’re trying to reach some hard-to-reach audience, you may need to use paid-for advertising. The good news is that, even with a small budget, you can actually reach an audience if it’s targeted well. The bad news is that you have to learn how to target an audience well and there are loads of different ways to do that.
We won’t go into detail, again it’s just worth practising. But it’s also worth practising and testing, again, in quiet times before the crisis hits to discover those snags that may occur when you need to get content out quickly.
For example, did you know that on Facebook, you will need political authorisation to run ads from a government account? That can take a week or two to sort, you don’t want to discover that in the middle of a crisis. But it’s also worth noting that paid-for advertising can have all sorts of extra benefits: it can give you extra metrics, for example, to assess how your message is landing. You can see, for instance, how long people are watching your video for, that can help you know what parts of your message are really landing.
And finally, if you’re engaging with the different ways of targeting audiences whether through paid or non-paid means, you can cleverly think of what sort of content and what sort of approach is best suited to your problem. If people are saying bad or misleading things about your department on the internet, on social media specifically, maybe you should be there in the discussing.
Using social media, Facebook, Twitter or other platforms to get your voice out there and be part of that discussion. But if the concern is that people are not talking about you, but perhaps might be looking for information about some confusing or problematic topic you deal with, perhaps you should be thinking of search content.
You should be using Google search ads to boost your GOV.UK page to the top of search results, making sure people see your reliable information quickly. Again, all of this comes down to exactly what you’re trying to accomplish, the topic you’re dealing with and your objectives, but you can’t do it well unless you know the landscape well. So, get practising and get investigating.
So those were the four steps of the FACT model: Find, Assess, Create and Target. As I said at the top of the recording, in principle, this shouldn’t be that new to you but social media and search, and other aspects of the digital environment do provide some new challenges. I mean you have to learn new skills, tools and techniques.
Everyone knows when Twitter or Facebook or Google search, the more appropriate tools and techniques so that it becomes part of your day-to-day. After all, social media and searches are just an extension of communication we’ve been doing already.
As I’ve said, the FACT model used well is just a way of getting government messages out further rather than something radically new. But that does require you test and practice and experiment over time, not just in crisis situations. It requires you to do things like build up your department’s Facebook page so that it has a decent following for when you need to use it.
It requires you to know what different tools are showing you when you have a graph so you’re not inadvertently misleading your colleagues. And it requires you to understand how different platforms have different users and how messages may spread about your department or subject area differently in different corners of the internet so you’re really up-to-date with where concerning trends and things you should you be responding to actually appear.
But overall, this shouldn’t be entirely new to you as a government communicator. It just takes a bit of practice and working with colleagues, and I can say from working in the Rapid Response Unit that we’ve had numerous secondees and numerous employees who turn up with government communications backgrounds feeling, ‘this sounds very technical’ and actually getting up to speed and learn the relevant techniques very quickly.
And in the end, realise this is all just about doing a little more in the digital sphere, understanding a bit more of the modern tools and techniques available, but ultimately it just comes down to basic principles of communication.
Thank you for listening to this module in our GCS Accelerate essentials learning series. Thanks for joining us.
End of transcript
By the end of this podcast, you will:
- know how to track online discussions, to identify emerging issues and potential opportunities for response using the FACT model
- be aware of how to monitor search and social media without paid-for tools
- understand how online tools can be used to communicate HMG messages in a crisis
Slide deck can be made available on request by email: firstname.lastname@example.org.