Digital, Culture, Media and Sport Committee
Oral evidence: Immersive and addictive technologies, HC 1846
Tuesday 19 March 2019
Ordered by the House of Commons to be published on 19 March 2019.
Watch the meeting
Members present: Damian Collins (Chair); Clive Efford; Julie Elliott; Paul Farrelly; Simon Hart; Julian Knight; Ian C. Lucas; Rebecca Pow; Jo Stevens; Giles Watling.
Questions 252 - 414
Witnesses
I: Stephen Collins, Senior Director, Public Policy International, Snapchat, and Will Scougal, International Director, Creative Strategy, Snapchat.
Witnesses: Stephen Collins and Will Scougal
Q252 Chair: Good morning. I would like to formally welcome you to this hearing of the Digital, Culture, Media and Sport Select Committee, as part of our inquiry into immersive and addictive technologies. We are delighted to have representatives from Snap join us for the hearing today. We have taken evidence looking at people’s use of games and social media and, as part of our inquiry, we are particularly interested in talking to Snap about the way people use Snapchat and how Snap monitors how Snapchat is being used.
As you know, in our previous inquiry we held substantial evidence sessions with other social media platforms, so the Committee has a fair amount of information about how they work. We have not previously had the chance to engage with Snap, so we are very grateful for you making yourselves available to give evidence to us today.
Could I start off by asking: do you monitor how typical users engage with Snapchat as a tool, how much time they tend to spend on it a day and how frequently they tend to activate it?
Will Scougal: We do look at some data in that regard, yes, absolutely. I want to say thank you for inviting us in to speak to you today about the platform and, in particular, augmented reality and its value to the UK and to us as a business.
When we look at our users, we can see that we have 186 million daily active users and around 60 million daily active users in Europe. About 70% of them spend about three minutes a day playing with augmented reality on the platform. We take the privacy of our users extremely seriously and we see that they value that in our platform. When they come to the platform and use it as a communications tool, they are in a private conversation and not one that is monitored in any great detail. The statistics and data that we look at are more of an overarching look at how people are using the platform.
Q253 Chair: You said 70% of users spend, on average, three minutes a day. Are the other 30% inactive or are they on it more often?
Will Scougal: Those 70% are spending those three minutes a day playing with augmented reality. On any given day we see users spending about 30 minutes a day on the app in total. That can be broken down over a number of different facets of what people are doing on the app. We do find that because Snapchat is a network for close friends, friends that genuinely know each other, who would spend physical time together—in the park or in a restaurant, for example—as well as digital time, that when they are on the conversation side of the app they are engaging in regular conversation through the camera, using the camera to talk with pictures.
They also use Snapchat to catch up with their friends’ Stories, and when people post their Stories only their close friends can see that content. Also, Snapchatters are engaging with the premium content side of the app where we have global publisher partners and brands, like Vice, Channel 4 or the Wall Street Journal and others, creating content specifically for Snapchat. We feel that that premium content space and that close friends' network creates an incredibly high-quality mobile environment for our users.
Q254 Chair: Typical users spend 30 minutes a day on Snapchat. That is an average amount of use. Do you have upper and lower bands of usage? Presumably there will be some people who barely use it at all, who have downloaded the app but are not very active on it at all. There will be others who are probably using it a lot more than 30 minutes a day. In your view, what would a heavy user’s usage of Snapchat look like?
Will Scougal: At the moment, I am afraid I do not have access to that data. I would not be able to give you an informed view on what that band would look like but, if the Committee is interested, I can go back and then send something to you in writing.
Q255 Chair: We would be very happy to receive that. Thank you.
Stephen Collins: It is worth adding as well that when people are on the platform they are actively engaged. As Will said, they are engaged in creating content or in communicating with real friends. There is not really the opportunity to scroll through anonymous newsfeeds or vast swathes of user-generated content. We see people, generally speaking, coming and going, not spending hours and hours and hours on the platform. We do not have the exact data here, but we can pull that.
Q256 Chair: People are either sending or receiving messages, basically.
Stephen Collins: Sending and receiving messages or creating—for example, using lenses—content they are going to share with their closest friends, which is a positive experience rather than a negative one.
Q257 Chair: From a design point of view, what is the purpose of the streaks function on Snapchat?
Will Scougal: Again, going back to that idea of the networks that people keep on Snapchat being predominantly small groups of people—20, 30 or 40 people that you genuinely know over and above larger groups of people that you may not have a relationship with—what streaks do and highlight within that conversation side of the app is a regular contact between close friends. It is not something that is a core feature of the app by any stretch. What we do find is that it is a celebration of that regular contact between close friends.
Q258 Chair: You will be aware of concerns that people have raised that it creates quite a lot of peer pressure, certainly for younger users, to be messaging their friends every day so that the streak does not end. Do you acknowledge that that is a concern that people have?
Stephen Collins: We want everybody to have a positive and enjoyable experience on Snapchat. What we see, in terms of streaks, is that it is a very small icon or emoji. It was part of a pack of 30 friendship emojis that were released in 2015. There is very little pressure, I think. It is not visible to other people; it is only visible to you and your very close friend. It is meant to be a fun thing. We have not changed that feature or tried to exploit it since its release. It was only meant to be a piece of fun between close friends.
We have reduced it in size by 30%, along the same lines as the Children’s Commissioner recommended—coincidentally and in parallel—and we are always looking at the opportunity to see whether we should sunset a particular feature. It is not a major feature or a key feature of the platform. That is the important thing.
Q259 Chair: You will be aware—and certainly evidence we have received from different people who have engaged with the inquiry suggests—that where streaks exist it can create a certain amount of peer pressure on friends to keep those streaks active. The reason I ask what the design purpose of streaks is within the app is that it is not clear to me what the purpose is of having a tool that tracks a conversation over a year or longer, as seems to be the case sometimes. It could create pressure on young people to keep that streak going for fear of it being seen as a reflection on the nature of their friendship, simply because they had not sent them a message on Snapchat that day.
Stephen Collins: I understand exactly what you are saying. The idea was to deepen individual friendships. It was not meant to create extra time on the application at all. We will revisit that.
Q260 Chair: Deepen friendships by making Snapchat central to their friendship?
Will Scougal: The core functionality or the functionality within the app is built around the idea of empowering people to have more expressive, more creative, more fun, more playful conversations through their mobile devices and, in particular, through the camera. Streaks are a feature of that. It is certainly not the intention or the purpose from our side for them to be something that creates a negative experience. As a platform, we are incredibly focused on the wellbeing of our audience of Snapchatters and making sure that what they are experiencing within the app is overwhelmingly positive.
Q261 Chair: I suppose the concern is: do streaks become like a measurement tool for friendship, as far as young people see them? While it may not be core functionality of how Snapchat works, nevertheless a by-product of its existence is that the length of the streak is somehow a measurement of the strength of a friendship and, if it is broken, that somehow has a real-world implication for a friendship. Most people might think that that does not reflect their own experiences of life but, nevertheless, for a young person it might.
Stephen Collins: That is a fair comment. We will take that away and we will look at it further.
Q262 Giles Watling: I am interested in the younger users. What is the demographic of the users of Snapchat, would you say?
Will Scougal: It is 13 and above, as a platform. Of the 186 million DAUs, over 75% are over the age of 18. I think a lot of people find that statistic surprising. From a core point of view, our core audience is very much in the millennial and Gen Z age groups, over 13 to 34.
Q263 Giles Watling: You have established that 13 and above is your age testing? You test for the average age. I am sure you are aware that CHILDWISE says that 90% of children aged 11 and 12 use age-restricted social media services. Would you say it is a fair comment that there are many who sign up to Snapchat who just bypass the age restriction and go on there?
Stephen Collins: Again, it is a fair comment. We do not want under-13s on the platform. We take steps to prevent them joining the platform but, you are absolutely right, the end result is that there is no foolproof verification system.
What we do now is, first, we have an age gate. People have to self-declare their age. If they put in under 13 obviously they are refused. If they are doing that on the web, we set a cookie to remember that browser and we also disable the back button so they cannot go back and put in a different age.
We also put the app in the teen section of the Google Play Store for Android—that is 80% or so of the market—and then in the 12-plus part of the app store for Apple. They are the basic safety features. Then we hope that parents and carers, of course, will help children adjust their phones to take advantage of those apps that are safest and most age-appropriate for them.
In the back end, we also do detect behaviour, consistent behaviour over a period of time that would indicate an under-13. If that is clearly the case, our Trust and Safety Team will investigate. If it turns out to be the case, we will delete the account.
Q264 Giles Watling: It would be fair to say, though, that in my experience the average young person can get around these things very easily. You are asking them to self-declare their age. How would you envisage being able to make it harder to access the platform for underage people?
Stephen Collins: Yes, it is a good question. Just to restate, we really do not want under-13s on the platform. That brings us no advantage. That brings us no commercial benefit at all. We want to make it an enjoyable place for everybody who is on the platform.
There are a couple of things that have been tried or mooted up until now and then there is something more for the future that I will come back to. Things like parental consent seem to be quite popular on a superficial level. Again, you are absolutely right that children are able to circumvent these things, for example, by setting up spoof email accounts or by taking a quick photo of their parent’s credit card details, if that is going to be the verification means, or a passport. There are a whole variety of ways that under-13s could still get around that.
The most sustainable solution would be some kind of central verification system, which is already being discussed. The Home Secretary has tasked the Home Office and related agencies to look into this. We are part of that working group. We met just yesterday. I cannot give you the details here because I am under an NDA, but we could furnish the Committee with those details in writing. It is a serious attempt to come to a proper conclusion, a fitting conclusion, to this conundrum that has been there for a long time.
Q265 Giles Watling: We would be grateful if you did furnish those details. I would like to move on very quickly to the ludic loops—which you have doubtless heard of—which are powerful psychological manipulations that look playful on the surface of it, and Snapstreak could be one of those. Call me a jaded old cynic but, fundamentally, you get very young people, you sit them into a loop where they have to keep going back to keep their self-esteem going or keep up their friendships or whatever, and you are sitting very young people in front of advertising, in a gambling or addictive sort of way. Would you say that was a fair comment?
Stephen Collins: Do you want to talk about the design side, Will? I think it is important.
Will Scougal: Yes, sure. Again, fundamentally, the wellbeing and the positive experience of our users are absolutely paramount. Because we are led by a product design mentality and aesthetic, our users and what they are doing on the platform very much inform the advertising we create. From the point of view of the three core formats in terms of video, filters and lenses, which are augmented reality, each one of those formats has been designed to create a positive experience for users.
