Oral evidence: Immersive and addictive technologies, HC 1846

Oral evidence, 26 June 2019.

Witnesses

Alex Dale and Adam Mitton.

Chair

Good afternoon. I welcome people 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 welcome representatives from the games company King, better known as the makers of Candy Crush Saga, to give evidence to the Committee. I would like to start off by asking a question that has taken the Committee’s interest during the inquiry so far. It is about the use of data gathering by video games companies to fine tune or improve the design of games. Alex Dale and Adam Mitton, could you tell us a little bit about the sort of data that your company gathers in order to improve the design of its games and also how it seeks to characterise players by game play?

Alex Dale

Thank you. We are happy to be here. I will give a little bit of background on the business, which will help give some context to the question you are asking. I will be brief. We have a lot of players, 270 million players worldwide every month all over the world in every single country. We are part of a US public company, so we have to be very careful about what data we share. We have thought about what is available to share without commercial consideration and we will try to be as helpful as possible with that. The most obvious point is that we are a free-to-play games company. That means that the game in its entirety can be played and completed without payment, and that is the case for the vast majority of our players. I will switch now to market data, so this is not King data and you should not assume that King is either better or worse than the data I am about to give but it is indicative of the category of games that we are operating in, free-to-play games. Roughly 3% to 5% will pay in any one month.

Chair

On that figure, just so we are clear, is that 3% to 5% of total players?

Alex Dale

Of all players, yes.

Chair

We have had evidence from other games companies as to what people who play regularly might constitute. I would be interested to know how you would define “regularly” and what percentage of regular players would pay.

Alex Dale

I will explain why people play, which will then inform the answer to that question as well. Essentially, because you can complete the game without paying, the reason you pay is primarily to complete the game more quickly or to progress more quickly. Therefore, there is not an obvious relationship between the amount of time you spend in the game and the amount of money you pay in the game if you are a payer. The top 20 spenders of last year play less than me each day. They will spend a couple of hours playing every day. That is the order of magnitude.

Chair

How much do you play?

Alex Dale

I probably play an hour of Candy and an hour of other games minimum every day.

Chair

Please continue.

Alex Dale

Going back to the data point, when we make a game we are thinking about the audience and our core audience is female 35-plus. That is our core demographic. We do not make games for kids or market games for kids. It is their mums, grandmums and, to a reasonable degree, their dads and grandfathers too. The games are designed to be played primarily on mobile in bite-sized sessions—that is the phrase that we use a lot—and an average player will play for 38 minutes a day.

Chair

I think this is important because you have millions and millions of people who have registered. There probably will be a large number of those that are dormant accounts. I think I have heard that only about half the people who are registered play frequently. For these statistics to have meaning for the Committee, I think we need to look at people who are regular players of the game. Let’s say they play every week. How many of those people who play every week will spend money playing? Of those people who play every week, how long will they spend on it? I think the average stats for all users does not give us a true picture of what a typical player who plays frequently will do.

Alex Dale

We will try to deaverage it. The 270 million is active players. It is not dormant accounts.

Chair

How do you categorise an active player?

Alex Dale

They need to open the app. They may or not play but generally they do. A lot of people will have the app open while the television is on.

Chair

How often do they have to open the app?

Alex Dale

Once a day.

Chair

Active players are people who open the app once a day?

Alex Dale

Yes, but generally they play as well. They might be doing something else. If the telly is on and they have Candy Crush open, the adverts come on, they switch to Candy Crush, the adverts finish and they go back to the telly, or they are playing on a commute and they don’t want to miss their stop, obviously. Deaveraging it, 3.4% of our daily actives play for three hours a day or more. If I was playing only Candy Crush I would be in that category myself. It is 0.16% of the 270 million play for six hours or more, but there we have this issue of the app being open, multiscreen usage being not uncommon[1]. We should also remember that a fair number of our players are retired. We have players in their 60s, 70s and 80s and the oldest player I have been in contact with was 92 or 94 years old. It is people who have plenty of time on their hands as well. Back to the data, we make games. We are also a business but we make games. We want to have games that are fun and also require genuine player agency—you have to think in order to play. We also want games that are fair, that the rules apply to all the players. When it comes to the use of data in terms of the design of the game, we don’t differentiate between players. Level 500 of Candy Crush is going to be the same for everyone. We are not altering the design based upon the profile of the player or using algorithms or anything like that. What we will do is try to balance a level in a game so that it is appealing and sufficiently challenging at the same time. When Candy Crush was first growing there was the infamous level 65 and level 65 was just too difficult. We had to change level 65 because people were dropping out of the game. We changed level 65 to make it the right level of difficulty. When we speak to our players they very much appreciate that there are some levels that are difficult and some levels that are hard and they enjoy that variety in the game. We have a lot of information about how people are playing the game. We use that to make the game fun and challenging in the right mix but we don’t differentiate with that data.

Chair

In changing the user experience of the game?

Alex Dale

Correct, yes.

Chair

The data is designed to analyse levels of difficulty. Do you make assumptions based on enjoyment, based on the amount of time people play?

Alex Dale

Yes, essentially. We want to get the right level of difficulty for each level. That is the way it works.

Chair

You have data based on the usage player behaviour of each player?

Alex Dale

Yes, although typically we look at this data in the aggregate, so it is statistically analysed. It is not looked at at the individual player level as individual players. It is across millions and millions of players: how are millions and millions of players enjoying or not level 65, level 100 or whatever?

Chair

Presumably you can identify individual players and you can identify data about their game, about their player patterns of behaviour?

Adam Mitton

Let me chip in there. As part of the compliance with GDPR when that came into effect last year, we did a big restructure of the database so that the individual identifiers were separated from the aggregated game play telemetry type data. The majority of the business has access to only the aggregated telemetry level data and not the individual identifiers. Some bits of the business do need to have access to individual accounts, such as customer services and things like that, but generally speaking the business works with the aggregated data rather than the individual player data.

Chair

What partners do you work with, do you share data with?

Adam Mitton

In what sense, sorry?

Chair

If you have player data, do you work with partners to share the insights and information about player usage of the games?

Adam Mitton

Not that I know of, no. The way our players play our games is pretty sensitive to us and something we protect pretty closely.

Chair

We will come on to talk about the data you share with Facebook later on. That is certainly something we will be raising. How important are data gathered about games play to the design of the game, to adaptions and the future design of the games? Are you learning from the way people are playing to improve the design of new levels or new aspects of the game?

Alex Dale

We are in the free-to-play games business so it is very easy for players to switch from one game to another because they are free to play. We want to keep the players we have entertained. We want it to be fun and exciting and challenging for them. That is both through the game itself and changes we might make to the game and through the side quests or side challenges that exist within the game. Maybe I should explain a little bit about how the games are designed in the first place. Would that be helpful? Game design is a very peculiar art and science. It combines left brain and right brain. The chap who invented Candy Crush trained initially as an accountant, then he ran an online dating service and then he decided to be a games designer. There was not a degree in games design for him to follow.

Chair

He ran an online dating service before?

Alex Dale

Yes.

Chair

He has a history of data gathering, data analysis, behavioural insight, psychology.

Alex Dale

Well, he is analytical and also creative and that is the left brain-right brain side of things. When we are designing a game we really start with what is the core concept here. It could be matching candies, it could be clicking blockers. That is called the core loop and we are thinking about that in the context of a target audience. As I mentioned earlier, that will be typically female 35-plus. We then go into a prototype stage where we have built a rough model of the game. In the Barcelona studio, for example, which I look after, we will have something called a mothers’ day. This is when the developers and the designers and the artists bring their mums into the studio to get initial feedback on whether the prototype concept is appealing or not, whether the rough art is appealing or not.

