Tackling Audience Experiences in Games

[Music] so thank you so much for being here we came from austria as you might hear eventually and we're super excited to bring our knowledge to you guys today we want to talk about tackling audience experiences and games so not looking at games from the typical player's point of view but there are also other people who might be interested in the game but first things first let me quickly introduce ourselves so my name is johanna bieker and my name is rainerman i'm a researcher and i'm a lecturer at the grad university of technology in austria so i'm coming from the academic side and i'm ceo and co-founder of the austrian-based indie studio rear byte so my background i have a phd computer science i love doing research i love doing data analysis like machine learning i love looking at the players point of view through this data analysis i also do ai

research so everything very academic versus very researchy well i'm not a doctor but i've been working a lot in games and our company has published about 25 titles on all the major platforms so we're into creating developing and publishing games um all i can publish is papers as an academic which is also lovely because i can gather knowledge and bring it to the people by publishing those papers and to publish those papers i obviously need data but again when i do not develop games or publish games i don't have often access to this data so yeah luckily our games produce a lot of data that can be used for analyzing them so we found that being a researcher at the research institution and working together with a indie game studio is a very very good benefit for both of us because i can use this data to help their studio to improve the

games on the other hand i can also use this data to publish interesting contributions for the community so we found this is something which we have in common but there's also something else we do have in common we both love streaming so i love streaming games i'm a streamer myself i'm not a good streamer or a popular streamer but i enjoy it myself and what i found most interesting actually was looking at the chat while streaming because looking at the chat there's something going on which is so different compared to what we have when designing games so there's the entire population of spectators and we wanted to find out more about those spectators through data analysis yeah well i also gained a lot of experience in streaming it's more about streaming development and play testing of our games so um yeah today we want to talk about this

spectator who is a spectator and how can we design to engage spectators and last but not least we want to give away some evaluation strategies that we use together in this field so we will give this talk a little bit from the researchers perspective and then also from the designer developers perspective um so when we think about our spec data we need to first find the proper definition for spectator as big data can be found in so many different and several places a spectator can be someone watching a stream without any action just being there and watching a streamer but it can be also someone who is really involved and really engaged constantly interacting with the streamer and with other spectators through the chat for example asking questions making comments it can be also someone who is just sitting at home maybe a parent watching over the shoulder of the child playing a video game maybe being invested maybe not at all

but it could be also a fan in a stadium filled with thousands of fans enjoying together um a starcraft play so spectators can be found in so many several places and we also need to consider the different types of spectators with different backgrounds for this talk we want to use a definition of spectators but spectators are basically people who follow the gameplay in real time without being the main player in them of the game eventually they have impact on the game to some extent which we'll hear here later but it's not the main focus yeah but first let us take a step back and see what is important for the player so the player wants to have a good playing and streaming experience and a player wants interaction possibilities with the viewers that are there in the chat or viewing on site for example it's all about entertaining those viewers and retaining those viewers and in the end the goal is to build a

community around the game the player and the stream for example when you design a video game for spectator however you need to think of different aspects because as soon as a player would enter a stream the spectator must immediately know what's going on understand the gameplay mechanics and so on you should think of different interaction possibilities with either the player or with the with the other viewers they want to be entertained surprised so they have different motives for coming to the stream or they just want to be part of the community as a social aspect we also need to know that fans and spectators is something different a fan can become to have a speculator can become to a fan but they don't start off as a fan immediately so fans are usually motivated by a lot of achievements however data can be motivated by so many different things without the necessity that they are fans they can just enjoy the aesthetics of the game this

drama or the skill of players there are different strategies which we which are definitely interesting for us as game designers how to make viewers to fans but that's probably a different talk today we want to focus on the motives for the spectators and what's interesting about that so spectators are motivated by different aspects of a streaming or game watching experience so some are eventually more interested in the game itself some are eventually interested in the streamer or player or team esports team as a person or as a team and others are interested in the social experiences and want to interact with other spectators or the community so in literature we can find different motives for spectating interestingly enough there's not a lot of research yet on streaming or in