First and foremost, I would say that fundamentally each one of those is very clearly called out as advertising. The second point I would call out is that they are predominantly based on choice, as in the user has it within their power to choose to engage with an ad or to choose to move past it or to not engage with it.
For example, in the case of augmented reality, you open up the app, the app opens onto the camera, you scan your face or the world around you and it launches a series of AR experiences, one of which might be an advert but the rest of which might just be experiences that we have created to enhance and augment conversation.
From the point of view of video as well, when you are on that premium content side of the app, again, the power is very much in the hands of the user. What we have done from a design perspective is to try to create the best-in-class video ads that we could. From our point of view, that meant that the video was full screen, it was sound on, and that the user could tap to skip past it or choose to swipe up on it and experience more of it. More than anything, we have tried to create ads that contribute to the overall experience, again going back to that fundamental mission of making it a positive experience as opposed to something that may be forced upon them.
Within our shows content, we do have commercials that play for six seconds. Again, when we think about how users are using the platform, what we saw within shows was that Snapchatters would watch the whole thing, much like they might watch a television programme, so it made sense for us to experiment with the idea of a six-second commercial. That has been working very well in that our users are watching them; they are not leaving the shows. For our mind, that is an additional format that was a long time in development and design that has ended up contributing to the overall content experience.
Q266 Giles Watling: The addictive cycle is something I have on the wall of my office. In previous panels we have asked the likes of Facebook, Twitter and Instagram about this addictive cycle, where you Tweet or you do something and then you have to go back and see whether you have been liked and then, a little while later, if you have been liked that is satisfying and then you have to go back. Would you say that that is deliberately built into Snapchat?
Will Scougal: First of all, yes, those metrics and those behaviours do exist on other platforms because they are built like that. When we think about Snapchat, and when we talk about Snapchat as a platform to partners, we always frame ourselves as a camera company first and foremost. There is a fundamental difference between the way we position ourselves and the way the app is built, versus platforms that you might instinctively compare us to.
If you open up a feed-based platform, you are prompted to consume the content. You scroll to consume. When you open up Snapchat, you open to the camera. You are given a vision of the world around you and a tool with which to capture it, augment it, enhance it and share it with your close friends.
What do not exist on Snapchat are traditional social metrics, such as a like, a retweet or a follow, because we are not looking for users to create content that is then validated with that kind of metric. They are creating content to use as communication. They are creating pictures using the Snapchat camera to communicate with their close friends. For that reason, we do not see ourselves as a social network. It is not a ‘broadcast to many, validate the content’ experience. It is more of a broadcast to or, rather, to communicate with fewer people who you genuinely know and who genuinely like you. It is more of a conversation than a broadcast platform, in that sense, when it comes to social metrics.
Q267 Giles Watling: You clearly feel that it is advantageous for you as a company to keep people using Snapchat, to keep them online as long as possible. That is your aim.
Stephen Collins: To take it at a higher level and just give some more context—which I think perhaps is helpful—the application was created initially as an antidote to social media. It was not created as social media. The founders, who are still essentially running the company today, wanted to create the most natural form of communication in a digital space without any of the pressures they felt that social media was putting on users.
What they created was an application where it was primarily set for close friends and family, a narrow, deep relationship tool rather than, as Will says, a ‘broadcast out’ tool superficially to thousands of people. The evidence of that is clearly that most users have a list of friends in the tens—not in the hundreds, not in the thousands, not in the millions, it is really in the tens—and they engage with those people in a more natural way.
The reason it opens to the camera is to allow more natural forms of communication, reducing barriers between people. The only thing between you and the person you are communicating with is the camera and the camera lens. Ultimately—and I am sure Will will talk about this later—we may able be able to do away with the camera itself, or at least the smartphone. We can talk about that as well.
The other key feature that is important to stress is that, from the outset, Snapchat messages were ephemeral. Once they are seen, they disappear. Again, it reduces pressure on people to be something they are not, try to be more than they are or try to be something artificial. If you know that if you send a picture or you send a text, it can be funny, it can be silly. If you know it is going to disappear afterwards, you are more likely to engage more deeply in a more authentic way.
What we have done with the platform is created, I think, a much more qualitatively strong experience rather than a quantitative one. Again, statistically, if you look at the number of users that we have compared to the largest social media platforms, for example, the numbers do not bear any relevance at all. If you look at the subjective wellbeing research that is being done, people generally do feel much better using our platform than regular social media. That is an important differentiation that came from the very start of the creation of the app.
Q268 Giles Watling: As you are probably aware, there is always an issue with particularly young people sending inappropriate, probably sexualised images across social media, because it is of a temporary nature on Snapchat it is there, and it is gone. But there is still the issue of screenshots and people just saving the image. I am sure you would agree there is a danger there. Is there any way you can approach and deal with that sort of media being—
Stephen Collins: The actual nature of a screenshot is in the device, not in Snapchat. The device manufacturers create that functionality and the user is able to capture anything that is on the phone so that the screenshot is there. What we have done to protect people or make them more aware is that if a screenshot is taken: first, it is not really the etiquette to take a screenshot of the Snapchat community but, secondly, a message comes up to let the sender know that the picture they sent, the recipient has taken a screenshot. That is there as a matter of record.
Q269 Giles Watling: They are alive to the danger?
Stephen Collins: They are able to see that.
Q270 Ian C. Lucas: Mr Collins, if someone is below the age to qualify for Snapchat, you said that cookies were put in place to flag that particular account.
Stephen Collins: If they apply on the web, correct. If they sign up for an account through the website, that is correct.
Q271 Ian C. Lucas: Through the website?
Stephen Collins: Through the web browser. Cookies are set in the browsers.
Q272 Ian C. Lucas: On a phone like this one?
Stephen Collins: If you went through Safari on the iPhone there, yes.
Q273 Ian C. Lucas: I downloaded your app on Snapchat here and I have just signed up with an account. When I first signed up, which was during the hearing about five minutes ago, I put in 2008 for a date of birth and I received a message that said, “You are not eligible for Snapchat”, which is fine. I then put in my true age, which is before 2008, and it allowed me to sign up and I have just opened an account.
Stephen Collins: Yes. If you sign up through the application, because there is no browser it is not possible to set a cookie. It is an easier route. That is correct.
Q274 Ian C. Lucas: What you told me earlier was that you would be flagged up. You did not mention the distinctions of how you signed up. You told me that there was a flagging of an account. Someone who is 12, if they cannot get in because the qualification age is 13, if they learn that, is just going to put an earlier date of birth in.
Will Scougal: Right. In the instances when something like that might, or may or may not happen, there are other signals, inference signals that are picked up by the platform, by us, and we actively look and assess whether or not a user is of the right age. If we deem them to be of the wrong age we will eject them from the platform.
Q275 Ian C. Lucas: Sorry, you will what?
Will Scougal: We will remove them from the platform, block their account.
Q276 Ian C. Lucas: When would that happen? Do you have a review period for someone who is using Snapchat? Is that because it seems to you that, from the account, they are below 13?
Will Scougal: I will have to come back to you on exactly the detail of what we look for in that particular instance because I do not have it with me right now, but there are a number of different signals we will look at. When we think about how users use the platform and the experience they have, we build primarily with privacy and safety in mind first and, if there is ever an inference that either of those are at risk—for example, if someone is underage—there are systems in place to ensure that they do not, for example, get served advertising.
Q277 Ian C. Lucas: What proportion of users sign up to Snapchat through a mobile phone?
Will Scougal: I am afraid I do not have that. Well, all of them will sign up through a mobile phone because it is a mobile-only platform.
Stephen Collins: You can sign up through the web on your phone or on a PC; you can create an account either way.
Q278 Ian C. Lucas: Hold on, I just want to be clear about signing up through the web and how the system would work if it was signed up through the web. That it would not work on a mobile phone. Well, it is working on a mobile phone because I have just signed up on the web, despite the fact that I put my true date of birth.
Stephen Collins: There is more than one way to sign up, you are absolutely right.
Q279 Ian C. Lucas: I want to know—and I am amazed that you cannot tell me this—what proportion of people sign up on a mobile phone as opposed to on the web.
Stephen Collins: Snapchat is a mobile app.
Q280 Ian C. Lucas: So most will be on a mobile phone?
Stephen Collins: Yes.
Q281 Ian C. Lucas: Your system does not work for most people who sign up, for age verification?
Will Scougal: What I would really like to be able to come back to you on is a full—
Q282 Ian C. Lucas: Mr Collins, you are pulling a face. Why are you pulling a face?
Stephen Collins: I am trying to think of how to phrase this without it sounding negative or misleading. There are different ways to sign up to create an account. You have identified one, which is a popular way, I am sure. We will find out exactly the figures. There is more than one way—as Will has described—of identifying under-13s, primarily by their behaviour on the platform subsequently to signing up, but as I also said to the Chair earlier—
Q283 Ian C. Lucas: We want to stop people from signing up before they sign up if they are not qualified, don’t we?
Stephen Collins: What we really need is an age verification system that is robust and works. We are working hard with the Home Office and with other agencies connected with the Home Office to create that. We have to hold our hands up and be honest. Anybody who works in the internet industry will tell you that it is not possible to have a fool proof way to keep under-13s off any platform.
Q284 Ian C. Lucas: Why is it not possible? You said that the method I used to sign up is a popular way of signing up to Snapchat. We do not know how popular that is, but you are going to tell me how popular that is because I think that is important. Your age verification system does not work for a popular way of signing up to Snapchat. Do you agree?
Stephen Collins: On the initial sign-up, certainly, yes, we agree.
Q285 Ian C. Lucas: That is not acceptable, is it?
Will Scougal: When we look at the industry as a whole, the age—
Q286 Ian C. Lucas: No, I am asking about Snapchat. You are responsible for Snapchat.
Will Scougal: Yes.
Q287 Ian C. Lucas: You are responsible for the age verification system and it does not work on the popular way of signing up to Snapchat. We get a lot of issues from our constituents concerning children and the impact of social media. Age verification is really important.
Will Scougal: I absolutely, 100% agree. With the signals that we have that we look at on the platform, from a sign-up point of view through to the way people use it, there are ways for us to be confident that the users that we have on the platform are of a certain age. Certainly, when we think about it, from an advertising point of view, as Stephen said, there is a broader point about the need for there to be a robust age verification system that we can all get behind.
Q288 Ian C. Lucas: Do you know how many people you have taken off Snapchat on an age verification basis after they have been on in the first place?
Will Scougal: I am afraid I do not have that number now.
Q289 Ian C. Lucas: Can you please tell us?
Will Scougal: Yes.