Chair

You are a company that acquires a massive amount of data for games play and there are many people who work for the company whose job it is to analyse that data and use that data to design new functions of the games and so on.

Alex Dale

Yes.

Chair

I suppose the concern for us would be is that data used to encourage people to play more.

Alex Dale

There are two points in response to that. First of all, I should stress that it is both an art and a science. You cannot create a compelling game with data and science alone. There needs to be creative design spark and intuition that is very much part of what we do. Second, we do want people to play more. We don’t want them to switch to other apps, and I have given you some of the statistics about how much people play. There is an e-mail we received a year or so ago that really makes the point quite clearly. I can paraphrase it. This was “What King did for my mother”. It was sent to our CEO, “My name is Y and my mother E passed away at the age of 48. She was sick with a rare cancer. She was invalided for three years. I convinced her to get a smart phone and she has been playing Pet Rescue, Candy Crush, Candy Crush Soda and everything like that from you. She is at level 900. She had a positive outlook on her disease at all times and your game helped with that and kept her sane. Thank you for what you’re doing.” I am paraphrasing. There are going to be people who like to play our games a lot and we don’t know why they are doing that but we see that they do.

Chair

How many people work in the player insights team at King?

Alex Dale

In player insights? Probably about six.

Chair

Who do they work with across the business?

Alex Dale

The game design teams and the marketing teams.

Chair

Their job is really to aggregate this data and use it to help design games.

Alex Dale

No, they are doing the qualitative research. They will be doing things like talking to our players, particularly those who are either looking at a new game or a marketing campaign. That is very qualitative, focus group, survey type work. Then we have business performance people who are looking at the numbers and the business, and that is a separate group.

Chair

I was looking at this earlier. I see that the insights team is looking to hire people to work in it, particularly a senior player insights manager, so I was looking at the job spec for that and the role within King. There it says that, “You are required to develop learnings by consolidation and synthesis across multiple datapoints and creation of thought pieces”. That would sound like it is pretty quantitative not qualitative.

Alex Dale

No, the multiple datapoints there will be from surveys, from qualitative surveys and quantitative surveys but not the sort of online data I think you are referring to.

Chair

It also says that, “Your job is to incorporate behavioural data and research, leveraging our internal understanding of players and their gaming habits and practices”.

Alex Dale

Yes. You would be doing the job badly if you did not acknowledge the behavioural data from game play as well, but the specific job you are talking about is a qualitative research type job.

Chair

It says “incorporate behavioural data” and also “leveraging our internal understanding of players and their gaming habits”, which would suggest that you do have profiles of individual players and their gaming habits that people can analyse.

Alex Dale

There are two points. Again, it is both art and science and on the art side I would put very much the qualitative focus group type.

Chair

I think it is fair to say we are focused on the science side.

Alex Dale

On the science side, we run the games for the average player. We don’t split them into clusters or segments or similar. We will differentiate between a payer and a non-payer, absolutely, but we don’t divided the overall 270 million into apples and oranges and pears and clusters like that. The game is run for the average.

Chair

It also says here about the skills that people need to work in this job, “Significant quantitative experience, ideally including regression analysis and modelling”.

Alex Dale

Anyone in a survey orientated—like someone working for Nielsen today would be expected to have those skills. We are not looking for a PhD in mathematics or computer science, which is where the behavioural people would sit. The job you are referring to is very much on the qualitative research side.

Chair

It barely mentions qualitative research here at all. I used to work in an advertising agency. I have some idea of what qualitative research is and it does not really mention that at all in this. It seems to be all about analysis of company data, quantitative metrics, quantitative research.

Alex Dale

That role is about things like Nielsen and MORI. It is equivalent to Nielsen and MORI.

Chair

It does not really mention working with those sorts of public datasets either.

Alex Dale

No, but that the type of dataset we would be creating from surveys and similar in that role. But if you want to talk about the other roles I am quite happy to.

Chair

What is interesting is that you look at this and what it suggests is that there are people—and I am not saying there is anything wrong with this—whose job it is to look at company data and from that company data, which is based on insights of players, to work to improve the company’s understanding of players and the design of the game around players. I would imagine the regression analysis is to understand when people start to disaggregate from the game or increase their activity, to understand people’s behaviour within the context of the game, not just basic data about them.

Alex Dale

To your first point, absolutely, yes. We have lots of people looking at how we design the game to make it more appealing. There is no doubt about that. I used to run the team that this role is being hired into and the regression analysis will be about taking survey data and doing basic regression analysis on that. That player insights role is really on the soft side of marketing analysis. On the quantitative side we have lots of data scientists in King who will look at that behavioural data and analyse.

Chair

How many do you have, roughly?

Alex Dale

I think we have about 50 or 60. I am not sure.

Chair

Their job is to analyse the data of player behaviour?

Alex Dale

Yes.

Chair

We have spoken to other companies who have similar functions in their businesses and they create multiple different categories of types of player, based on a profile not so much of individual players but there might be 100 different player profiles. Do you have something similar at King?

Alex Dale

No, we don’t. As I said earlier, the game is optimised for the average, which is a very technical way of saying we make the game fair, the rules are the same for everyone. We will look at people who spend in different tiers of spend. We will also look at what type of handset is being used, because that can make a difference to the coding of the game: do you make a game that is available to relatively older models of Androids or iPhones or iPads or similar or not, which obviously affects the audience potential?

Chair

You are categorising people by spend levels and you obviously can identify people who play for longer periods and shorter periods from the data you have given us. We discussed this last week with EA and Epic at some length. What responsibility do you have or monitoring do you have in place to identify people who may be playing for excessive periods of time or spending excessive amounts of money within a short period of time?

Alex Dale

I will answer each of those in turn. On the excessive time one, given the story I read about the hospital experience, it is very difficult to know what excessive is in that context. If we had a block per day of X hours, that would have cut off that patient from her ability to enjoy the game and, as I mentioned, we have a fair number of people in their 60s, 70s and 80s playing Candy Crush. I gave you the statistics: 3.4% play for three hours a day and that is daily users. It doesn’t mean they do that every day. It means that in any one day 3.4% will be playing for more than three hours and 0.16% for more than six hours, or at least they will have the app open for more than six hours.

Chair

There are press reports of people who have played for 18 hours straight, people who may have a serious addiction and play for excessive periods of time. There will be people who are spending excessive amounts of money and your systems can spot this. Do you think you have any obligation to intervene and say to players or even to suspend the account of a player who may be behaving in an excessive way?

Alex Dale

I did not answer your money question, so I will come back to that. We look at how people spend. The highest level of spend in Candy Crush in 2018 worldwide was $2,600 in a day. That was the top. That sounds a lot and it is a large amount of money, but what that player bought was 25 $106 bundles that they then used over a seven-month period, so that works out at roughly $370 per month. After they had depleted that—because when you spend you are buying gold bars that you only use in the game—they bought another 10 packages of $106 each. There was a sale on at the time, so they were making a rational decision, because they are obviously a Candy Crush fan, to spend that money in one go in the game. It is down to player choice whether that is what they want to do and they decided to buy that level and rationally deplete it and then buy again. We have customer support available in, I think, 24 languages. From the 270 million, we have between two and three contacts a month from people who are concerned about having spent too much money or time in the game.

Chair

That is people self-referring. What I am asking is: is it is your policy as a company to look at a player who is suddenly spending a lot of money that they did not spend before or whose game play has gone on a path where they are playing for longer and longer periods, well over six hours a day, or spending increasingly large amounts of money? Is it ever your policy to contact that player and say, “We may suspend your account. We may intervene to see are you content with the amount—are you aware of how much you are playing and how much you are spending”?

