Beyond Games as a Service with Live Ops

[Music] hello and welcome to this sort of virtual GDC presentation beyond games as a service with live ops we're gonna get started right away one of the benefits of doing this virtually so we'll start I'm going to introduce myself my name is Kristen Cox I've been making games for a little over 17 years mostly worked as a game designer and a game director and I've spent pretty much my entire career making large-scale mmo's so I worked at Walt Disney internet group online parts of the Caribbean online I then worked at Nexon America where I operated the global version of maplestory and then I worked at ArenaNet and I came on to ArenaNet to help develop launch and lodge Guild Wars 2 and then I stayed with doors 2 through to large-scale expansions and

then also the launch of Guild Wars 2 in China a little over a year ago I left a Rena net and I came to join Microsoft at Xbox currently the director of live operations for Xbox game studios Publishing and one of the biggest reasons I stepped away from development and more into publishing was that I care really deeply about live games and the way that the industry is moving forward with live games especially in the triple-a market and I wanted to take on a role where I would have a larger purview and touch more games so at x GS publishing I actually work with a lot of different game teams on helping them understand how to do great live games and how to make their games successful after they're live so today I'm gonna talk largely about live ops I think what it is how you do it and also some of my own personal experience

designing games around it but mostly this talk is about while I boss matters and what it can teach us about designing games for modern audiences and how I think can help us break through the limitations of games as a service I'll talk a little bit about what I see as the limitations of the games as a service ideology for design but before I dive into all of that I'm gonna start with some context so we're all on the same page so this is some fancy numbers about the games industry basically the games industry today is really really big last year generated about a hundred and forty billion dollars and revenue that makes it really one of the biggest entertainment industries and they're all far outpacing things like film and television in most other traditional media there's 2 billion gamers in the world what that means is two billion people touch games on a regular basis in

the world so that also of course expands to being global games are everywhere pretty much everywhere you go in the world you're going to find games and even more specifically video games of some kind now you're all in the video games industry so you probably know this already and this may seem like yes of course games are a big deal of course they are but this if you've been in the industry for a while you know that this has not been true up until pretty recently actually and games have changed a lot in the last thirty years so this is just a graph looking at the size of the industry we're just sort of using revenue as one thing to look at the size of the industry but it's I think tells a pretty compelling story the industry has seen exponential growth in the last 20 years so you know there's a little oven down starting in the 70s and 80s but really it starts to explode right around the turn of the century and that growth in gaming is intrinsically linked to the

growth of connectivity in the internet and that sort of led to this big boom in gaming that also subsequently led to this big explosion in mobile games sort of see that part of the industry get bigger and bigger and bigger as mobile games internet you know internet connectivity even in our pocket goes up and up in so this reality that games industry growth is intrinsically connected to the growth of the Internet has led to this narrative about the games industry that's really deeply impacted the way that we approach game design especially in the last 10 years and that narrative is that this growth was unlocked when we shift in our design focus from games as a product right these are discrete packaged goods that are designed to be consumed and instead we started thinking of them as services so evolving applications that are designed to be engaged with over and over and over again so this narrative is pretty

compelling right when we look at the industry today we do see this shift has happened and when we look at the most successful games today they're very rarely product style games and so that also makes sense with this narrative if we look at the top grossing games worldwide from 2017 to 2018 we see a list full of games that have service components and I think what really drives this point home is when we look at the fact that these games are actually very similar in 2018 as they were 2019 and we're still kind of sorting out the details of 2019 but I think it's fair to say that almost certainly 2019 will share many of the same games as well and this really speaks to the idea that these games are not consumed media they are really meant to be played over and over and over again in fact every single game on this list has a service component and many of them

are also mobile which sort of speaks to the state of the industry today and the way that the industry is composed today so that all seems to make a lot of sense right so this growth really does come from this shift in design focus from product to service my problem with this narrative is that it isn't really about design even though a lot of times design gets cited in this narrative and our shifting focus it isn't a design narrative this is actually a business narrative this is the story of how our business focus changed over the last 20 years you can see that when we look at these two approaches right when we look at what it really means to be game as games as a product and then what it means to be games as a service these philosophies really speak to how games are delivered not how they are designed the two approaches definitely suggest things about design and I think we can all imagine games that really only work