Q290 Ian C. Lucas: Turning to the issue of the legal position, you kindly wrote us a letter dated 15 March before this hearing about data control. You said in that letter, “Snap’s use of data is controlled by our main US entity, Snap Incorporated”, and you went on to say, “We have agreed to disclose data directly to foreign law enforcement agencies in good faith in a very limited set of circumstances consistent with US law”.
Chair: Just for the benefit of the record, this is obviously in reference to the Breck Bednar case that you kindly wrote to us about in response to the public discussion on that. The matter was raised in the House of Commons previously. Just for the record, that is the context of these questions. We just have a few questions about the letter in that case.
Q291 Ian C. Lucas: “We have also agreed to disclose data directly to foreign law enforcement agencies in good faith in a very limited set of circumstances consistent with US law”. Is that your policy or is it the legal position? In other words, are you choosing to apply the principles of US law to cases in the UK, or is that because you are legally obligated to do that?
Stephen Collins: I will take that. We understand it is our legal obligation to comply with US law and the US entity is the data controller. We have had discussions with the ICO concerning which entity would be the data controller for the UK. They agree it is the United States and, therefore, we have to abide by United States law. There are exceptional circumstances, which are in the letter, for terrorism cases, child sexual exploitation cases and imminent threat to life situations, where we will act in good faith and directly respond to the requests.
In other cases, we have a Law Enforcement Operations Team that works 24/7 and we have a Law Enforcement Operations Guide that we provide to law enforcement agencies, including in the UK. We have a very positive relationship with UK law enforcement generally. It works very well. Co-operation, generally speaking, works very well.
What is frustrating for everybody, including us, is that the mutual legal assistance treaty between the US and the UK is a very slow process. There are things that can be done in exceptional circumstances, I understand, to expedite that process. That is an inter-governmental process, though, not a company one. We undertake to respond very quickly to any requests we receive from UK law enforcement through the process or in those exceptional circumstances, directly.
Q292 Ian C. Lucas: Your view is that you are constrained by US law as to what information you can disclose, despite the fact that, for example, this particular case occurred in the UK? Is that right?
Stephen Collins: Clearly, we can disclose through the MLAT process—
Q293 Ian C. Lucas: You are constrained in disclosing certain information because it goes beyond US law to disclose that.
Stephen Collins: Yes, that is correct. There is a potential conflict.
Q294 Ian C. Lucas: That is not your choice. That is in place because of US law.
Stephen Collins: As I understand, yes. The other thing to say is that this is a temporary situation because the US has introduced and agreed the CLOUD Act. In the UK there is the so-called COPO Act, the Crime (Overseas Production Orders) Act, which went through the House last month.
Both of those have, as the first priority, UK–US mutual recognition or expeditious responses to requests from the other Government, from the other party. That will resolve what has been an issue in the internet space for 20 years, and not just between the UK and the US. Wherever there is an MLAT, a mutual legal assistance treaty, the processes are so long and involved and the personnel on either side so stretched it sometimes takes months. That could be between the UK and Brazil, the UK and Argentina or wherever. It is a known issue and one that has been there for 20 years. It looks like it will be resolved, at least between the US and the UK, in the next few months.
Q295 Ian C. Lucas: Can you just help us by saying one action you have taken to help the police investigating the messages sent to Breck Bednar’s family?
Stephen Collins: Yes, of course. As always, in this case as well, we responded very quickly to the two requests the police made, the Kent police, the Folkestone constabulary. They asked first what our process was. Then they asked a second question about one clarifying point on that process. We never received any formal request for data at all from the police.
Ian C. Lucas: Thank you very much.
Q296 Chair: For the benefit of the record, the Breck Bednar case is particularly distressing. The boy was groomed and brutally murdered and the family have subsequently been receiving messages—they believe from the murderer—with details of the boy’s death. Those messages have been received via Snapchat. That is the context of the case.
Obviously in your response to the request for information from Folkestone Police Station—which is in my constituency—you set out the company’s policy through this treaty arrangement, by which an application for data could be made, but no data request has been made to Snap by the police. Is that correct? That is correct. If Kent Police did make a request directly to the company for that data and information, as part of their investigation, is there anything to stop you giving it to them?
Stephen Collins: No, but they would have to go through the process. It is a UK–US governmentally agreed process.
Q297 Chair: What would be the consequence of Snap responding directly to a request from the UK police in this regard?
Stephen Collins: I do not know but, potentially, we would be in breach of our obligations under US law.
Q298 Chair: Is the request in breach of US law? If this was an American police organisation making this request of Snap, would there be any reason in law why it could not be complied with?
Stephen Collins: I am not a legal specialist in this area. I can come back to you on that. It is my understanding that it would be.
Q299 Chair: It would be in breach of the law?
Stephen Collins: It would be. That is my understanding, but let me clarify that and I will come back to you.
Q300 Chair: We would be very grateful if you could because, if it is not and there is some discretion as to whether you would comply or not, a lot of people would suggest that, in a tragic situation like this, rather than directing people through a long and bureaucratic procedure, a more direct response might be more appropriate. The key question here is: do you have discretion over whether you comply with a request directly or not when it crosses the jurisdiction of the US and the UK?
Stephen Collins: Let me come back to you on that point. I want it to be absolutely clear because it is an important one and it is a very, very distressing case. I would like to take this opportunity to extend our sympathies to the family. As a parent I cannot imagine, to be fair, what they are going through.
I think we have done what we could on the legal side and also on the trust and safety side. It is important that you know that we have been in direct touch with the family. We have helped them understand how they can report to us to have us block accounts that are harassing them. If they report through the app we can preserve that data for police investigation. Also, through the MLAT process, if the police request us to preserve data while they go through the process, so that it is not lost, we will do that.
I would say as well that we have advised the family on how to restore the default settings on the Snapchat account of family members, which of course would allow them then to only receive messages from family members and to refuse friend requests from people they do not know. That does not take away the hurt, but it removes future harassment.
Q301 Chair: I understand that and a lot of that is set out in the letter to the Committee. Would you be happy for the Committee to publish the letter you have given us as part of our inquiry, so that it sets out for the record the issues that you have set out in the hearing today? You can tell me now if it is a yes or a no, or if you want to get back to us on that, that is fine.
Stephen Collins: Can I get back to you? I think it will be fine and I do not want to sound negative or not constructive—
Q302 Chair: No, I understand that.
Stephen Collins: —but let me just check with our legal department. I think it will be fine, but I will come back to you.
Q303 Chair: Thank you. I understand the process that you set out. To me, it would seem that that process is inadequate in a case like this. I suppose, what we want to determine is if that is inadequate because the legal framework is inadequate and needs to be changed, or whether in situations like this companies like yours do have some discretion in how you respond. If you could let us know, in point of law, the company’s position on that. That would be very helpful.
Stephen Collins: Yes, I think that is fair.
Q304 Paul Farrelly: Just on this point—and I know we have lots of other things to explore on the immersive reality side—it is cases like this, of course, that give companies a very bad name. It does not just affect you, clearly. There is Apple and access to its devices, and of course there are privacy and other considerations across the world depending on the different jurisdictions, some of which are more repressive than others, but this is not a repressive jurisdiction.
In your reply could you also include to what extent you have discretion in respect of your terms and conditions for users in different jurisdictions, to have that discretion to comply when reasonable requests are made by law enforcement in situations like this? I cannot simply believe the default is purely US law.
Stephen Collins: Again, it is a pertinent and fair question. Of course, a change in the terms of service to increase law enforcement access to user data generally is something that would be very seriously considered because there are definitely privacy implications. You are right, this is not a country with a repressive regime and unfair law enforcement agencies, but there are countries like that. Again, I will come back to you on the terms of service.
Q305 Paul Farrelly: Just regarding the issue of age and identity verification, I do not want to give you too much of a hard time because it is the same issue for all the platforms. I am sure that if children under 16 had to go through parental consent, they would just say they are 17 anyway. It is an issue and your platform has not been associated with some of the disgraceful vigilante-type scandals that WhatsApp has been hit by, but you have to think if you do not get it right, there but for the grace of God do you go until something else potentially happens in the future.
Stephen Collins: Yes. The other thing to emphasise, another point of differentiation, I guess, from some of our competitor platforms, is that we have a very high degree of privacy built in. It is privacy by design. From the very earliest stages of product development, we have had a Privacy Council and a Product Council working with the engineers all the way through the product development lifecycle. Wherever there is an opportunity, for example, to minimise data, to protect user privacy or not to over-collect data that we do not need to provide the service, the data minimisation principle, we will do that.
Some of the problems that occur on other platforms are more down to the widespread sharing of personal data, either with third parties or on the platform, and then the usage of newsfeeds, for example user-generated content newsfeeds. Whereas on our platform, the public side—all the news, the entertainment and so on—comes from professional publishers. We do not have constant feeds of user-generated content being broadcast out. We are a very poor broadcast platform for user-generated content and that is deliberate.
Q306 Paul Farrelly: If people think that on your platform images disappear very quickly—but in fact they do not always do that because they can be copied—then intuitively you would think that people might be encouraged to behave in ways in which they would not normally behave and, therefore, that is a risk of using Snapchat if those protections are not effective, particularly in terms of sexual imagery. I see that you are very popular with young girls.
Stephen Collins: In terms of one-to-one usage, we do not monitor people’s private communications, just like Vodafone or British Telecom do not monitor people’s one-to-one phone calls. We do not do that. We are trying to create a natural communication between two people.
If there is anything in a conversation that somebody is not comfortable with or that breaches our terms and it is reported to us, we act on that very quickly. We have a 24/7 Trust and Safety Team. You can flag any kind of uncomfortable content or what you think is bad content. It will be flagged to our Trust and Safety Team and they will act quickly. They will investigate, take a snapshot of what that is, and we will enforce our terms. Our community guidelines are very strict when it comes to things like pornography, offensive content or hate speech. All these kinds of things are forbidden by the community guidelines, so if we are notified we will act, and we will act decisively.
Q307 Paul Farrelly: I think other members of the Committee want to come in on to some of the statements that your chief executive has made on privacy and data protection, as regards you and the likes of Facebook. One final question: we do not dwell on compliments very often, it is always complaints. In terms of the complaints that are made to you, what are the most frequent categories of complaints?
Stephen Collins: Do you mean in terms of trust and safety?
Q308 Paul Farrelly: You receive complaints. Do you categorise them, enumerate them?
Stephen Collins: Yes. We can provide you—
Q309 Paul Farrelly: Could you get that to us?
Stephen Collins: I do not have the details here, but we could, I am sure, rank certain categories for you that have come up. That would be things, I guess, like hate speech or nudity or—
Q310 Paul Farrelly: Yes. What do you do about them? What do you do in terms of the complaints procedure?
Stephen Collins: We will always act on the complaints. We will suspend an account. We will delete an account.