Alex Dale

We used to do that and we would send an e-mail out when a player’s spend was $250 in a week for the first time. It was an e-mail that said, “We notice you are enjoying the game a lot at the moment. Are you sure you are happy with this?” E-mail is very effective for the most engaged players, including those who spend a lot of money. We got back, “I wouldn’t spend the money if I didn’t have it” and things like, “I’m fine, please leave me alone”. We felt it was too intrusive so we stopped doing that. Obviously with the WHO announcement we should look at this whole area again, and we will do that, but we have tried it before and players said they didn’t like it.

Chair

There was a survey done by “Time” magazine of 1,000 Candy Crush players, 30% of whom said they thought they were addicted to it. What was your figure for daily players? Was it 270 million?

Alex Dale

Yes, monthly players.

Chair

If 30% of those were addicted to it, that is a lot of people. Even if we just took the general theory that about 1% of people who play games are addicted to them, that is still a lot of people who are addicted. What I am not getting is any sense that you feel you have a responsibility as a company to identify people who may be addicted and are either playing for excessive amounts of time or spending excessive amounts of money. You are happy for people to refer themselves to you as having a problem but you don’t see you have a responsibility to identify them yourselves.

Alex Dale

We also talk to players who are high spenders and spend a lot of time in the game.

Chair

How many do you talk to? You have 270 million people playing a day.

Alex Dale

Remember it is a very small number who spend at high levels or play at high levels—0.1% for six hours a day[2]. When we speak to them they are happy with what they are doing. We have player meets where they might come into the studio, meet the people who make the game and it is generally mums and sometimes their husbands come along and they are happy with what they do. We also survey. We did a recent e-mail survey of players who are in the high-spend category. The majority believed that it is value for money and is a great hobby—they like it.

Chair

By your figures, even taking 0.1%, which I think is rounded down slightly from 0.16% but it is easier maths, that is still 270,000 people a day who are playing it for over six hours.

Alex Dale

A lot of people use the word “addicted” in a colloquial fashion, “I am addicted to ‘Game of Thrones’, I am addicted to ‘EastEnders’”.

Chair

It is not the same though, is it?

Alex Dale

No, of course it isn’t, so we need to understand where the science is in this area. We absolutely acknowledge the seriousness of the WHO decision. We note that the American Psychiatric Association does not agree with the WHO. I think you had a Professor—

Chair

What does King think? Does King agree with the WHO?

Alex Dale

We are not psychiatrists. We are happy, however, subject to normal commercial confidentiality and data privacy, to work with experts in this area if they want to discuss it with us.

Chair

I think your answer is no, you don’t agree with the WHO, because if you did I think you would be behaving differently.

Adam Mitton

The WHO classification is a relatively recent classification. The WHO itself acknowledges that time spent on its own and money spent on its own does not necessarily mean addiction. We are working with our industry bodies and we will—

Chair

Let’s just be clear, by your numbers you have over 270,000 people who play the game for over six hours a day who could be considered to be addicted to it. Even though you track game play in terms of money spent and time spent on the game, you don’t necessarily think you have responsibility to intervene if someone is showing the signs of being affected by a gaming addiction.

Alex Dale

I will come back to my example and the other players we have spoken to. If someone is watching television for six hours day and has the Candy Crush app open at the same time, we should not assume anything about their intention. It is perfectly possible that they are enjoying both.

Chair

As I said earlier on, what you can do, using the data you have on each player, is monitor patterns of behaviour and your systems should be able to monitor someone who is suddenly massively increasing either their spend or their game play.

Alex Dale

Drawing on your point, we do look at how people are spending, as we discussed earlier, and when we see someone move into a high level of spend, it is not bang like that. It is measured over time and then it decreases at a lower rate over time. It is very unpredictable as to when it will start and there is lumps in it on the way, but it is not like suddenly, bang, we see a massive change in behaviour.

Chair

What we have discussed now—and I know other members want to come in—is that you are a data rich business, you gather data about each user. You use that data to enhance the game, which probably means people spending more time on it, people playing it more frequently and people spending more money on it. That is the nature of your business. What you don’t do is use that business to intervene if you think people are spending too much time or money on it.

Alex Dale

The question is what is too much.

Simon Hart

As a continuation on the same theme, we have heard quite a lot from you about the analytical element of what you do to understand your players more thoroughly. From the exchange you have just had, to me there does not seem to be very much inclination or evidence that you, on a broad scale, explore what the potential consequences are arising out of addiction. You do or you don’t accept the World Health Organization. What work do you do to try to measure the full extent of online addiction within your own particular products?

Alex Dale

I will just refer to the earlier comment. This is a very new judgment and we are happy to co-operate with independent academics to understand it more. We are not psychiatrists.

Simon Hart

To be honest, that is not really the point and we have made this point when we have had other witnesses. You are producing a product that consumers are obtaining in many millions. If you were producing a physical product you would not be sitting here saying that you will co-operate with academics at some stage in the future to ascertain what impact those products might have on your customers. You would have had to do that before you released them to the market. You would have had to satisfy yourselves and the market that there were no adverse effects even if it is a tiny proportion or the risk is minimal. You would have had to demonstrate there were no adverse effects for people using and enjoying your product. I don’t want to put words in your mouth, but what we are hearing now is not dissimilar to what we have heard from other witnesses, which is slightly a case of, “That isn’t our problem”. I would suggest that it is your problem. Even if the player profile that you have described to us is perhaps more adult than in some of the other instances we have looked at, there is still none the less the risk that some of your products have addictive qualities that could significantly impact on people’s lives. Yet it seems to me that you have made no attempt to research what the extent of that is or what to do about it if you were to uncover some difficulties.

Alex Dale

I will answer that in two stages. Adam can take the second stage. First of all, we have no indication from our meetings with players or the customer care contacts we have or any of the surveys that there is a problem in this area beyond two or three people contacting us every month. For the second point, and Adam can give more detail, we absolutely follow all the consumer protection requirements and the OFT guidance as well.

Adam Mitton

Exactly, and there is consumer protection legislation in the UK and the Europe and OFT principles that were put together with the industry five or six years ago about how to market games to people.

Simon Hart

Can I go back a step? As the Chairman said earlier on, to some extent you are relying on quite an informal way of ascertaining the extent of the issue here, including self-referral. What I am suggesting is that that may not be a particularly scientific or effective way of measuring the effect of this addiction or the addictive qualities of the products that you put the market. Surely you should be undertaking research that is a great deal more penetrative than simply relying on people to get in touch with you when they feel like it, which as we know with other addictions is a highly unlikely scenario anyway. Surely you should be undertaking the qualitative and quantitative research to such an extent that you can absolutely satisfy not only us but anybody who is coming into contact with them that there is absolutely no issue with the addictive qualities of your product and that, if there are, there are means of dealing with them. Yet that is not what you are saying you do.

Alex Dale

What we are saying is that we don’t have evidence that there is a material problem here.

Simon Hart

But you don’t have evidence that there isn’t either, because you have not bothered to do the work.

Alex Dale

As I said, we ask our players, “Are you enjoying the game? Do you feel that you get value for money from the game?” We have two or three contacts a month saying, “I think I’m doing this too much”, at which point we offer to block their account.

Simon Hart

We know with every other addiction that all of us have to contend with from time to time that that is a very unreliable means of measuring the extent of the problem.

Alex Dale

I am not sure it is fair to say all other addictions. Absolutely we should take the WHO seriously and we are reviewing our policies, particularly around proactive contact, which is the one we dropped a number of years ago. We are absolutely prepared to work with researchers on this topic. But expert opinion is divided and I am not sure we can see that if Grandma X is playing for six hours or 10 hours or 12 hours because she has the app open for that length that means she is addicted. We can’t see that.

Simon Hart

That is because you are not bothering to do the research that would demonstrate that one way or the other.