in one category or another but they're more about the way the game is sort of delivered to players and the kind of business you're going to do with the game and a lot less about really what the game is even so more and more game teams are being asked every day to make games as a service games even design teams are told design a game that's gonna be a games as a service design a game that's going to last forever things like that are being asked of these teams because there's this idea that games of the service tells us all kinds of things about design but there's a huge limitation to trying to apply a business strategy to game design and I think we're starting to see that a little bit of friction happen there between sort of player expectations disconnects between

what the game delivers and what sort of players expect and a lot of that I think can be traced back to attempting this sort of forced design through this business lens when we look at games as a service they tell us it tells us a lot because the service has a lot to say but it's primarily focused on good service quality and optimizing the experiences through iteration but it doesn't really have is a lot to say about what those experiences should be we don't really get a lot of clues about how to design the game by looking at games as a service ranking as a service says be reliable deliver valuable services awesome that's a really great but what services should be delivered we don't really know then it says polish and iterate those experiences again a fantastic thing to try to do it's a good process but what experiences are we polishing and what are we iterating towards and then it says you know drive towards KPIs or key

performance indicators right pick targets of things you want to drive towards great we should all do that but which ones and why you know design has to ask these questions these are fundamentally the questions of design what services what experiences why are we doing this what are we driving towards and games as a service just doesn't really have the answers it has processes but it doesn't really tell us what we're trying to do besides really successful game designs are not centered around services I think we all know that they're supported by services and they're supported by this kind of thinking in this kind of process but they're not built around those services design puts service first it can result in really polished but often co soulless games or even worse coldly efficient engagement extraction machines that don't really connect with players on any kind of deep emotional or motivational level games as a service just doesn't provide designers with what they need to

understand how to make great live games it's just missing a lot in that area and we look back at this list it is true that all of these games use a games a service business model and that's great but what's more important for designers looking at this game is not a business model buts more important is that all of these games are designed around intrinsic motivations rich experiences that have all been changed over time and long-term engagement that can become a part of a player's life that's what we should really be taking from these games success the ability to become more than just consumed media that's what keeps these games at the top of this list from the player's perspective year-over-year and some of these games like League of Legends or dungeon and fighter online they've been at the top of this list long before we started I'm out games as a service at GDC they've been doing this with a sort of

different approach before we started categorizing things as games as a service this might seem like a fine line that I'm drawing here this difference between this business approach and this design approach this difference between focusing on optimizing service and focusing on long term player experiences that might seem like a really fine line but in my experience it can be incredibly striking so as an example I'm going to talk a little bit about some of the things that we saw in Guild Wars 2 so these are really scrubbed numbers from Guild Wars 2 they're just really tional and what we're looking at is average revenue per user and retention numbers and on the Left we just have all players so everyone playing Guild Wars 2 this is this is the average or mean of the ARP you and retention for these players and it's fine but on the right we have cohort number one and so corn is

just a group of players that share some characteristic so the subset of all the players and they share some characteristic together and this card has significantly better ARP you average revenue per player and retention than just the general average so these are better players in this measurement anyway so who are these people right what is this attribute they share that makes them so much better than the average player well in this case this is players who spend time in town so town in the worst who are no combat social hubs in the game now it's not as if you really need to spend time in town and Guild Wars 2 in fact there's exceedingly little that actually requires you to return to town the spending money on micro transactions player 2 player market place guilds chatting party systems these are all

accessible anywhere in the game you do not need to go to town to do these things this is very little in the game in the game Loup that drives players to town and yet spending time in town is clearly super important and if we dig into this it's not an outlier group right if we just narrow down all the players that are engaged in the game so players are actually actively playing the game we see that actually majority in several counties a lot of them spend time in town and if we dig deeper in this curve on the right you can see the likelihood of being a townie grows over the players first 100 days plenty the game even though the amount of things that you in town does not increase as you progress in the game and certainly not to this level in this increasing tendency to spend time in town corresponds with this increasing likelihood to become a spender so really for us just being a great highly engaged player seems to be strongly correlated with spending this