Will Scougal: One of the things to flag again is back to the idea of the quality of the experience we are trying to create. If a piece of content is flagged as unsuitable or something that is offensive or something that is not within our policy guidelines, it will be reviewed by a person. It is reviewed swiftly, and it is removed from anyone being able to see it until it has been reviewed by a person.
Q311 Paul Farrelly: We do not want to be spoilsports but if you have things in your response you can stress that you do that others do not, that is positive for you.
Will Scougal: I think the swiftness of the response and the fact that it is human eyes looking at it. There is not an automated system that will make a decision; it will always ultimately be a person that reviews it.
Stephen Collins: Sorry, Chair, could I just add one point to that, which I hope is interesting and useful? We have published a twice-yearly transparency report since around 2015 or so. That has focused up until now on our co-operation with law enforcement around the world. We are in the process of developing that further and broadening it out in anticipation of the Online Harms legislation that will pass through the House in coming weeks and months.
Of course, at European level, there is the non-binding, voluntary best practice that we have discussed with the European Commission, so some of the data that you are referring to will also become public in the coming months. From our perspective, it is important to give as much information as we can and be as transparent as we can about some of the not so good things that do happen on the platform.
Q312 Chair: Before we move on to some other topics I want to pick up on one or two things that have been discussed so far. We talked about streaks earlier on. If I was a parent who was concerned that my child was using Snap and was becoming unduly stressed by the pressure to keep streaks going for a long period of time, can I disable the feature?
Stephen Collins: No, you cannot disable streaks on its own, no.
Q313 Chair: Why not?
Stephen Collins: It just is not in the feature set.
Will Scougal: It is not a function that exists.
Stephen Collins: It appears. If they stop using it for 24 hours, it will disappear.
Q314 Chair: I appreciate that but if I was concerned, as people are—people have raised this with us, and people are concerned that streaks create undue pressure for continuous usage of the app—I either use Snapchat or I do not? You cannot use Snapchat without streaks?
Will Scougal: Streaks are not something that come up for every user as a default. You have to be particularly good friends with that person and have regular contact with them for a streak to appear on a daily basis. For us, again, that goes back to the celebration of that friendship rather than it being intended to be a point of stress or anxiety. In the design process that we are undertaking, we are certainly always reviewing all of the features within the platform to make sure that what we are ultimately building is a positive experience.
Q315 Chair: It strikes me that when a feature is baked into the app in that way—to such an extent that you cannot use the app without that being a feature of it—that suggests it is quite integral rather than a nice to have. As I said earlier on, people have raised concerns with us about whether this creates pressure for continuous usage and children become anxious about streaks ending. In some ways, I think streaks work in a very similar way to the way likes work on Facebook or the way that retweets work on Twitter. There you might measure the value of a post by the size of the audience that it reaches. I appreciate that you do not reach your audiences in that way on Snapchat, but what you might say is the value of a relationship is measured by the length of the streak.
Will Scougal: That is a valuable piece of feedback that we can take back internally. I do think that the nature of Snapchat as a platform and the content—the fact that messages are ephemeral; the fact that on the Discover side of the app, where we have our premium publisher partners and our broadcast partners, the content that they create really only exists for 24 hours; the fact there is not really any content on the platform that exists for longer than 24 hours; the quality of the conversation you are having between friends, the quality of the content on that side of the app—for us is fundamentally what drives people to come back to Snapchat on a daily basis.
Q316 Chair: Is it not the case, say, if I was using Snapchat and I had a streak with a friend, if my friend is messaging me, that my smartphone would start to buzz to indicate that I was about to receive a message on Snap, that someone was typing a message for me before it had sent? Is that the case?
Will Scougal: Sorry, could you repeat the question?
Q317 Chair: If a friend of mine on Snapchat is sending me a message I get a notification—probably in the form of some sort of buzz on the phone—that someone is in the process of sending me a message, even though it has not yet been received. That is correct, is it not?
Will Scougal: No, not in that sense. Again, back to the idea that everything on Snapchat is in the power of the users to create the experience they want to have—
Chair: Unless you are using streaks, obviously.
Will Scougal: —notifications are off by default. If you turn them off you can have a different level of notification. That really is in the hands of the user to create and that can be specific to a particular friend, a particular group or a particular piece of content.
Q318 Chair: Can a user override any of the notification settings on their device if they do not wish to receive notifications?
Will Scougal: Not as far as I am aware. They can turn them off and they can turn them—
Q319 Chair: If I was a parent, say, or a user who is concerned that I find these prompts and notifications distracting, constantly driving me back to use it, if I would like to use Snapchat but I do not want to be constantly reminded all the time that I have to be using it, you cannot turn these notifications off?
Will Scougal: You absolutely turn them off. I have mine turned off.
Q320 Chair: Sorry, I thought you said a moment ago you could not.
Will Scougal: Sorry, if I did, I misunderstood the question.
Q321 Chair: I suppose there are two sorts of notifications. There is the notification you get on your smartphone if it is sitting idle and you get a message to say you have received a message on a particular app, but then there seem to be in-app notifications that exist. You cannot turn those off, is that what you are saying?
Will Scougal: Yes, you can.
Chair: You can?
Will Scougal: Yes, you can turn them off. Again, the experience that you create within the app is designed to be in the power of the choice of the user. If they want to have notifications on, they can turn those on within the app. If they want to have them off, they can turn them off within the app. They can opt in to have notifications on particular pieces of content. For the majority, notifications are off by default. That is a decision by design to not be intrusive and to be respectful of the user experience. Obviously if the user wants to choose to be updated and notified then we allow that, but we very much want them to be able to choose whether or not they can, or they cannot.
Q322 Chair: It seems odd to me that you provide that level of choice on notifications—which I think is the right thing to do—but you do not on things like streaks, which has been a concern about the way Snapchat is used.
Stephen Collins: It is a fair comment. We take on board the streaks, the strong comments from the Committee, we take those away, and let us come back to you with some thoughts on that.
Q323 Chair: Yes. It is fair enough if people enjoy the function but, if other people do not or are concerned about it, it is whether they should have the right to turn it off.
Stephen Collins: It is a fair comment, Chair.
Q324 Jo Stevens: Can I just stay with streaks for the moment? Earlier you said—I cannot remember which one of you, I am sorry—“We are incredibly focused on the wellbeing of our users”. In May last year in our disinformation and ‘fake news’ inquiry, we had Tristan Harris come and give evidence to us. He is from the Centre for Humane Technology. I just want to read to you how he described streaks, to see whether or not you agree with it.
He says that streaks is, if you have two friends and you have a third party, like a puppet-master, who has made these two human beings feel like they are only friends if they keep this streak going. It is like tying two kids on two treadmills next to each other, tying their legs together with string and then hitting ‘start’ on both treadmills.
Those kids are running, and they are trying to keep up with each other. They are sending the messages back and forth every single day, and if one of them does not send a message back the other one falls off the treadmill, but it is not that the child has fallen off the treadmill. It is their friendship that has fallen off. That is how they feel. That is how it makes them feel. Would you agree or disagree with that analysis?
Stephen Collins: I think the point you made at the beginning is that we do value extraordinarily highly the wellbeing of our users. That is the principal point that we should keep in mind. I think it is an exaggerated description. We have acknowledged that streaks may not always be used in the way it was intended, which was supposed to be as a light-hearted small emoji that would appear as one of 30 emojis. Nobody ever talks about the other 29. They were just in a pack, little faces that appear next to friends. It is meant to be carefree and enjoyable. If that is indeed not the case we will revert to the design team. We will talk about that, discuss it, and we will come back to the Committee with a view. It is very hard to comment on such an extreme kind of picture.
Q325 Jo Stevens: It has been on your platform since 2011, I think, or later?
Stephen Collins: No, they were around 2015. The emoji pack came out, I think, in 2015. We are talking about a small minority of users. I do not want to minimise the concern. Clearly, it is a concern of the Committee. We have heard it before from others. For what it is worth, we have already reduced the size of the tiny little flame. It is just a tiny icon, probably 4 or 5 millimetres high. It does not do anything. Nobody else can see it apart from the two friends. Let’s look at it further. We absolutely take your comments on board and appreciate them. We will take that away and get back to you.
Q326 Jo Stevens: Just briefly on another aspect of your evidence, something else we have heard this morning that certainly concerns us is about the age verification system, which clearly does not work based on what we have heard this morning. If it does not work, how does Snapchat protect children under the age of 13 who access your platform?
Stephen Collins: It is important to look at the opportunities and the whole value chain as well. We have acknowledged, of course, the shortcomings that the industry has on age verification and how we are working hard to find a verifiable solution, not a solution that creates a false sense of security like perhaps parental consent would do. In some of those cases, of course, the person the child is trying to avoid would be the parent, so it is hard to always have parents involved.
If you look at the overall opportunities in the value chain to protect children that are not being used but that could be used—and this is not to apportion blame on anybody; it is a recognition and educational thing—all phones have various settings that can be set by parents, carers or educators to restrict access to certain content and certain sites.
Both app stores, Apple and Android, have family suites of tools to allow parents, carers and educators to control a safe experience for people on those phones. It could be screen time. It could be whitelisting of sites or blacklisting of sites. It could be restriction based on age to the app stores and apps. Our apps are only in the over-13 categories of the stores.
There is an opportunity as well when phones are bought for children and when data plans are bought for children. It is adults who are buying those, and they do not know that these tools exist. There is an opportunity for operators to ask, “Are you buying this phone for an under-18? Can I help you configure it to what you think is most suitable?”
Q327 Jo Stevens: I am sorry to interrupt you, but what frustrates me—and it is the same for lots of the evidence we hear about different things about social media and digital activity—is that you do not take any responsibility before the event. It is always afterwards. Something bad has to happen before you think about the consequences of the product that you produce. Why not think about these things before you do it?
Will Scougal: You raise a really important point and it is worth considering the differences of Snapchat to other platforms that you might instinctively compare us to. The environment that exists for friends to connect is very different to social platforms. It is predominantly close-knit small groups of friends who connect with each other. To go back to the previous question about what we do to ensure the safety of someone who signs up and is underage—
Q328 Jo Stevens: You accepted this morning that that must happen very regularly.
Will Scougal: There are signals we look at and are constantly looking at to make sure the age demographics on the platform are correct. For example, it might be flagged that someone is connected to people who behave in a certain way, look at certain content and are in certain locations, which might infer that they are of a certain age. That would put them into a group of people who, for example, would not receive advertising at any point. It would put them into a group of people who would be flagged as potentially underage. As I said previously, I would value the opportunity to come back to you with some specifics on what we do in that instance.