Adam Mitton

The WHO classification of gaming disorder is a relatively recent development and it acknowledges that time spent on its own and money spent on its own does not equal addiction. We do not have data about our players beyond that so it is hard for us to extrapolate from our data where there is or is not gaming disorder. It is a relatively recent classification. As Alex said, we are absolutely prepared to work with our industry bodies and experts to understand what can be done in this area more practically.

Simon Hart

The frustration to some extent is that when we have had similar discussions with other witnesses, as soon as we get on to this area of line of inquiry, which is quite an innocent one from the Committee’s point of view—it is actually trying to better inform the debate and about us trying to understand it. It is not about us trying to close down the gaming industry or us trying to embarrass you or Epic Games or anybody else. It is trying to acquire information that we can use usefully and, when it comes to legislating in this area, we can apply in evidence and principle fairly. Yet when we get on to this area there always appears to be a sort defensive mode that you get into.

Alex Dale

If that is your objective, my recommendation would be that either the Department or this group works with the industry body and independent experts to do a proper scientific analysis. That would be the right way to approach it, I think, rather than rely on any individual company’s set of players. Of course you would need to use their data to do that, but you are going to get different answers depending on which games and services you look at. I think pulling back, looking at the industry as a whole and working with multiple services is the right approach.

Simon Hart

You would co-operate with any investigation that—

Alex Dale

Subject to normal commercial protections, yes, absolutely, and data protection, obviously.

Simon Hart

Clearly and data protection, and make a financial contribution to that research?

Alex Dale

I would have thought so. We would have to find out how much it is first. I am not in a position to make a commitment like that, but within the bounds of normal market research, yes, I would have thought so.

Simon Hart

That is a preferable option to you undertaking this work yourself?

Alex Dale

We must be wary of false positives here. Presumably you would be looking to have a set of rules that works across all games companies or all service companies rather than just the match puzzle category. In order to do that, it necessarily needs to be industry level research.

Chair

My slight concern is that when you strip out the commercially sensitive data and information and data that you might otherwise say is covered under GDPR, there would not be a lot left for academics to analyse.

Alex Dale

Not necessarily. We have done research with academics before on what makes a good game level, so a good level in Candy Crush. We are not sharing personal information because it is about devices and how they play a particular level, but the research is still useful.

Ian C. Lucas

Mr Dale, you have been very helpful in providing a lot of figures and information. What we are concerned about here is unintended consequences. I am not suggesting for a moment that you are trying in any way to ensnare people or create addiction or anything of that nature and I have been impressed with much of the evidence that you have given. But our concern is that in other areas in which we have made investigations we have found that we are closing the stable door after the horse has bolted. You have 270 million users. Do you employ any psychologists in the business at all?

Alex Dale

No.

Ian C. Lucas

Don’t you think that, bearing in mind the success of your business, it is incumbent upon you to at least consider that there may be adverse consequences to excessive use of the game and to take advice internally on that by, for example, employing people to look into it?

Alex Dale

That would be marking our own homework. I think this is a very important area and the right way of dealing with it is at an industry level and we would be happy to co-operate.

Ian C. Lucas

I don’t think it would be marking your own homework. I think it would be taking your responsibilities as a very successful business seriously. The scale on which you operate is huge and you have been very successful, and that is a good thing. We are concerned because in other areas we have found difficulties in dealing with issues that were unintended but have had adverse consequences.

Alex Dale

What we are focused on in King—and we spend a lot of energy talking about and discussing this topic—is that we are a business but we are also a game. What does being a game imply? It implies that there are rules that are the same for everyone and that there is genuine player agency involved in what we are doing. There is lots and lots of debate about what we should or should not do in the game. The simple test we have, which we call the TV test, is are you happy going on to national TV in your country—and we are talking about Sweden, Spain, UK, Germany here—and explaining, “I’ve designed this level or designed this mechanic or designed this product and I am proud to say I have done that. This is a great product, a great level, a great design and it is fun and it is part of what normally would be a considered a game, that is that it is fair and there is genuine player agency involved.” That is what we spend our time talking about.

Ian C. Lucas

On the basis of the figures that you have just given us, over 9 million play your game for over three hours a day. It just seems that is a long time.

Alex Dale

I do that. People who commute—what is the average commute in the UK? It is probably over an hour each way plus Tube. When we look at when people are playing our game, in Europe there are peaks in the morning commute and peaks in the evening commute, so people are playing as they are commuting. In the US people drive to work so they are not playing Candy Crush. They tend to play last thing at night, quite often as a married couple, we understand.

Ian C. Lucas

Do you not agree with me that it would be helpful for you as a business, given the financial success that you have had and the scale of numbers of people who play the game, to have some independent research within your business about the possibility of addiction for use of your game?

Alex Dale

I totally agree with you. The key word there is it needs to be independent and we are happy to participate in that.

Ian C. Lucas

That could be internal.

Alex Dale

It is not independent if it is internal.

Ian C. Lucas

I think it could be. You could commission some academics. It may be seen as criticised but if you had come here today and you had told me that you had commissioned research into this issue, I would have been pleased to hear it. I am disappointed that you have not.

Alex Dale

Okay. I am saying very openly and publicly that we are happy to do this.

Clive Efford

How do you determine the market value of your games and how might that differ from how value is determined for a game that is paid for upfront?

Alex Dale

Essentially, every player will be deciding for themselves whether they wish to pay to progress faster or not. The minimum spend at the moment is $1.99 for 10 gold bars. Those 10 gold bars are used at the player’s discretion. They can then decide to spend more to get more gold bars to buy additional lives or boosters or whatever in the game or they decide that it is not value for money and they spend their money somewhere else. The level of spend they want to make in the game is very much player driven.

Clive Efford

In making that choice of do you make it a game that people pay a single fee for upfront or a game that people buy extra lives for, how do you assess that value? What comes into that?

Alex Dale

A bit of historical context is important here as well. Ten years ago games, including mobile games, were generally young—and by young I mean 25, 35—male, typical console audience. Then we had the growth in casual games, particularly on smart phones. That coincided with, or necessitated really, an ability to play in bite-sized chunks, that is you could play for three or four minutes at a time rather than need to learn complicated moves on a console and train in order to be able to play the game, which would mean you needed a lot of spare time. That is what led to the growth of the mobile casual games audience and it also led to the emergence essentially of female gamers in the mainstream, and that is really the market we have been going for. Part of that growth was free to play as opposed to premium. Premium is when you pay a fixed amount upfront although sometimes you have to pay extra as well in those games. Free to play is as I have described. Our philosophy has always been to make sure the game is truly free to play and that design principle is very important. The vast majority of people at the end of Candy Crush, so level 4,990-something now, have never paid. We were accused early on when we were growing of fixing the levels to try to convert people into becoming spenders. That is absolutely not true. The levels are the same for everyone. People who are paying are paying because they want to progress faster through the game and that is their choice.

Clive Efford

How do you make the choice yourselves between a game that is a premium upfront to pay one fee and then you have access to the game, albeit that in some games you can have additional payments as you go along, and a game where people have free access and then you try to monetise their engagement in that game? How do you assess how people make that choice and whether it is value for money for you to do that, to organise your game in that way?

Alex Dale

We do not offer premium games. We do only free to play and historically what you saw around 2012, 2013, 2014 is as free to play became the normal way of playing, the number of companies offering premium games and the number of games offered in that way went down and down and down. Players preferred free to play because you could try the game without any upfront payment. It is obviously appealing.

Clive Efford

Is there anything about players that you look for that makes them more likely to be someone to spend money with your games?

Alex Dale

It is very difficult to understand that in advance. We do not know and the conversion from being a free player to a paying player can happen at any point in a player’s lifetime with us. It could happen in the first week, after a year, after five years. It is very unpredictable.