esports behavior it's growing and becoming more important but the most research we found and which is very similar to what we are interested in can be found in sports events so going to a football game to a soccer game for example and understanding the spectator's motives there can help us also understand the motives for people watching a stream or going to an esports match so the some of the motives can be for example their identification that the spectator really wants to be associated with a successful player and also wants to celebrate the success their statics this is really really important they really enjoy the beauty of the game we have drama and excitement this is this idea that you experience this thrill maybe stress but a very pleasant form of stress while watching a game escape escape would be that you escape

from this daily routine and you come home from work and you just turn on a scream a stream and watch the stream just to getting relaxed again knowledge eventually you're a gamer yourself and you want to learn more about the game you want to experience through the skills of the player and you want to learn the social interactions especially on twitch this is so important that you're able to communicate with other spectators this evolution to some sort of a community to some sort of a group and then the team or the player support that you're part of a community who is supporting a player similar to player types which all of you might know there are also different types of spectators this is one form simple model we built where you have to plan the player and the game as main interest points for spectators could be also other spectators and some people are really want to interact some however want to

consume so those for example who want to interact with other players or with the players are socializers those who want to consume and learn from the skills from a player for example can be learners those who want to be part of the game make some sort of impact on the game can be participators and those who are consuming the games they're just enjoying the beauty and based on that in literature we find nine different personas and what we all know when we create games it's always important to think about those different personas who we are designing the games for same can be done for spectators so there's for example the bystander the bystander is someone who is still very not very interested in the game yet who just happens to be there this one is either uninformed about the game or uninvested and not really engaged yet priorities can change

if there's something happening which is engaging for example some drama or some excitement the this one can become to a curious one the curious is suddenly interest in the game and then wants to learn more about the game it starts with a small fascination and they eventually start watching and watching more and this one eventually becomes to the inspired the inspired feels like oh that's a cool game i also want to play this game this one would eventually after the stream go to steam and buy the game him or herself this is already a very very nice process when he got to inspired and as soon as you already know about the game you eventually become the pupil a pupil this one wants to learn this one wants to understand specific techniques tricks of a player or wants to learn about specific maps within the game he's or she is very interested in

learning details so this one would ask question would i would interact with the player or would interact with other spectators and then there's the unsatisfied aka the grumpy one who'd rather like to be playing him or herself which comes from a desire an example for that would be a dad playing with his son and the son wants to play and the dad has to watch but actually wants to play himself next there's the entertained and the entertained would rather prefer watching over playing so there's this difference that that content is being watched for entertainment basically like moving then we have the assistant the assistant wants to sort of impact the game progress they want to help they want to be some sort of advisor the this one would be for example the older brother brother sitting next to me and trying to help me through the levels but

this could be also someone who wants to tell me about things which i don't want to know because i want to explore the game myself then there's the commentator this is a very interesting one this one is a spectator and a performer or entertainer at the same time this one wants to entertain other spectators this one would be constantly talking about the game give him comments in the in the chat in a twitch that also constantly talking about the game um this can be some sort of a camera and telling a different perspective and last the last one is a part of the crowd this one wants to be part of the community if something insane is happening and the crowd would go wild or in a stream the entire chat will will be flooded this is what the part of the crowd would enjoy yeah so now that we've learned about these different types of personas of

perspective spectators how can we design to engage the spectators well first of all our main goal as a game designer would be to create a playable game but speaking of spectators we also want to create a watchable game so that means viewers need to be able to understand what's going on in the game and they need to be able to keep track of everything that is going on in the game at any time and also we want to retain those viewers that means we want them to come back to our stream or to our play session and they should keep watching what we identified were those three phases of spectator engagement first we want to attract people to your game or to a stream and we want to teach them about the game so what they get what is the game about how is the game played what are the rules of the game next we want to retain those viewers so that they come back and maybe tell others and watch again

keep watching and third we want community involvement and engagement so that means we want people to interact with streamers and with each other and this is all to support the spectators motives that we have heard before so uh for example we can attract with the aesthetics of a game or with drama or just with basic simplicity um so that people are um yeah they know how the game works and we can use drama for retaining viewers for example and we can try to build up knowledge so that people come back learn something more and then we have social interaction for involvement and engagement we can um encourage people um to teach others about the game um and we can support this affiliation that viewers will be supporting a specific team for example or a special