Q329 Jo Stevens: Perhaps you can help me here because I am slightly confused. You said earlier that you do not monitor people’s communications on Snapchat. How do you monitor for inappropriate content or is there some content that you moderate and some you do not? I am unclear about what you do.
Will Scougal: A user can flag whether a piece of content is inappropriate. If someone was to send something to them that was offensive or that they felt was inappropriate, they can flag that to us. That is point one. We do not monitor the one-to-one communications between people on the platform. That is fundamentally core to the design for privacy and design for safety elements that we adhere to.
Q330 Jo Stevens: Right. I want to go on now to data collection and protection. Your chief executive was quoted last year as saying that he thought Facebook could do well to copy your data protection policies. What makes you better than Facebook in terms of that?
Stephen Collins: It is good that you quote our CEO. There is a culture of data privacy from the top of the company down. We believe that, even though the short-term gains for us commercially may be tougher, the long-term wellbeing of the company is best served by having a data minimisation policy rather than a data maximisation policy. It is important to understand that we do not collect vast swathes of data to create complex data graphs of our users that we sell on to third parties, for example.
Q331 Jo Stevens: What do you collect under these minimisation principles?
Stephen Collins: There are two key areas of data that we collect. One is the data that is necessary to provide the services. Users sign up and give us basic subscriber data and then, for example, we will collect their location data if they opt into a location-based service. We cannot serve them a geofilter—a kind of switch that is located geographically—without them telling us their location. They would have to opt in to location data being collected for such a feature to work.
We also collect behavioural data, which we use for targeting of advertising. That is the behaviour of users on the public side of the platform when they are interacting and engaging with professional publisher content. There, again, it is inference. Maybe Will wants to talk about the intricacies of advertising, but we cannot do microtargeting and we cannot offer advertisers that granular detail. We can offer them a qualitatively better experience with our users because they are more engaged and are happier on the platform. What we lack in data we make up for in creative, positive engagement, and that is why advertisers advertise on the platform. We should be clear that the platform itself employs data minimisation techniques rather than data maximisation in terms of collection.
On the private side of course is the ephemerality of the content that is shared one-to-one. In one-to-one communications, it disappears after it has been consumed. In fact, most of the user-generated content on the platform is gone either once it is viewed or within 24 hours. Again, those data points we are not collecting. We are not maximising data collection to create complex data graphs to microtarget our services or third-party services.
Q332 Jo Stevens: Has Snapchat ever been used for political advertising in the UK?
Stephen Collins: Very little. We do have a strong political advertising policy. Every one of those ads is checked. It is tagged. It is clear what it is. It cannot be from a foreign backer. Whoever pays for it has to be a UK citizen. The policy is on our site. I can send you the political advertising policy. You may know more about that than me.
Will Scougal: Yes, on the political side, there have been political ads, but they are rare.
Q333 Jo Stevens: When did you start taking political ads?
Will Scougal: When we started our advertising business. That has been a drawn-out process, in terms of when political ads are served and when they are not served tends to be around specific moments that are relevant to the political calendar, but they are few and far between.
Q334 Jo Stevens: When? Can you tell me which year?
Will Scougal: I would not be able to give you a specific year. Rather than make a mistake, I would rather come back to you.
Q335 Jo Stevens: Do you want to write to us on that?
Will Scougal: Yes, thank you very much.
Q336 Chair: Bernie Sanders used Snapchat in the 2015 primaries and 2016 presidential election. That was pretty well reported. He was probably one of the outliers in the use of Snapchat as a tool for political communication. Given that, if it was being used in the United States at that time, was Snapchat used during the Brexit referendum as an advertising tool?
Will Scougal: There was content relating to the referendum on the platform at the time. The thing I wanted to confirm was whether or not it was paid for advertising or whether it was organic content served by us to draw attention to the fact that there was a vote. On the day of the vote, there was a filter running across the UK. I am not sure whether it was paid-for advertising or whether we as a platform were doing it.
It is worth noting that political content—from an organic point of view and from a premium content point of view—has existed on the platform for a while. There have been political shows. There are news programmes that cover political content. As a subject matter, it is certainly something that our users have been interested in and that they engage with.
Q337 Chair: As far as you are aware, there was political content on Snapchat related to the referendum period and you will write to us and let us know whether that was advertising or organic content?
Will Scougal: Yes.
Q338 Jo Stevens: If it is possible to say who paid for it that would also be helpful.
Will Scougal: If it was advertising and that is within the policy, you will have that, yes. I will check what we can share.
Q339 Jo Stevens: Going back to data collection and retention, how do you ensure Snapchat users’ data is secure when it is shared with third-party apps?
Stephen Collins: I will give you the overarching principle and Will can give you a few lines on Snap Kit.
The default is that we do not share data with third parties, period. However, we do have a product called Snap Kit, which allows a very limited amount of data to be shared with select third parties, which are individually vetted by us by a company. There are not hundreds of thousands of companies, maybe only 15 or 20 at the moment. It is a very small number. The only data that is shared with them is the user’s display name and the Bitmoji, which is the little avatar character of each individual. Beyond that, no further data is shared.
Q340 Jo Stevens: If I am on Snapchat and I go onto a third-party app, do I know when I am doing so that you are sharing my data—my name and my Bitmoji—with that third party?
Will Scougal: The experience is centred on the user having the choice and the power to choose what is happening. To give you an example, the Snap Kit suite of experiences centres around taking some of the Snapchat content and some of the experiences people have on the platform and distributing those more broadly; for example, taking the augmented reality experiences that people might have in the Carousel on Snapchat and placing those within video-based experiences on other platforms.
Q341 Jo Stevens: Do you specifically and explicitly ask me to consent to giving that data to those third parties?
Will Scougal: Yes, absolutely, but it is worth saying that we are not necessarily sharing specific user data. It is more about the experience happening on Snapchat that is accessed and made available to other platforms.
I will try to illustrate with an example. On Snapchat we have augmented reality. We have through Snapchat ARKit made those experiences available to some other platforms. For example, on Twitch, which is a gaming platform, you are able to stream your gaming while wearing a Snapchat lens. That does not need your user data at all without your given consent. That is the same with all of the ARKit experiences.
Q342 Jo Stevens: You mentioned that you only have 15 or 20 companies that you give data to through Snap Kit and you individually vet them. What are the criteria you use to vet them?
Will Scougal: First and foremost, from a policy point of view and from a values point of view, values are incredibly important to us and that is certainly something that is led from the top down from our CEO. First and foremost, the values of that platform need to be in line with ours in terms of being built for privacy and built for safety. We also look to align with that platform on that. There is a team vetting each individual partner and they work very closely with them to ensure the values and the build principles that we adhere to and champion are very much in place and present within that partner.
Q343 Jo Stevens: Are all your policies and values about privacy and safety on your website? Can I read them if I look?
Will Scougal: Absolutely.
Q344 Jo Stevens: I want to ask you about gaming. You mentioned Twitch. It has been trailed that you are going to become a gaming platform as well. How are you going to decide or determine which games developers you will work with?
Will Scougal: Again, similar to our partners with Snap Kit, while we want to democratise some of the features of Snapchat, we very much want that to be with partners who are aligned to our values. That would be predominantly our first step within that.
In reference to becoming a gaming platform, there is no official statement as to whether we are or are not heading in that direction. There are playful elements to the experience that exists on Snapchat but, over and above them at present, being aligned to what might be considered a gaming experience, I would say that they are more aligned for the most part to give users the ability to be more creative and more expressive in the communication they take part in through the camera, as opposed to what might be considered an immersive gaming experience.
Q345 Jo Stevens: What about gambling? On the basis that I am under 13, I can easily go on Snapchat—as has been proved this morning—and can go on to third-party apps via Snapchat. What stops me being able to do online gambling on Snapchat or via Snapchat?
Will Scougal: First and foremost, while we do not use data in the same way that platforms you might instinctively compare us to do, our targeting is as robust and rigorous as any of them. Certainly, while we are absolutely focused on the wellbeing and positive experience our users have, we also want our advertisers to have the most premium and quality experience they can have as well. For us, that means that their ads have to be delivered to the right people and they have to be in the right target group.
To focus on that advertising side for a second, it is absolutely important to us that not just the targeting works but that the quality of the ad works as well, so the quality of the experience that the user is being exposed to or targeted with. In that regard, you are not able to take part in a gambling experience on Snapchat. That is not something that we allow.
Q346 Jo Stevens: You are saying to me that I cannot access a third-party app through Snapchat that will enable me, as an under-13 year-old, to gamble online as part of a game?
Will Scougal: As part of the advertising standards laid out by the industry, our targeting is in line with and robust enough to allow gambling advertisers to advertise on our platform. In that sense, those ads can be served to people of the right age.
Jo Stevens: Or the wrong age.
Stephen Collins: Can I jump in? We are talking at slightly cross purposes. If I understand you correctly, I think you are talking about third-party gambling sites using the API to directly target our users to gamble on our platform using that third-party gambling site. We do not allow that. There are no plans to allow that. There is gambling advertising, yes, but, as Will has described, we take strong efforts to have the advertising of gambling appear to over-18s. Should a user click on one of those ads, they would need to have an account off Snapchat, nothing to do with Snapchat, on the gambling site or application. It is not part of our environment at all.
Q347 Chair: Will Snapchat allow games that feature in-app purchases to be accessed via Snapchat?
Will Scougal: Do you mean that the game launches within Snapchat and you can purchase things through the game?
Q348 Chair: Yes, those sorts of games where you can purchase things through the game. Would you have a relationship with a games developer like that? You cited Snap Kit and how you have relationships with developers where data and information is shared, but would you allow that with a developer that is creating games with in-app purchases as a feature?
Will Scougal: Snap Kit is very much about the Snapchat experience existing off Snapchat and that would exist within that developer space. Again, they would have to be aligned to our values and principles. In terms of the types of games that may be developed in the future, I am not able to confirm what may or may not be coming down the line and, again, would revert to the fact that we have not made any official statement about gaming becoming something that exists on the platform.
Q349 Chair: There is not a policy either way at the moment. It has not been ruled out but there are no set plans. Would that be a fair reflection?
Stephen Collins: It is a hypothetical because we currently do not have games of the sort you describe on the platform. Any future business decisions or business developments would take into account the safety by design and privacy by design principles we employ all the way across the platform. We would absolutely employ those as well. I know it is not an entirely satisfactory answer, but it is a hypothetical question. It is hard to answer.
Q350 Chair: I was interested to know whether there is a principle saying, “No, we would never do that”, or whether it is, “We would judge it on its merits if such an opportunity came along”.
Stephen Collins: Yes, and neither of us have been party to those conversations.