Clive Efford

You have no way of recording data or any information you exchange perhaps with a platform like Facebook that would assist you in identifying those people who might be more likely to spend money with you?

Alex Dale

There are two different points there on customer acquisition. First, Facebook and Google and our other partners will have identified devices that they believe are highly engaged gamers, although it is not certain because it is statistically based,. Obviously those are people in the casual space that are of interest to us. Secondly, if a player stops playing Candy Crush and leaves King, maybe they would be interested in another game for King, so we might offer another game from our stable of games to them through advertising.

Clive Efford

That is advertising that you purchase through Facebook or some other platform?

Alex Dale

We buy the advertising ourselves, yes.

Clive Efford

How much does it cost to build and maintain Candy Crush? How much do you need to make from each player on average to make it a viable game?

Alex Dale

That is something that I think would be going too far in public data sharing. What I will say is that the value of a player varies enormously by country, by device, by time, and there is no single answer to that question.

Clive Efford

You would not be able to tell me what the average monthly revenue per player is?

Alex Dale

I have some data here that I can share, if you like. This has been “approved”. There are 270 million players a month; 2 billion net revenue to King, which is 62 cents per active monthly player on average; 60% Americas, 30% Europe, Middle East and Africa. That is the King published data. Moving to third-party market data, we have 3% to 5% of players spend monthly. That does not mean they spend every month but it is in any one month and they will spend on average $10 per month in any genre. That is a cross-genre statistic. What we have done for the Committee is looked at third-party market data for match puzzle, and this is the category that we operate in. It is games from companies like King, Peak, Playrix and similar. What we see there is that of all payers the annual mean is $170. That is slightly above the all genre number and the median is $21. The top 8% of payers, that is 8% of 3% to 5% of 270 million, represent two-thirds of the match puzzle spend in any one month. That is the core of the market for us and these are players who are spending $30 a week. That does not mean they are spending $30 every week but at any point in time that is the core segment of value to us.

Clive Efford

Just going back a bit, do you use Google advertising?

Alex Dale

Yes.

Chair

Does Google provide you with any information regarding players who are likely to play and spend money?

Alex Dale

Indirectly yes, but things have changed quite a lot since I was in charge of marketing. It is very rapidly evolving technology.

Clive Efford

Is that good or bad?

Alex Dale

Probably good. Anyway, my successor is doing a fine job. I think I will get back to you in detail on the types of service we get from Google.

Adam Mitton

I think I am right in saying that they do not hand over their data about their customers to us.

Clive Efford

No, they probably wouldn’t but they could identify for you people who it is worth your while targeting. Is that how it works?

Adam Mitton

We would buy advertising inventory and the partners typically try to target out segments.

Clive Efford

It is not just a trawling exercise, is it? It is more sophisticated, isn’t it? Doesn’t Google sell itself to you on the basis of its ability to get in the right market for you?

Alex Dale

Yes, it is very sophisticated, but I think the important thing to remember is that the vast majority of our players do not come through advertising. They see someone playing Candy Crush, they are interested in it and they go to the app store. We call these organic installs. They know that the game is reputable and King is reputable and they download the game. The vast majority of our installs are essentially free; they are organic.

Giles Watling

Gentlemen, you must be aware of the famous addictive spiral. I have a copy on the wall in my office, and it applies particularly to Twitter. You make a tweet and then X minutes later you feel the urge to check if anybody has responded and if they haven’t you go round again and you might retweet and then you go round in circles. Did this accountant/dating agency fellow who came up with the idea of Candy Crush build an addictive spiral into the game?

Alex Dale

Candy Crush is a compelling game. On the prompts you are talking about, we do send e-mails to players and we do send push notifications to players, but what we find is that it is the most highly engaged players who respond to the e-mails and the push notifications. It is not the e-mails and push notifications are making them more highly engaged. It is that the more highly engaged players respond to those e-mails and push notifications.

Giles Watling

Earlier you used the term “when people complete the game” and I understand that Candy Crush is not a game you can ever really complete. You have thousands of levels but you do complete sections of the game at a time. Is that designed in an addictive way, so you complete one and you go on to another?

Alex Dale

The game is designed to be fun. I go back to the principles I was talking about earlier. We are making a game with genuine player agency and consistent rules and we want it to be fun. We do not discriminate between different types of player in the actual game play. Fun is the mantra, make every day more fun, and that is what we believe.

Giles Watling

I thoroughly concur with that. When something is hugely fun, there must be a point where you want somebody to invest. You need to make a living and you are a business. Is there a built-in signal within the game that now you need to invest, you need to make—

Alex Dale

No, absolutely not.

Giles Watling

There isn’t? It is just something that people want to do because they get the urge to do it.

Alex Dale

We encourage people to spend but it is not a built-in signal in the game. We might put together a conversion offer and this will be seen by all players. It is a bundle of lives and boosters that we would like you to take. It might be $1.99 or something like that. We are a business so we need to do that.

Giles Watling

But you are not proactive, you are reactive on that. It is the player’s desire.

Alex Dale

Someone who has not spent before might receive that. The rule is that simple.

Giles Watling

Right, as simple as that. Can I ask you about these gold bars? I understand the Candy Crush store requires people to buy these gold bars. You are turning real world currency into gold bars. Is that an attempt to disguise what people are spending so that they are now removed from real world currency and don’t realise what they are spending on the game? Is that the idea?

Alex Dale

No. It is very common.

Giles Watling

Why do you do it then?

Alex Dale

We want people to get good value for the game, and I am now talking about our core payer base. We will put together offers for them and they will get some boosters and they will get some gold bars. If it was just cash, that would not be possible and that is good value for the payer.

Giles Watling

We have had evidence before us of one particular young man who managed to spend 32 hours in one sitting playing a particular game. Clearly that is not an ideal situation. Is there a way that you would limit play at all?

Alex Dale

I will go back to my earlier comment. We have a lot of people who are retired. We have the story of the person in convalescence, the person in hospital. It is for them to decide if they should be playing only two hours a day or six hours a day. We have not heard of people playing Candy Crush for the sort of length of time you are talking about.

Giles Watling

The other limit, of course, is would you consider building in a cash limit that people can spend on the game over a period of time? You mentioned that somebody spent the figure of $2,600 in one day. That is a serious amount of money. I do quite understand that it was a sale and they were buying for the future and all of that, but do you think there might be some sort of advantage for you, as you said, as an honest broker in the games world, to have a limit so that people can’t get carried away as they can with other forms of gambling?

Alex Dale

I will give some quotes from players. These are from the meetings we have with our high spenders and highly engaged players. This is from Sarah, who over, I imagine, the full seven-year period, has spent $7,000 in Candy Crush. She filled in a questionnaire and we sent her a Candy Crush-branded monopoly set to thank her for filling in the questionnaire. She says, “My husband and I play all three Candy Crush versions every day and we are very excited to be rewarded for our loyalty. We both spent a fair amount of money on boosters and were thrilled to have something fun we could physically play at home and share as a group. Thank you, King, fabulous idea.” Another one would be Lauren, at a US event we held earlier this year, with a lifetime spend of $3,500. She was a management consultant, an MBA graduate at an Ivy League university. She had an accident in 2013 with a traumatic brain injury and several neurological conditions. She could not work, she couldn’t follow a conversation and she could not watch TV. In 2016 a friend encouraged her to start playing Candy Crush and she noticed over time she was better able to concentrate, speech improving and the doctors noticed improvements. I quote, “I am absolutely convinced that playing Candy Crush helped me repair my brain injury and greatly improve my neurological conditions. My changes in functioning began shortly after I started playing Candy Crush and there really is no other explanation for why those improvements happened.” Both of those players spent a lot of money. It is incredibly rare that people will spend that level of money. It is a tiny fraction of a fraction of a fraction of our players but I do not think that anyone would deny that they have found great value and pleasure in that.