player or stream so what happens when a viewer walks into a stream first question that pops up is what is happening here and as i said before the optimum thing would be that the game is just very easy to pick up and from the very first second you actually know what uh what's going on but we want to have viewers learn about the game while watching the game or the stream so our goal is first to design to attract viewers um and that is that that should be seen in all the different kind of areas of the development of the game coming from game design then graphics ui all those features they should give spectators visual hints and experiences about all the actions that are currently happening in the game people should know about the tactics and even emotions the players are going through a good example for for that is in league

of legends there's a color scheme indicating what is happening at the time so you just jump into stream and then you see a certain color and you know okay this is the current phase or this is maybe what's going on at the moment um and people or viewers should understand the current situation of the game for example if the team is winning or losing next we want to design to teach viewers um this is an example we did in one of our games at rarebyte we we have just there's a difference between ui done for players versus ui for spectators so for a player it's maybe important to really focus on the player's character and see what's going on with the character and around the character and give a lot of feedback there but for spectators maybe they want an overview of all the players in multiplayer and you want to be able to compare players skills or players stats

one element which is very exciting here is one method called information asymmetry here the idea is that the player and the spectator have different pieces of information very obvious one as a streamer or as a player i do know for example what strategy i'm planning at the moment this is just in my head so obviously this is something i as a player know but the spectator doesn't i as a streamer can now decide whether or not it makes my stream more interesting to tell the play the spectators this piece of information or to keep it for myself to keep the speculators full of tension and excitement what's happening next so keep this element of surprise as a game designer i can also add elements to make this happen for example what we just heard to add specific ui elements or give specific overviews which are just in seen by spectator so for example i

can see an entire map of all the different players competing to each other and i know what's going on but the players do not so this is also helping me to get excited or i know as a spectator there's a specific trap over there and the player is just running over there and i as a spectator get very excited about that so why do viewers stay or come back to a game or stream well first there's just this fierce competition that you know from sports people like to see competition and see teams or players competing and and basically playing against each other and then there is drama of course you never know who's gonna win this game is it this player is it this team is it another one and then there is this affiliation that we want to support that viewers get engaged and will support their favorite streamer performer player or team what we want to do is we want to design for drama so we want to create

an intense playthrough and intense matches to support that drama and if you think of esports that's basically what is happening there is um that players that are very similar in skill play against each other so there is a natural drama coming from that but if you have games that are played by just random players this is different and maybe you have to think about integrating mechanics to rebalance different player skills there's this infamous example of rubber banding in mario kart where you can basically if you fall behind you can catch up because you get just get faster we even use that in a hoverboard racing game you just had to like move on to a trail and then you were able to catch up and there was just an intense race coming from that with completely different kind of players or skilled

players also we want to design to create the unexpected so a viewer comes to stream and should never know exactly what will happen next and that's one of the reasons this viewer will probably stay to find out we can add random events or story twists to our games to change this kind of balance within the game we can add surprise boxes uh crates or loot to do that for example this here is from we are screwed the game we're currently working on um we have procedural content generation and um random events in that game and this is a space code game where people take on different roles and we use this kind of those random elements to force people to change roles all the time something which is also very interesting is how to make it happen that viewers can be part of the experience

so there could be for example a passive passive mechanic to make streamers give streamers the option to involve the spectators in the in the game so for example what what i use from in darkest dungeon there's this option that you have those different characters which you can play and i would always look at the stream and look who's most active and then i would name the in-game characters based on my spectators based on my viewers so this is they're basically getting involved in a very passive way but i as a player i have the mechanic to involve them um on the other hand um being being as a spectator active in a game experience can be also very engaging this would be this would mean that um spectators can vote for specific um things happening in the game or now it becomes increasingly interesting to also make mechanics that you can for example cheer for specific players within a game and then they get