Q351 Chair: I want to clarify a couple of things for my benefit on Snap Kit. You said there were 15 to 20 companies that are the work partners?
Will Scougal: It launched with around 10 to 15 companies. It is something that we want to become more popular, and so I would not take 15 to 20 as a static given number. It is part of our business goals to take more of the experiences off Snapchat and make them present on other platforms. We fundamentally believe that augmented reality, for example, is an incredibly exciting and interesting space, not just for us as one of the most scaled AR platforms in the world but also from the point of view of other businesses.
Q352 Chair: What sort of companies did you start those relationships with when Snapchat started?
Will Scougal: We have had relationships with Twitch, where we have launched augmented reality into that platform and enabled gamers to wear AR as they are gaming. We have also had partnerships with other apps, such as Uber and Tinder, where elements of Snapchat have been shared from a user’s profile—with their absolute consent and with them opting into that experience—to become a part of the experience on that app as well.
Q353 Chair: To take Tinder as an app, for example, the way Snap Kit works is the Snapchat about the user is shared with the developer but not their friends’ data?
Will Scougal: In that particular example, it is the story of the user. Every 24 hours a Snapchatter has an option to post one-to-one to a friend to communicate or they can post to a story, and that small group of 10, 20, 30 or 40 people that they are connected to can then watch that story. That exists for 24 hours. The Snap Kit application on Tinder takes that story and puts it on the user’s Tinder profile, so it becomes a way of them showing more of themselves on Tinder. I believe that on the app at the moment there are static images. With the integration of Snap Kit, you have the ability to show your story from that day.
Q354 Chair: Data relating to that share is shared with the developer as well, so shared with Tinder in that case?
Will Scougal: At the users’ approval.
Q355 Chair: Does that default, though? The user cannot opt out of that?
Will Scougal: It will not happen without your approval. You have to opt into it for it to happen.
Q356 Chair: You could still use this tool without sharing your data back with the developer or do you just say, “I am not going to use it in that way”?
Will Scougal: The user would have to go onto Tinder and go through the steps to integrate their story into Tinder. There are two or three very considered steps to make that happen. It is not something that just happens.
Q357 Chair: If Snapchat user data is shared with developers in these cases are there any restrictions on what the developer does with that data as well? Tinder has a pretty universal data reciprocation policy with Facebook. If you have Snapchat users who are also Facebook users that are sharing data with a developer that works with both companies—
Will Scougal: We are absolutely not sharing user data in that way with anyone.
Q358 Chair: The Snap Kit page says, “We share only your data. We don’t share any of your friends’ data when logging in with Snap Kit”. That has become fairly industry standard. We have looked at that issue with Facebook at some length. That implies that your data is being shared with the partner that you are sharing that data with.
Will Scougal: If they are going to opt into the experience, there is the understanding that there is an element of data sharing. They are opting to share their story on another platform to that degree.
Stephen Collins: The concise point here is that the data that needs to be shared to provide the service will be shared. Otherwise, the service could not be provided. It has to be done consciously by the user. Also, reading from the site, “Apps connected with Snap Kit only need the basics”, no email, no phone number, no location tracking, no friend graph, and no demographic data. It is quite categorical. We cannot have a Snap experience on Tinder, for example, without that data being on Tinder, but the rest of the data that would perhaps in other more liberal APIs be shared is not the case with Snap.
Q359 Chair: I have a couple of things from me on advertising. Presumably, if I am an organisation that has a load of phone numbers, I can quite easily see how many of those phone numbers relate to users on Snapchat to create a Snapchat audience I could message. Is that correct?
Will Scougal: Sorry, if you as a user—
Chair: If I was an organisation—let us say Bernie Sanders’ campaign—as a consequence of being Bernie Sanders’ campaign, I have loads of mobile phone numbers. Presumably, I can take that information and see how many of those mobile phone numbers are also Snapchat users.
Will Scougal: We do not engage in that type of targeting, which I think is what you are talking about.
Q360 Chair: What is stopping that? Say I am a campaign and I have a massive contacts book. I open a Snapchat account and, presumably, I could find loads of Snapchat users straight away, just on the basis of having this enormous contact book of numbers.
Will Scougal: No, it would not work specifically like that. You are talking about looking through contacts and then trying to find users based on their phone numbers. That organic experience for someone to go and track down users is not one that exists. You would need to know their Snapchat name and you would need to have more information about them.
Q361 Chair: If I opened a Snapchat account now, I can ask Snapchat who in my contacts list is also on Snapchat so I can then link up with them.
Will Scougal: Yes, and you would offer them a request to become connected. They would need to accept that request. The inference is that, if they are a contact in your phone, they are someone you know.
Q362 Chair: Yes, but what if I had 5 million numbers in my phone?
Will Scougal: It would take you a very long time to message each individual person to become a contact.
Q363 Chair: It can be done organically; it is just a laborious process?
Will Scougal: Each one of those people would then need to accept you and say, “Yes, I know Mr Collins. I accept that person as a friend”.
Q364 Chair: Once you have built up a network like that, you can presumably organically message as many people as often as you like from the network in your contacts list?
Will Scougal: You could, yes. I would suggest that you were creating a nightmarish Snapchat experience. It is interesting and, again, it is a really good illustration of why we are not a social platform. If you had 5 million contacts on your Snapchat account and you did something funny or snapped something particularly interesting and then suddenly had 5 million people snapping you back, you would probably spend a year and a half trying to work through those snaps, watching the snaps and replying to the snaps. Then, after you have replied to them, they disappear. It would be like having a conversation with 5 million people as opposed to five really close friends.
Q365 Chair: Finally from me, for companies and organisations wanting to engage in promoted content shared through Snapchat, what are the categories that you offer to advertisers? I guess it would be based on location, if a user has shared location data with you and some basic demographic data, I would imagine if they have shared that with you as well. Then you said the way they engage with other forms of public content as well. What, for example, would that include?
Will Scougal: There are a mix of different signals that we can use to make sure that advertising is delivered to the right people at the right time. Our video targeting, for example, will look at the type of content that you watch to help us understand whether or not you are interested in sport or music, or maybe you are interested in sport and music and we want to take both of those things together.
There are inference signals and there are demographic signals. Also, we can combine that with what I like to think of as more of a branded experience within augmented reality that people would choose to take part in. You can own the carousel by being the No. 1 branded experience within that carousel by targeting the whole of the UK, or you can become a bit more refined with your targeting by employing some of the other signals that we get from the rest of the platform.
Location is also another really valuable signal that we think makes Snapchat quite unique in the way that it can deliver advertising experiences to people, and they will choose to take part in them. For example, over 90% of our audience share their location data with us because what we give them back is a value exchange of layers of creative that will frame where they are, what they’re doing, and they can use that again as a tool for self-expression in the communications they have with their friends.
Q366 Chair: Would it be right to say that you do not consider the content of people’s private messages with each other? For example, if someone discusses football with their friends through Snapchat that is not data that is picked up and is used for promoted messaging later?
Will Scougal: We do not have keyword targeting as an option, no.
Q367 Paul Farrelly: I am sorry; I have never used your little ghost. What drives me mad is people signing me up to WhatsApp groups without my permission. It also drives me mad that having downloaded Messenger on Facebook I cannot get rid of it, so I get bombarded by unsolicited people who I have no intention of replying to. It is more than I can handle, quite frankly.
I want to ask you on advertising, let’s say that I am a reputable toy company and I have a product that I think is going to appeal to 13 to 15 year-old girls. What is my route to advertise that with Snapchat? What is the route?
Will Scougal: There are a number of different ways you can do that. For example, if you had a TV ad and you had a minimal budget, you might want to use Snap Publisher to go and re-edit that TV ad to be suitable for our platform, to make it vertically formatted and sound on. Then using our self-serve tool you would be able to select your target audience. You would obviously sign up for an advertising account, select your target audience, select the flight time, select the number of impressions you want to deliver, and then be able to launch your campaign without necessarily dealing with anyone directly at Snapchat.
If you were looking for something on a slightly larger scale—perhaps, across multiple markets that involved a larger budget—you may work with a media agency that might have a relationship with us, and they would engage us alongside you, and they would work with our sales team to establish what is the right thing to do and which product formats we might employ to reach—
Q368 Paul Farrelly: Do you have the real-time auctions mechanism?
Will Scougal: Yes. There is an auction that exists for video advertising and for AR, and from a filter point of view as well.
Q369 Paul Farrelly: If I am a toy company who thinks I have a product that is suitable for 10 to 12 year-old girls, presumably we just pick the 13 to 14 year-old bracket, because most of them are 10 or 12 anyway?
Will Scougal: We err on the side of caution more than anything else. Safety and privacy are fundamental to the way that the app has been designed and built and to our values as well. If there was an inference that as an advertiser you were nefarious, first, advertising would not reach those people because of the signals that we pick up. They do not get served ads. Secondly, if you were doing it through the relationship that you would have with us directly we would be advising you in a different way.
Q370 Paul Farrelly: What, if any, process do you have to make sure that the advertising is appropriate?
Will Scougal: All advertising goes through a monitoring process and is reviewed. There are certain tools that we have internally that will pick up on the appropriateness of an ad. Again, we err on the side of caution and of anything is flagged as potentially inappropriate it will be reviewed by human eyes, not by anything automated.
Q371 Paul Farrelly: I raised with Facebook that I made the mistake of clicking on something that seemed to be a report but was clearly some sponsored content. Since then I have been hounded by ads that are not only just annoying, they are fraudulent, because the devices do not exist. There is a mechanism where you can report them. It is clearly not seen by humans or they pay lip service to it. Then you can say, “Do not show me ads like this any more”. That stops it for a little while but then the bombardment starts again. It is like being pursued down an endless beach by a bloke who will not take no for an answer. In this case, there are no sunglasses in the case. It is just relentless. These ads stalk you, basically. Can that happen on Snapchat?
Will Scougal: It is an interesting analogy. There are a number of different ways to target people, and one of the things that we advise, as a team and as a business, is that the quality of your creative is what will drive effective business results. If we go back to the idea that the platform is designed to put the power in the hands of a user, they can choose to engage with an ad or chose to skip past it relatively quickly without effort. The onus is on the advertiser to create advertising that contributes to the experience that people come to Snapchat for, whether that is amazing video content or incredible stories that sit within that premium content.
Q372 Paul Farrelly: If I do not want to see particular ads, can I effectively stop seeing them?
Will Scougal: Yes. You would just tap and skip past it. It would not feel like it was hounding you because it is totally within your power to just move past it.