Giles Watling

That is great news and I think those glowing endorsements are splendid, but do you have any examples, for instance, where somebody has gone the other way and they have come to you and said, “Please help me, I am carried away by this”?

Alex Dale

Yes, we have two to three people a month contact our customer care to say this is too much, and we will help them turn off their account if that is what they want to do.

Giles Watling

Picking up on an earlier point made, if you are getting two or three people a month like that, do you not think you need some sort of psychiatric care for these people built in within the company to guide them out of whatever problems they seem to be entering?

Alex Dale

Customer care is very easy to find. It is a click on the screen. I can show you after the meeting if you like, a click on the home screen. When we look at the volume of contact and at the surveys and the meetings we have with our highly engaged players, we do not see this as an issue. That said, as I talked about earlier, we are happy to participate in industry research on this topic.

Adam Mitton

The two to three contacts that we have a month worldwide, when we get back in touch we respond to the contact that we have and we give them the range of, “We can block your account” or explain to them how to limit purchases on their device. The operating systems on mobile devices have the ability to turn off in-app purchases and things like that. I think one player in the last 18 months has asked for their account to be blocked in response to that.

Alex Dale

Eighteen months, in the UK.

Giles Watling

Would you accept that perhaps this might be the tip of an iceberg? As you earlier said, these are self-referred; there may be many more out there. Is there any reaching out you feel you should be doing to your players?

Alex Dale

It is not just self-referred, it is also the meetings we have with the players and the surveys we do. As I mentioned earlier, we used to send out a proactive e-mail and essentially we were told, “Please do not do this. I am fine; it is intrusive”. Given what the WHO has said, we will look at that again. Absolutely we need to take this issue seriously and we want to be a good company.

Chair

I would like to go back to something we discussed earlier on, going back to the job advertisement for the player insights manager. There were three parts to this. I think we covered two of it but not the third.

Alex Dale

You are very interested in this job.

Chair

I am not. What I am interested in is that you and your colleague have been slightly more open with us today than some of your competitors or other games companies have been, but this does give some insight into the way data is gathered and analysed within the company, and that is what interests me. I read this line earlier on where one of the roles within the job was to, “Incorporate behavioural data and research, leveraging our internal understanding of players and their gaming habits”. Incorporating behavioural data and research—you referred to that—could be external commercial data, the Nielsen data or other datasets and could be—

Alex Dale

And the quantitative data. Both.

Chair

Yes, understanding their gaming habits. You said that obviously the company gathers data about levels of spend and so on. The one thing I did not think we quite covered was, “Our internal understanding of players”. It says, “Our internal understanding of players and their gaming habits”. I am interested in what the internal understanding of players is based on.

Alex Dale

That would be things like you have this core audience of 35-plus female players. What are they like? Create a mental image. You will be familiar with this from your advertising experience. Create a mental image of who this person is, what sort of supermarket might they go to, where might they eat, do they have pets, children and so on. I think that is what that refers to.

Chair

It says, “Our internal understanding”. You can get that data about people of that demographic commercially and in all sorts of places. But “our internal understanding”, what is that internal understanding based on?

Alex Dale

This is when we are asking questions like that of our players. It is not bought in from MORI or whatever. We have brought the players in or we have done a phone survey or an e-mail survey and we are building a mental picture of who they are.

Chair

What sort of questions do you ask them?

Alex Dale

A range of things. We are interested in what sort of games they play, what sort of games besides King games do they play. For example, we learnt that people play other puzzle games, word games, that sort of thing, from asking our players what they like to do. We are interested in helping our game designers have a mental picture of who they are.

Chair

Do you know roughly how many questions people get asked as part of these questionnaires?

Alex Dale

No, I would need to get back to you on that.

Chair

A dozen, 20, 30?

Alex Dale

We try to make them short and sweet.

Chair

It could be any one of those numbers from that point of view.

Alex Dale

No, 30 is a lot.

Chair

More like 20?

Alex Dale

I would just be guessing.

Chair

If a person does an omnibus survey they could answer 30 or 40 questions.

Alex Dale

I can send you a copy of the last survey we did.

Chair

Could you? I would be very interested in that if you could. Presumably those answers are then linked back to individual players so you have a piece of research that says, “Here is someone’s gaming habits and here is some insights about them as a person”.

Alex Dale

Yes, we can do that and we would do that for gaming habits in particular. I think for the other stuff, like do they have a cat or not or what pets do they have, that would be largely irrelevant. The cat stuff is really for understanding what sort of picture we are painting of a customer.

Chair

How is that valuable?

Alex Dale

It is to help the designers imagine who they are building a game for.

Chair

It is interesting because you said you do not employ psychologists but this to me sounds like psychology, which is not just designing a game around gaming habits but designing a game around what makes someone tick.

Alex Dale

It is common marketing. You want to know your customer, and we do that.

Chair

Yes, but there is a behavioural science behind this as well as just understanding game play. This is about trying to get a deeper understanding of a type of player, which you must analyse against their game play, otherwise it would be largely pointless.

Alex Dale

We want to understand game design. We are game designers. Obviously not all of us, a relatively small number of people in King are game designers. That is the art or one of the arts in what we do. We want to provide as much colour and insight into our audience as is reasonable for a game designer. But a game designer will not be saying, “It is female 35-plus, therefore I must do A, B and C”. They will be using that information as part of a palette in order to create their game design. It is not linear.

Chair

No, but presumably that data is only worth something if you can say, “We understand this is our core demographic of who plays the game and what sort of people are they and what are they interested in”. You might have growth sectors and say, “We would like to understand why 25 to 30 year-old men do not play the game more”. I do not know what your growth sectors would be. Presumably this data is only really valuable if you can link the insights you are gaining through surveys, the psychological insights, into people’s game play as well. Then you can say, “Here is a typical person and this is what their world is like, this is what they are like and this is how they play the game”.

Adam Mitton

If you mean doing that on an individualised basis, I do not think we do that. I am pretty sure we have not done that. Alex used to run the department but I remember from not long after I joined the company one of the questions we asked the audience was do they consider themselves as gamers. A lot of the people in our demographic, 35 year-old women, would say they were not gamers, yet they were playing Candy Crush. It said to us that we need to make sure that we have a game that is easy for them to understand and pick up quickly and maybe we need to have tutorials in there. It is that kind of level of insight that we can get from the surveys and that might be an example of how we use it.

Chair

Are the surveys anonymised?

Alex Dale

I am not sure I know.

Adam Mitton

We will get back to you on that. I imagine they are. We follow normal data protection requirements so I will get back to you on that.

Chair

If you have a known player—you approach someone who is one of your customers—you have their games play data and you ask them to fill out a survey about themselves, if it is not anonymised I would imagine the purpose of that is so you can link the psychological profile of someone to their games playing.

Adam Mitton

As Alex said, we do not change the game design on an individualised basis.

Chair

No, but you are gaining insights to inform game design in the future. You are not creating a unique game for each individual player but you are using, I would say, psychological insights as well as insights on game play to improve the design of the game for everybody.

Alex Dale

Going back to your earlier question, we do not employ behavioural psychologists or similar. We have game designers, who are a unique bunch with very varied backgrounds, whose first concern is about games and they are driven by a passion for games. They will be looking for insight and clues as to, “Is this concept I am working on going to be fun for a particular audience or not?”

Chair

Sorry to interrupt. To me the distinction you make there sounds like someone saying, “I do not employ a chef in my home but I do employ someone who cooks my food”. You are saying, “I do not employ a psychologist but I do have loads of people who are gaining psychological analysis and insights into our players, which they use to help inform games design”.