time in town I don't really find this particularly surprising mostly because we saw a really similar effect in maplestory when I was working on that so this was a huge shock to me because spending time in towns and non-combat social zones usually means focusing on social interactions like chatting and boating or simply being seen by other players it also needs downtime which can be really important for players that are trying to spend a lot of time in your game to not always be actively engaged to have some downtime where they can relax and plan and just sort of talk um design the point I'm making to bring this up right is the design you have to be thinking about the players experience and what's engaging to them in the game what kind of experiences engagement the game is giving to them not just how to engage them more if we focus too much on just delivering services we might just get bogged down in just increasing engagement you know both Guild Wars 2

and maplestory have lots of progression systems and rewards and gameplay loops and combat and all of these things that we associate with active engagement and we've spent a lot of time trying to increase players engagement with those systems but what we actually saw was that as players invested more time the game they were more motivated by relatedness to other players in the joy of wasting time just chatting in town now I'm not saying that every game chat social space is to increase revenue or that there's some magic connection between spending down time in town and spending in every game but I just want to emphasize that engagement is more likely to come from the actions that feed whatever the game's intrinsic motivator is not necessarily the actions that feed really easy to measure and optimize loops for maplestory end goers - the motivators in the game and we renew this are largely social they're largely relatedness for other games is going to be other things but the point is design has to think about those

things from that direction from the direction of what are the players experience going to be and what's going to motivate them not just from the direction of how is my service optimized so games in the service it's been really great for business it's been great for strategy I think it does actually have a lot to offer to game development as far as processes and just a way of thinking about delivering to players but it's mostly confusing and frustrating for design because it just doesn't fit into the things that we're trying to do and yet more and more teams will continue to be asked to make engagement focused live games right we want to make engagement focused live games but games as a service isn't really gonna help design because it doesn't really show us in a design direction luckily there is an approach to game development that's been around for a long time it's been employed by those working in live games for quite a long time and it is a lot more useful for games for game design and I think it has a lot more to say

about the direction of design and that is of course LiveOps surprise so line ops is a way of thinking about game design and games service that's been around since muds which are multi-user dungeons these are sort of the precursors and memos they were text-based multiplayer adventures that were around in the 80s largely in 90s and that sort of where live ops grew up was in managing these online communities of people playing muds together and it's great live offices at supporting games that are focused on long-term engagement because it's sort of what it grew out of at the heart of my mops it's really about taking care of a game and taking care of the games players over an extended period of time so you can see that in the word itself it's live it's a connected experience it's persistent and there's an operation component where you

evolve and change and take care of it with continual investment I have spent pretty much the entirety of my 17 years in the industry working in LiveOps framework so a lot of experience with live ops and I've seen the approach in in my time go from mostly servicing relatively small niche communities and virtual worlds online to being a big driving force behind the free-to-play explosion and now having deep influence on some of the most successful games today and also the triple a space so live Ops has grown in its importance to the game industry in general and I will say as I work with teams who are transitioning to thinking about making live games and in service based games teams that have some tradition of live offs or some understanding of live operations tend to be a little bit better prepared for what's coming when they make this transition so that's live ops but of course we want to know how do you do live ops it would

not be super fair for me to come and say hey live ops is great it's so much better than all these other things you're doing and then not actually explain how to do it well like most things in game development it's a journey and not a destination just like pre-launch development right line ops is a set of steps so just you know pre-launch development and this is a very nice simplified version of that journey pre-production production launch great we've all been through it unlike this part of development limas tends to be additive instead of sequential so you never actually stop doing the basics you just keep adding more and more on right so we start by doing basics and then we add optimizations then we add master you just sort of keep layering it on um of course lava actually does begin well before the game is launched I know this sort of looks that this charts it makes it seem like you make the game and then you live offset and there's a lot of that but in order to do really great