Q373 Paul Farrelly: As I have mentioned Facebook, I have a final question. We have explored some of the issues that your chief executive was referring to when he said, after Facebook copied one of your features, “We would really appreciate it if they copied our data protection practices also”. We have explored that. Could I ask you: what is your experience of competing against Facebook? Do they compete fairly?
Stephen Collins: That is a hard question to answer succinctly. We focus on how we compete in the market rather than how our competitors are competing. If they are not competing fairly, we would expect competition authorities to hold them to account.
Q374 Paul Farrelly: Are there any instances where you have felt they have not competed fairly?
Will Scougal: I think it is interesting to define what we mean by “competition”. To a certain degree, we are flattered by advertising formats such as vertically formatted, turn-on videos and such as augmented reality being adopted by other platforms, because we see those as best-in-class mobile experiences, both for the users, for Snapchatters, and for brands as an advertising format.
To see those formats adopted elsewhere and to see them become popularised for us is a positive thing, because it just spreads a more positive experience and makes those platforms hopefully a better place. I certainly think that we are a platform led by innovation. We are not a platform that looks over its shoulder at what people may or may not be copying and we focus on moving our own platform forward.
Q375 Paul Farrelly: That is really succinctly dodged. They are not breaching your copyrights or your patents. They tried to buy you. You did not want to sell. You tear your hair out at them, but all is fair in love and business. However, are there instances where you have felt they have not competed fairly?
Stephen Collins: I think you will understand this is a very sensitive topic. Perhaps we can write to the Committee giving our perspective on the overall competitiveness of the digital marketplace in the UK, if that is helpful?
Q376 Paul Farrelly: That will be helpful. Clearly, there are anti-competitive practices that you do not see that might be terms and conditions for advertisers, and so on and so forth, or you might have some frustrations where you might think that the competition authorities might be better on the ball.
Stephen Collins: If you will allow us, we will write to the Committee with those thoughts.
Q377 Rebecca Pow: I want to focus on apps in particular, but in particular the apps that can make one look more beautiful. Can you quickly encapsulate how one gets holds of these apps to make one look more beautiful: thinner, smoother skin, or bigger eyes?
Will Scougal: First, I think you are centring around augmented reality.
Q378 Rebecca Pow: Yes.
Will Scougal: If it is a broader thing that you are talking about. Obviously, there is software like Photoshop or apps like PhotoME that are built around the idea of perfecting and beautifying photo imagery. There is a broad landscape of those types of apps.
Q379 Rebecca Pow: I think you can buy them for £4 an app; to get one for smoother skin, or one for bigger eyes. Is that right?
Will Scougal: To be honest, I am not sure when it comes to the broader landscape of those types of apps. Photoshop is a very expensive piece of software, for example, but if you are referencing augmented reality on Snapchat specifically—
Q380 Rebecca Pow: You then use it on Snapchat?
Will Scougal: You do not necessarily use those apps on Snapchat from a beautification point of view. I think what you are referring to is some of the augmented-reality experiences that exist on that bottom carousel.
Rebecca Pow: Yes, and the filters and things.
Will Scougal: There are a number of different elements to those AR experiences that drive use. They are fundamentally about enabling the user to add a layer of expression; a layer of fun to the communication they are having with friends. Augmented reality for us represents an incredibly exciting and interesting space, both from a Snapchatter point of view and an experience point of view, but also from a brand point of view.
At the moment, you are seeing things like you can add an element of style to a lens, to a photo, with some sunglasses or a flower crown, or you can add an element of fun through puppy dog ears, for example, or a user might choose not to augment themselves at all. They might choose to augment the world around them, for example with the dancing hotdog or another piece of animation.
I think this is moving into a really interesting space in that we are starting to see more utility-based AR come into the platform, so AR that has a functional use over and above enhancing or augmenting communication, which again makes it an incredibly exciting and interesting space to get into.
Q381 Rebecca Pow: I understand that, from a tech point of view, it must be exciting, but I gather from reading that there are concerns that some of these apps that can alter your body image are potentially affecting young people’s self-esteem. Particularly people with lower esteem would be more attracted to these images of themselves that they could “improve”. It is leading to concerns about what is happening to these young people and potentially sending them down the line of, first, spending more money on cosmetics and things to make themselves look better, but, secondly, potentially it has even been suggested by some experts that there might be a link to teenage suicides with this whole self-esteem body image being part of this. Is there anything that you are addressing or have come across?
Stephen Collins: Maybe I can jump in on a slightly higher level point about lenses and lenses that are available on Snapchat. Of course, a lot of people are not familiar with opening straight to a camera lens. The lenses are not only a playful and fun way to interact with friends. They also break down barriers of expression as well, because people can use them to do fun things.
I am not sure that there are any lenses that really are beautifying lenses. They are usually funny, or bizarre, strange, and they are not altering a person’s body. The Snaps that are sent are ephemeral, so they disappear. It is not like you are posting to a social media page where the perfect image is kept in perpetuity. These are Snaps that are sent one-to-one between friends.
Q382 Rebecca Pow: I have some young people watching this, and they say, even when you put the puppy dog ears on, it makes your face look prettier. You are instantly affecting people’s image. I am led to believe that this is having a real negative impact, particularly on the image of women and young girls. Clearly you just think it is a fun thing with no other impacts.
Will Scougal: Just to take a step back, I think that body image issues and anxiety among young people around how they look is obviously something that is an important issue. It is something that we see in celebrity magazines, we see in advertising, and we see in other pieces of content a lot in terms of images being doctored and Photoshopped to look like something they are not. Those are being displayed and distributed as fact or as reality. Whereas, what we look at with Snapchat and with lenses is an environment where augmented reality is adding a layer on to the real world that is not—as Stephen said—specifically about beautification. It is about adding a layer of fun, or expressiveness, or creativity.
Q383 Rebecca Pow: I have also had it put to me that, increasingly, cosmetic surgeons are being presented with these pictures of celebrities who have used various apps to make their bodies look in a certain way—maybe you might call it beautiful—saying, “I want to look like that”. It is an increasing issue and they are finding this very difficult to handle. Do you accept no responsibility that you are causing any of this?
Stephen Collins: We do not want to just sit here and say, “From a technical perspective everything is fine”. I think it is a fair point to make. There is research being undertaken. Our chief medical officer recently was looking at mental health and time spent online. These are really valuable insights rather than anecdotal insights. More could be done there. If there is anything that we are doing that empirically is seen to affect young people in this way, we would of course change the product. There is no doubt about that.
Q384 Rebecca Pow: It is interesting because it has been put to me as well that part of this society we are creating, through all these devices, is giving a whole different image to women, especially young girls. It has put to me by quite a few young girls that it gives the impression that women and girls are never good enough. They always have to do something to improve themselves. Are these apps exacerbating that, and should you be thinking about this a great deal more?
Stephen Collins: We can only talk to Snapchat itself rather than the wider ecosystem. As Will has described, there are certainly a lot of applications out there that allow manipulation of images to a quite extreme and sophisticated level. On Snapchat the lenses cannot alter your body—it is just a facial thing—and they are fun. Even with the skin smoothing there are dog ears, or bear ears, or something. There is an element of fun in all of it, and if the fun is stopping, we should address that.
Q385 Rebecca Pow: Also, rappers are starting to pick it up. There is a rapper called Kendrick Lamar, and he has been referring to creating a culture where women cannot be trusted because they basically Photoshop themselves. I think there is a great deal more going on than you are perhaps realising with the influence of all these apps, and lenses, and filters.
Stephen Collins: That is right. There is a multitude of things going on. It is right that we are moving in a technological direction that is pretty fast. I am a parent. I have two daughters and two sons, and of course I am super conscious of this kind of thing.
Q386 Rebecca Pow: Me too. My son uses it and has a great deal of pleasure out of it. However, I know for other people they could easily be influenced if they were vulnerable. That is my point. That leads me neatly on to the fact that it is a very fast-changing technology. One might say, from a tech point of view, it is clever, it is fun, and it is potentially going to make money for people. You can change that very fast.
However, the people who are using it, who are then being influenced by it, cannot change fast. You are perhaps having a knock-on effect on their lives without really addressing what you are doing to think about it. Do you have any sort of policy about this Snapchat dysmorphia, as I think it is called?
Stephen Collins: There are a couple of things. I think we need more evidence and more research; I really do. Not just for Snapchat, I mean across the industry. It is very important. I think that should proceed.
Q387 Rebecca Pow: You would wait for more evidence. I have given you quite a lot of evidence. There is some information here from Newcastle University. There is a great deal of evidence coming out.
Stephen Collins: We also take feedback from our users, of course. If users are unhappy, they will tell us. We do focus groups as well. All these things are tested to the best of our ability, and to the best of our knowledge what we provide with the AR carousel is overwhelmingly a positive experience for our users.
Q388 Rebecca Pow: Finally, Will, I believe on your write-up one of your roles is that you are working with businesses to realise their potential and their objectives. Do you ever consider that you might be doing that to the detriment of some of the people that are using your system?
Will Scougal: In terms of the campaigns that we create or the creative we create, in my opinion it is absolutely fundamental to the success of the campaign for that creative to be contributory to the experience that the people are coming to the platform for. Any campaign that may cause a negative experience in my mind is one that is set up to fail. It is absolutely not something that we would seek to do in any way whatsoever. Outside of that, we very much work in partnership with our brand partners to make sure that what we are delivering is both positive for the users and positive for the brand.
Q389 Rebecca Pow: If you are encouraging, for example, young girls who want to look more beautiful—like these celebrities who have doctored themselves up with the apps—and then are buying more cosmetics, your business is winning, isn’t it? It is playing on the weaknesses of a young teenager and it is ending up with somebody selling more cosmetics, and you are working on somebody’s vulnerability.
Will Scougal: I appreciate the point of view, and I will just illustrate a beauty campaign for you by way of helping to hopefully explain and allay some of those concerns. Within beauty advertising—and I have worked in advertising for a long time—there are historically campaigns that have run over a number of years, possibly since the dawn of beauty advertising, where the role of the ad is to show the viewer or the audience what the product is, how you use it and what it does. Those are fundamentally the three core ingredients to beauty advertising.
For an example on Snapchat, you could bring in a video ad that may show you how to use some makeup, and it would show you what it did through video. That is a format that has existed on television and it exists on other platforms and across the internet, and for my mind is standard in that regard. What we have been able to do with augmented reality is turn advertising that you would sit back and watch—for example, you may watch an influencer doing a beauty tutorial—into something that you can take part in yourself, enabling you to try the makeup on through augmented reality.