Alex Dale

When you are playing a game, whether it is Scrabble or Monopoly or Candy Crush or something else, that game has been designed by someone. There will be feedback incorporated into it. I was reading about Game of Life, which is a traditional game. It has been through multiple iterations and the game will have evolved as people’s appreciation and understanding of different elements of the game have changed over time. That is a quite normal product development/marketing thing to do. Every business would want to do that.

Chair

Thank you for agreeing to send a sample survey to us. Obviously we do not expect that to include any individual’s personal information, just the questions that someone was asked. One of the reasons we are interested in that, in the work that we have done looking at the work of companies like Cambridge Analytica and Facebook, is that these sorts of surveys added to data are used to gain psychological insights about people. Obviously that has been done to a games design but that is something we take an interest in because from our point of view—as a company within the rules that exist you can do as you please—we want to make sure that if people’s data is being used in this way it is something they have given their informed consent to.

Alex Dale

Going back to how we design the games, I mentioned mothers’ day. The next thing we would do is what we call a play test, where we take a game—it could be at a relatively early stage—and allow people to play it through something called Test Cloud. There they know exactly what they are doing. It tends to be biased towards expert gamers and we will see how they are playing this game. Then we might do more of a quantitative play test where the game is released to tens of thousands of people in a particular market and we will look at how that game is being played. All that quantitative feedback will be incorporated into the design of the game.

Chair

A final one on this. How many of these surveys do you think you do in a year?

Alex Dale

I don’t know. We will get back to you on that. Not many. I do not know if it is one a month or more than that. I don’t know.

Chair

One survey. How many people?

Alex Dale

I have not been involved in the marketing for two years, so I will get back to you with a proper answer.

Chair

The number of respondents is the thing rather than the number of surveys.

Alex Dale

Okay, you want the number of respondents. Fine, we will see if we have that.

Chair

Thank you.

Ian C. Lucas

Do you know what proportion of Candy Crush players log in via Facebook and play directly on Facebook?

Adam Mitton

I cannot give you the numbers. Let me say, though, that a significant majority of our players do not create any kind of registration with King to play the games at all. You can download the game from the app store and just start playing. We create, effectively, an anonymous account for you so that we can make sure that the sessions work and that if you buy something through the platform that you are playing on we can deliver that purchase to you and so on, but you do not have to give us a username, an e-mail address or anything like that.

Ian C. Lucas

What proportion of your users do give an e-mail address? You need not give me numbers; give me a rough percentage.

Adam Mitton

A significant majority do not.

Ian C. Lucas

Do not. Do you know if that is 30%, 10%?

Adam Mitton

I cannot really share that because of the sensitivities around public information. I am sorry about that.

Ian C. Lucas

Okay. Can you give me a percentage of what proportion of your Candy Crush players play directly on Facebook?

Adam Mitton

I do not know that. Do you know that, Alex?

Alex Dale

We can get back in confidence. I do not know that.

Ian C. Lucas

Is that commercially sensitive information?

Adam Mitton

Yes, I think so.

Ian C. Lucas

Do you share data with Facebook?

Adam Mitton

If somebody plays Candy Crush on the Facebook website, from that platform, Facebook will get some information about the fact that they are playing the game. It is its site.

Ian C. Lucas

It will know from the site that somebody is playing Candy Crush through that?

Adam Mitton

On its site it will know that its customer—it is its website—has played Candy Crush. It will be able to see that. On that platform the payments run through Facebook so it will know whether its players have spent or not.

Ian C. Lucas

Do you share any data with Facebook from Candy Crush players who are not Facebook users or who just play Candy Crush through you? Do you pass on any of that information?

Adam Mitton

Only if someone playing on a mobile device creates an account on Candy Crush, or whichever game it is, using Facebook’s log-on API. It is a facility that we offer to enable a player to create an account. The reasons that they might do that is so that they can track their progress across multiple devices, typically, or they can restore progress if they get a new device or something like that. Some players do it also so that they can see that their friends who are Facebook connected are also playing Candy Crush. They can see them on the saga map in the in the game so they can see each other’s relative progress in that sense. In that scenario Facebook would get some information about who has logged in and when but not much else.

Ian C. Lucas

If someone has no connection with Facebook whatsoever on their account and they play Candy Crush, do you share the fact that they play Candy Crush with Facebook?

Adam Mitton

No, except that we have some ads in the game and if Facebook is the ad network that delivers the ad, it will see that—the inventory is available in Candy Crush as the inventory source.

Ian C. Lucas

Facebook would know that a Candy Crush player was playing Candy Crush even though an individual had no relationship with Facebook?

Adam Mitton

I think I am right in saying that Facebook only serves ads to people who are also its customers elsewhere, so, no, it would not. It would be someone who has a relationship with Facebook. But the information that it gets in the advertising scenario is what the inventory is. It is somebody playing on a mobile device playing Candy Crush and the dimensions of the ad inventory that is available.

Ian C. Lucas

You mentioned GDPR earlier on and how you differentiated as a result of GDPR. Has your relationship with Facebook changed because of the introduction of GDPR?

Adam Mitton

Yes. Facebook changed the number of the ways that it shared data with us and with everybody else as a result of people using their Facebook login. It changed its APIs and, as a result, our relationship has changed.

Ian C. Lucas

That was a result of GDPR?

Adam Mitton

GDPR and Cambridge Analytica happened at the same time. We would have to understand what happened.

Ian C. Lucas

They happened quite close together.

Adam Mitton

You would have to ask it what prompted what.

Ian C. Lucas

Before GDPR and before Cambridge Analytica you had a relationship with Facebook. At that stage, was Facebook able to access business through Candy Crush?

Adam Mitton

What do you mean by access business?

Ian C. Lucas

Access individual users and identify individual users.

Adam Mitton

If a user connected to their account using the Facebook login, a function that we have in the game, yes, Facebook would know who that player was and when they were playing, because that is what the login allows it to do.

Ian C. Lucas

What about if they did not access it through Facebook, they accessed it just through you?

Adam Mitton

No.

Ian C. Lucas

You did not have a data-sharing agreement with—are you a Facebook developer?

Adam Mitton

We have an app on its platform, yes.

Chair

So that I am clear, Facebook customers who log in to Candy Crush via the Facebook login, Facebook will know who they are, it will know when they play and it will know when they make a purchase?

Adam Mitton

Not if it is on a mobile device. On its website Facebook will know when they have made a purchase, because it processes the payments on its site.

Chair

Yes, I understand that. In its developer agreement for games, it could often know what level within a game someone has reached as well. Would that be data that King shares with Facebook? From the Facebook developer portal, that is listed as one of the sources of data that Facebook would get.

Adam Mitton

I am pretty sure that it does not get that from the mobile integration. It may see that on the website, I am not sure.

Chair

Also, gathering this data means that if I was an advertiser and I wanted to target people who are Candy Crush customers who frequently spend money on Candy Crush, Facebook could give me—Facebook would not give me the data, Facebook would allow me to market to people like that.

Adam Mitton

For people who are on its platform, yes, it would know that.

Chair

I want to be clear on the data you get back from Facebook. Could you tell us what your understanding is of the data you get from Facebook, from people who use the login?

Adam Mitton

I have a list of people who use the login. We get a username from them. We get an e-mail address from the user but that is user-configurable. They can disable that if they want to. We get an access token. We do not get their Facebook password, we get an authentication token from them. In the player’s option we can see their friends list but only those players who are also playing Candy Crush and have logged in and agreed to share their friends list. Both players have opted in, basically, to enable us to see, and placed each other on their friends map. We see that.

Chair

If they are both Candy Crush customers but they happen to be friends on Facebook, they will see each other on Candy Crush?