a specific boost in the game as a game mechanic or a specific item in the end it's about building a community around the game and making it accessible to all those viewers all those different kind of spectator personas the audience should feel welcome at all times while players should still feel safe and the goal is to let viewers be a part of something next we want to go or quickly go through some strategies we used on our projects so first so there are so many evolution strategies i mean i'm coming from ux background there are so many ways how you can think of evolating the player experience and all those can be also used basically to evaluate the spectators experience those can be things like observations think out loud tests you are working with data in-game data in-game metrics interviews you name it but we want to

introduce two specific models which we used in our collaboration basically and the first one is that in this point of view yeah let's start with the indie point of view and of course this is called the party test and it looks exactly like in this picture here um this works in a way and this is very simple um we would just invite 15 to 20 people to our office and do something called a play party and we would just um start a game and then have some people play this game and other people's watch those people play and then we evaluate and uh we record that and we collect data basically and we can identify issues like with onboarding with usability and playability and this is both for player and spectator engagement and experience and my point the academics point of view i love again i love working with later

and i do a lot of big data analysis so i really try to gather a lot of data either qualitative or quantitative data but what's very interesting is for example to build different models or in based on the interactions of spectators with the game or with a spree stream with the player or with other spectators this can be usually can be data locked in the game can be locked in the stream can be twitched it or can be even twitter data which is just related to the game i wanted to introduce you two models which which we use are two processes we often use the first one is graph based analysis here we built based on the interaction of the spectators with each other with the player we built a graph this means if as soon as two spectators would interact in a chat at a very similar time we assume this might be some sort of communication

so we would make a connection between two nodes within this graph which means okay they're connected they have some sort of relationship in within this graph and we can make sense of the out of this data so looking at a lot of data this could be a graph this could how such a graph look like and what we can see there are some nodes with a lot of connections those nodes could be very relevant and important points to look at those can be for example those personas of commentators who want to entertain the others and are really important as an engagement starting point so can we use this information to give those a specific reward for example to make sure to keep them engaged and keep them in our stream or part of our game experience and then we can also identify the weaker points those with not a lot of connections and we can think of ways how to integrate those and

help them to connect to other spectators on the other hand this is one of the more traditional ways to do data analysis this is basically where you look at the number of interactions over time so this is when you look at the twitch chat for example you can see in minute minute 40 there was something going on something very important was going on and we can find out by looking at the in-game data here was a specific boss fight apparently also the streamer did something very well can we use this information to make the engagement even better for the next stream um so just to summarize we can also say that those different relation strategies can actually work together um again to so that i can do my data analysis i need a lot of data the point is that very very often when we work with studios it's becoming very expensive and it's

becoming very complicated because the data they the studio locked in the game just was not right was not locked in a correct way or just not locked at all yeah and the idea is here just to work together from the very beginning of the project and do a data definition together and then we as a studio would take this definition and build it right into the very foundation of the game and we then plan and prototype our game and when a build is ready of the game would start a play testing session and then we would be able to collect data that is then given back to the research institution and since the data already fits our what we sort of need because of this definition phase for us it's so much easier and much more cost effective to make sense out of the data and help the studio even it's um even if it's a small studio or a big studio to help to improve the game all right

um a short recap of what we talked about today um keep in mind not every spectator has the same motives there are different motives for viewing a stream or a game then uh also keep in mind that you need to design for different spectator personas and also as we just said sees the opportunity of collaborations between your indie studio or your game studio and research institutions all right um that's it thank you very much if you have any questions thank you so much [Applause] i think we have time for one or two questions afterwards we will be in the report area hi i was wondering if uh to combat the problem you're having with data sources like that being locked down by different studios are there any emerging standards that are being adopted across the indie platforms yeah that's a

very good point actually so that's one of the issues we're having right now we are actually trying to build one of such standards at the moment because yeah it is actually pretty easy and straightforward but yeah there's exactly a lack of things like that so thanks for pointing that out hi hi um so have you found that designing a game with these um making it spectator approved or liked that streamers will like quickly recognize that the game is made for that and will adopt it yeah yeah so the idea would be to to build features into the game that would just enable streamers to use that to engage more with their audience for example of course yeah cool thank you okay so thank you so much

and have a lovely evening