That in itself is purely demonstrating again what the product is, how you use it and what it does. It is following a formula of beauty advertising that has existed for a number of years. It is not in the beauty advertiser’s best interest, in that regard, to make anything happen as part of that experience that is not true to the product and true to the brand. As the brand, that is an example of how their influence and our partnership with them would ensure that that particular augmented-reality experience is not doing the things that you—
Q390 Rebecca Pow: Just on the videos, can you target the videos by gender and age?
Will Scougal: Yes.
Q391 Rebecca Pow: Could you target them to under-16s?
Will Scougal: Not if it was against standards policy, but if you are a brand that is interested in under-16s and you are advertising to them in general, and that is within policy, yes, you can target them.
Q392 Rebecca Pow: Stephen, you said that you were interested in some of the things I have raised as concerns. Would you share any of your research with us, or do you share what you have found with other regulatory bodies to help in this whole area that I have raised about women and girls, and the knock-on effects, and what I would call deleterious effects on society? Do you do that?
Stephen Collins: Yes, let me come back to you. Let’s see what we have, and I will come back to you on that.
Q393 Clive Efford: I was going to come in on that. Why should you go away and see what statistics you can make available? Why would you not make them available? You have accepted that there is harm being caused. You said so in answer to a question just now. You said, “We need more research”, so surely you would make your data available to ensure that that research can take place.
Stephen Collins: Make that data available to?
Q394 Clive Efford: To people who want to carry our research into any harm that may be caused.
Will Scougal: I think the point about there is research needed was more to address the fact that doctoring of images and the effect it has, maybe from a celebrity magazine point of view, or from apps that are specifically created to doctor images and physical appearance—things that are contributing to some of the points that were raised—rather than specifically research on just our side. It was more of a societal point on research than specifically Snapchat.
Q395 Clive Efford: If you accept, as Rebecca Pow has set out, that there are instances where there is, albeit a minority, a vulnerable group that can be harmed by the lenses that allow them to alter their images, if you were us, if you were this Committee, or you were the Government, what would be the thing that would be of most concern to you?
Stephen Collins: I genuinely think it is a wider societal question. I do not think it is focused down just on one application or one set of applications. It is part of a wider discussion, as the CMO is undertaking in her research to understand. If I were in your position, I would want that to have more empirical research to understand: where are these drivers? What are the triggers and where is this coming from? Is it from the application itself? Is it from other pressures in society? Is it because of technological pace? There are a whole variety of factors it could be. There is correlation here but is there causation? I do not know, and that is why empirical research is so important. I am not trying to hide and give you an evasive answer. The basis for any decision, I think, by Parliament is around empirical evidence.
Q396 Clive Efford: I keep coming back to the comparison with gambling and the excuses that they were making about the harm that, for instance, the FOBT machines were causing. It all came back as, “It is a societal issue and we need more research”; exactly the same phrase. If we know that a product is going to cause harm—albeit, to a minority—why wouldn’t we want to know about those consequences before we let that harm be caused? I go back to my original point. Wouldn’t you want to make all your relevant data available to ensure that researchers can look into what the impacts of your lenses and other applications on your site are causing?
Stephen Collins: So that I do not misunderstand you, what data are you referring to?
Q397 Clive Efford: For instance, first and foremost, we need to know who is using it and how you are targeting it. For instance, you have said that you can target adverts at young people by age. Are those products, for instance, beauty products that can be targeted at young women in particular? Who advertises through your lenses? Can particular advertisers ask specifically that they can send their adverts targeted through people who use your lenses?
Stephen Collins: By extension, could it be a beautification lens? Is that the implication?
Q398 Clive Efford: Exactly. Yes.
Stephen Collins: I had not understood that. Okay. I do not know the answer, but let’s take that away. I do not want to avoid your question. I hope I have not given that impression. We will take that away and we will come back with that data.
Clive Efford: You are going to think about it and write to us. Okay.
Will Scougal: Lenses can be targeted to a certain audience, and that could be a beauty product, yes.
Q399 Clive Efford: When someone comes to you wanting to develop a lens, have you ever turned one down? Have you ever said, “No, we cannot do that”?
Will Scougal: As a business, I cannot possibly make a statement on that. If I am just going to guess, there probably is an instance. I will go away and have a look and come back to you with confirmation, if that is okay. From our side, the business side that the team supports is there to help advertisers meet their objectives. I have not come across an instance yet where we have had an advertiser that we have refused to work with or help to understand the platform, use the platform to achieve that objective, when it is within the advertising policies that we have laid out and when it is aligned to our values as a business.
Stephen Collins: I cannot quote verbatim from our advertising policy, so again I am going to have to come back to you. I am pretty sure in the advertising policy there are criteria there that must be met, and if the policy is breached then the product would not be launched. Then I am sure there must be, on the advertising side, guidelines around any kind of harmful content.
Will Scougal: Yes. For example, in different countries there are different laws and standards that are laid out for different advertisers. For example, in North America marijuana has become legal, yet we have taken the decision as a business to not allow marijuana companies to advertise on Snapchat at the moment until policy is clear. That is one instance that comes to mind.
Q400 Clive Efford: Do you do any of your own research into what is going on?
Will Scougal: Yes, absolutely. We do. We have an excellent research team. They look at a number of different things. It really splits into two different areas, one specifically from an outwardly-facing point of view, the research behind campaigns and whether or not they have worked. We have third-party partners that we work with to do that, so people like Nielsen, Millward Brown, and Moat, who help us measure the efficacy of a campaign. That is a particular vein of research.
We also have research that we do that is more qualitative and empirical into user behaviour, so what people may do or spend time doing on the app and how they feel when they are doing it, the positive effects that the app might have versus other platforms, and what makes us unique in that space. We will continue to do that.
We have a core audience that is in some respects quite unique to Snapchat, the millennial, Gen Z audience is very active on our platform, very present on our platform, and we do feel that we have the ability to do research into that specific demographic to garner insight to help us build a better platform and a better environment for those people.
Q401 Clive Efford: Right. You do have quite a good understanding of who is using your site and your platform and what is motivating them?
Will Scougal: Yes, to a certain degree. We did a recent piece of research where we looked at Snapchatter behaviour—and it is no secret that people use multiple platforms—and what they may be feeling, or how they may be feeling, and what they may be using the app for. Overwhelmingly, that audience research came back as Snapchat is a positive place. It is a creative space and it is a fun space, for example.
Q402 Clive Efford: Yes. I would expect that, and I would accept that. The point is that we are not dealing with the overwhelming number, are we? We are always looking at those areas of harm that are among the minorities. That is where we have to delve into and find out what is going on. Do you have concerns around biases being reinforced by idealised images that filters create?
Will Scougal: If we boil it back down to: we are a platform that is cautious by design, built fundamentally with privacy and safety at the heart of the way we bring new experiences on to the app. There are teams that work alongside the design team to ensure that what we are doing is creating a positive environment, and that is something that we will constantly and perpetually look at. If there was ever a sense that there was a negative experience being created, it would be something that we would take into review very quickly.
Q403 Clive Efford: The overall thing is that this is fun, the lenses are fun, and people have fun messing around with this stuff. However, I go back to the point that there is a minority where it stops being fun. Your research must show that. There must be those areas that cause you concern. What is your reaction? How do you as a company react when you identify that from your internal research?
Will Scougal: That is an interesting point and possibly one that is again part of a broader societal issue, and possibly part of broader issues to do with the individual themselves outside of Snapchat. I certainly do not think that that is a point that is not important and not something that should be considered, not just by us but, again, the broader ecosystem.
Stephen Collins: Again, if I may, we will take this away. We could particularly ask the designers how they react to reports of negative experiences, for example. It is not clear to us here today. We do not want to give you the wrong answer. As part of our response we will come back. It is an interesting point at the end of that long tail; the vast majority of users are very, very happy but what happens to those at the end who maybe are not happy?
Clive Efford: This is not a question, it is just a point, but in a similar situation the gambling industry has ended up doing itself enormous harm because it did not recognise the damage that was being done to that minority. It just rumbled on and it kept coming back. I suggest to you that this is the same situation that the tech industries are going to find themselves in.
Q404 Chair: A couple of final questions from me. Obviously, you will be aware there was a lot of concern raised over the weekend about content from the Christchurch shootings being shared through social media, particularly Facebook and YouTube. Was content related to that terrorist act shared through Snapchat?
Stephen Collins: It was not. First, let me just say, obviously our sympathies to the victims of the shootings in Christchurch, the bereaved and the families involved. It was an absolutely horrific incident. It seems Snapchat was not involved at all. The user had no Snapchat account and, as far as we can see, there were no incidences of the video being shared.
Q405 Chair: When you say “as far as we can see”, how would you know?
Stephen Collins: Because we received one report from Norwegian law enforcement where a user there was sharing after the event. There was nothing live. There is very limited opportunity to share any live content on Snapchat. There is no content.
Q406 Chair: Most of the sharing was done after the event, anyway. There was obviously a component of live streaming.
Stephen Collins: The safety team now is fully vigilant. We remain vigilant. We are not complacent. I do not want to make out that we are complacent about this because Snapchat is not a good broadcast platform. That is a clear basis for why it was not used in this case. You cannot send generated contented to a vast number of people on Snapchat. It just does not function like that. There is too much friction.
Q407 Chair: You said you have one report from Norway of people sharing, after the event, content on Snapchat related to Christchurch?
Stephen Collins: Yes. Right.
Q408 Chair: Have you acted against the accounts that were sharing that?
Stephen Collins: Yes, we have.
Q409 Chair: They have been closed down?
Stephen Collins: Yes. We co-operated, of course, with the Norwegian police.
Q410 Chair: Other than you receiving reports of harmful content like that being shared, do you have any way of noticing as a company whether content like that is being shared?
Stephen Collins: Yes. On the public areas of the platform everything is moderated. That is algorithmically, and all the important stuff is done by human moderation. Essentially the Discover platform, which is the main public platform on Snapchat, is fully moderated, so it is closed. A user cannot post there without some form of moderation.
Q411 Chair: But user-to-user communications, if one user was sharing clips from the shootings with another user?
Stephen Collins: One-to-one?
Q412 Chair: One-to-one.
Stephen Collins: Yes, we could not see that.
Q413 Chair: You cannot see that. So, unless someone reports it to you, you have no way of knowing it is happening?
Stephen Collins: Correct. But that is the same on any messaging platform. Again, we go back to the privacy of communications for one-to-one communications and the assumption that people have—since the earliest days of telephony right up until the present day—is that private communications one-to-one are not blanket monitored.
Q414 Chair: When you talked earlier on about the signals you pick up about users and what they interact with, those signals are based on what they interact with in the public sphere of Snapchat?
Stephen Collins: Yes.
Will Scougal: Yes.
Chair: That concludes our questions this afternoon. Thank you very much for your evidence.