Adam Mitton

They both have to have agreed to share their friends list with us in their Facebook settings. We get the profile picture and we get country and language, so we know what language to serve them in. In some instances we get gender and we get some information on age range.

Chair

Do you get information about pages they have liked on Facebook?

Adam Mitton

We do not collect that, no.

Chair

According to the Facebook developer portal, that is data that is often given.

Adam Mitton

We do not collect that.

Chair

You do not collect that. Do you collect any data about videos that they have posted or been tagged in?

Adam Mitton

No.

Chair

What about places they have been to?

Adam Mitton

No.

Chair

Their hometown where they live, location data?

Adam Mitton

We get country-level information. I do not think we get hometown information.

Chair

Thank you. That concludes our questions. There are some things we want to follow up on. I want to say on behalf of the Committee we appreciate you coming prepared to answer in the level of detail you have some of the questions we have put to you. Thank you very much. Examination of witnesses

Witnesses

Chester King, Ian Rice, Dr Jo Twist OBE and Dr Richard Wilson OBE.

Chair

Good afternoon. Thank you to members of our second panel for joining us today. I will start off by asking my first question to Dr Jo Twist and Dr Richard Wilson. As you know, we have heard throughout the inquiry from individuals coming forward and by organisations like Game Quitters about problems some people have with gaming addiction and how they try to manage that. You will have heard as well our discussions with some games companies about the World Health Organization’s categorisation of gaming disorder as a medical condition. As you are the leaders of two trade bodies that are relevant to this sector, I am interested to know what your position is on that, both on the claim made by many games players that they feel they suffer from addiction and also the view of the World Health Organization.

Dr Twist

Thank you so much for inviting us to this session today. It is important that the WHO inclusion of gaming disorder in ICD-11 is about—ICD-11 is a statistical manual and it is a very critical tool in order to understand, attract, codify, compare and share data on public health globally in a consistent manner. It is important that we understand that because the WHO itself says, “Inclusion or exclusion is not a judgment on the validity of a condition or the efficacy of treatment”. We do agree with a large part of the scientific and expert community who do not necessarily think that the WHO is basing this decision of inclusion on robust enough evidence and they think that this is premature. Saying that, we are aware that people play to excess and what we want to do as an industry is to help find balance. We actively and will be redoubling our efforts proactively to make sure that people are empowered, that they understand that they can find balance, like the millions of players who do find balance. We also want to empower caregivers so they can protect their children, set parental controls and screen-time limits. We believe that there is a more nuanced understanding of different kinds of games and genres of games, different kinds of reasons why people choose to spend time in games, whether that be because, “I like a puzzle game to relax” or whether that be because, “I like a good story and good characters and good narrative”. We are in agreement and we would like to see more robust research done. This is why we will continue to engage and be in dialogue with the academic community and with health carers.

Chair

Thank you. I have one or two further questions but perhaps Dr Richard Wilson would like to respond as well.

Dr Wilson

Yes, thank you. Thank you for inviting me here today as well. I believe that anything can be done to excess. That is the first point. The second point is we have seen from evidence to this Committee that there clearly are some people who play games excessively. I was particularly struck by the two individuals who turned up who said they contacted Game Quitters. They were clearly playing games to excess and were addicted. I am also struck by the WHO’s definition of gaming addiction, which is very tightly drawn. I think we have to recognise the WHO’s decision on gaming disorder. It is important to recognise it because it has been drawn in a very reasonable and conservative way. I think we need to take steps as an industry, and obviously working with Government, to make sure we can minimise the potential for gaming disorder. I remember when we had our informal evidence-gathering session at the beginning of February, end of January. One of the things I said was I hoped it would identify good practice. One thing that has come out is that it is quite clear that we need to do more to minimise potential gaming disorder. One of the positive things that has come out is that most of the evidence we have seen seems to indicate that small proportions of people could be suffering from gaming disorder. That is not to minimise the problem but at least that is a positive thing to reflect on.

Chair

One of the things that has struck me during this inquiry, which has looked at both social media as well as gaming, not specifically gaming, is that they are both data-rich businesses. When video games started, a game was made, people bought it and they played it. The person who made the game did not have any data about how they played the game, necessarily, or certainly not live data. They just knew they had bought it and hoped they would buy another one. Now games are being designed around the data that is gathered about play. On that basis, what responsibility do you think companies have to proactively identify people who could be in the harmful category, who could be people whose behaviour patterns are likely to indicate that they could be suffering from some form of gaming disorder?

Dr Twist

I think the diagnosis is not there. The tools to understand how to diagnose what is excessive screen time in general, what is excessive technology and that kind of thing, need more research. What we would rather do is empower people, give people the information, give people the tools, help people have the discussion with each other and players and understand how to balance their lives. It is also fair to point out that there are many, many different kinds of games and hundreds of different genres. Not all games are the same. Not all games rely on live data. There are wonderful BAFTA-winning games that are 90-minute experiences that are pure story-driven narrative experiences that you play single player or you play on your own or you play without the use of the internet. Not all games relay on data but in a modern 21st-century society we are entering into a data-driven, algorithmic-driven world, and that is an internet-wide, 21st century, society-wide issue. We have to have a sense of digital literacy among young people in particular in order to understand how to protect their privacy, how to protect themselves and how to navigate this new world.

Chair

Yes, but the difference between now and the future, the difference between these data-based games and games that are not data-based, online with play data, in games where they make their money out of in-game purchases or advertising based on the amount people play, is a very different mentality for the designer. The designer’s objective is to get as many people to spend as long on that game as possible in order to maximise their return. Companies like King and Epic and EA are not charities. They make their money through people spending money on their games, EA obviously through purchase as well. That is the difference. One of the questions we have to consider is that change in business practice based on technology and the application of technology, which for many people will deliver something in their lives that is pleasurable but in some ways does create an incentive for people to spend as much of their time and money on the games as possible and the games to be designed with that outcome in mind. That is a big difference between the gaming industry now, probably, and the gaming industry 20 years ago.

Dr Twist

As an important part of the creative economy and as any kind of content business, which games are, you want to provide experiences just like a film franchise or a soap wants people to stay loyal to that soap or to come back and watch “Avengers” or that new Marvel film. We need to recognise that this is a challenge for companies. There are over 2,000 games businesses up and down the country providing employment, but not a lot of them are big companies and well resourced. We think this is really important to take into consideration as well. That is also part of the decisions that they make around what kind of game and on what platform they want to make a game and who they want to make a game for.

Dr Wilson

Can I chip in there? At TIGA we have slightly different figures compared to Ukie on the size of the UK games industry. What I could echo is that certainly looking at some data that we have examined over the last few months, looking at games that were released in 2018—there were 150 games that were released by UK game developers in 2018—we found that almost two-thirds of those were premium games. I am not suggesting a third were premium but it was quite interesting. Those two-thirds of games are going to be typically adventures that have clear start points and end points. They are not going to be designed for prolonged or extended use. It is worth pointing that out. Also to echo the point about the size of the industry—and I know we might come to this later—and the structure of the industry, it is worth pointing out that something like 66% of all the UK games businesses that we have analysed have four or fewer members of staff, quite small businesses. On the responsibility of games companies to address harms, the Government have issued an Online Harms White Paper and that will extend a duty of care to businesses, so this responsibility and the duty will be coming in any case.

Chair

We will come on to the White Paper in a moment. On the question about research and the lack of research, one of the problems with that is the data is not available. This Committee cannot do a study based on the patterns of play of someone who might play Candy Crush for that top end, that top point of whatever per cent, and how that journey occurs, how much money they spend. We cannot look at that and take a view because that data is not publicly available. The same goes for the rest of these companies. It is difficult for the industry to say there needs to be better research when the industry is not commissioning that research itself and it is not making that data available for others to do it. As representatives of the industry, what is your view on that?