LiveOps you actually want to start really early you know all the way back in pre-production you should be thinking about your progression design your retention design your monetization strategy then during production you've got to be thinking about what kind of goals you're going to be setting what kind of data you're going to be pulling in how you're gonna do live content how you're gonna do community feedback all of these things need to be part of the process of building your game then once you launch then we start getting into the specifics of these basics live instance process feedback management game iteration and then adding in player behavior analytics strategy adjustments segmentation and targeting breaking your game apart and really looking at how different players experience it all the way into community shaping brand retention and I'm gonna go really deep on all of these things in just a minute I'm actually gonna even though I do want to impress upon you that it is super important to start working on the live ops for your game all the way back in

pre-production you spend most of this talk talking about what happens after launch and focused on those areas where live Ops is really the primary thing that you're doing with your game and also going to be some of those things that are most different from what traditional game developers might be used to who mostly make games in the product so we're gonna focus mostly on that but I did want to make the point that doing these things well before you get to launch is going to set you up for success in life ops there's probably a whole talk there in just preparing for live operations so I said we're going to focus mostly on this part of the journey the process that starts with basics and then as ins mathematician it hasn't Sebastian I'm gonna go into detail about what I mean by those basics and what I mean about devising and what I mean by the the more advanced live us mastery techniques as we walk through this but one more point I want to make before we dive in is that this really is a process that everyone has to go through line ops can be very overwhelming

if you think of all the things that you could do with a live game and try to execute on them at once and so in my experience every game team every game has to walk this path and they have to walk it iteratively they have to add these things in slowly over time even if I'm working with the team that shipped a dozen live games before we still want to start with basics and then when we're good at that had the optimizing and then we've got that under control and the mastery so that we don't get totally overwhelmed so that's the only nine you want to say that even if this is the very first time you're using live ops or you've never not a lot of game before you're really not starting in any of a different position than anybody that's doing it it has done it many times before they might have some you know past learnings and experience to avoid some mistakes but you're all gonna have to crawl before you walk with every single game and that crawling starts with the basics so once your game is live the first thing you're gonna want to focus on is the basics which is really encapsulated with the idea of

iterating on your game this is the first and most fundamental LiveOps skill is making quick iterative updates to your game to improve it and when I say iterate I do not mean releasing planned DLC content that's not really iterative that's just a slow continual form of pre pretty long production I mean actually making small frequent changes to your game in response to player feedback and behavioral data to slowly improve the game over time this might seem obvious like yes that's what iterating a game is but many Triple A games still spend millions of dollars avoiding this step they really try to make the game as perfect as possible before shipping and sometimes inadvertedly that actually hinders their ability to iterate the game later because they don't really design the game to be iterated that might seem Taiwan to go but if you want

to iterate on your game you need to design it to be iterated on and that actually speaks to the architecture of your game content and the way that you are organizing all of what your game is as an example I want to talk a little bit about Guild Wars 2 and this is just sort of a story about not only iteration but managing player feedback so this is a map for Guild Wars 2 it's just sort of the UI version of a map maps are sort of the fundamental building block of the game world and Guild Wars 2 they are these massively large player spaces right they're filled with dynamic events that kick off based on player activity in the map and also some real world timers so when players are playing Guild Wars 2 and they're out questing they don't go get quest from NPCs they actually just engage with whatever event is active in the area they're in at the time and some of their actions might trigger an event to happen or they might

walk into an area where an event is already underway or events might be designed to trigger automatically every X number of minutes but the world is sort of this dynamic place where players are walking around and engaging with it so the Guild Wars 2 content designers crafted the maps to be these branching chains of events that are punctuated by fights with elite enemies called champions the design was the players would play through a sort of emerging story of events that culminates in the boss fight right was pretty traditional it's just more dynamic and that players would just sort of experience this inside of the map probably in small groups or maybe even solo so when the game launched we had made all our maps match sort of this philosophy and then nothing nothing about our plan actually ended up working that way because players did not do what we expected them to do shockingly instead players coordinated over chat to form these large groups of 50 to 100

players that would roam the map as a devastating zerg in this efficient loop defeating all the champions in secession and they called this the champ train and didn't take long for players to get these champ trains organized the champ train ignored the intended story driven loops of events and it removed a lot of the challenge that would come from battling by because urgh you know 100 players can simply steamroll basically everything in the map and so this was really nothing like what the content designers had intended at all and at first we tried to come up with ways to discourage this behavior and sort of force players back into the intended track move things around separated disincentivize it but the truth was as we were playing the game - and listening to feedback the chant train was really fun and players actually really enjoyed it it was a good time it's like fun to join a big surrogate players and just sort of steamroll over things and how you know all this sort of efficient moving around

the map you can just sort of join up and follow along and what we saw was actually players that participated had better engagement in the mass with good sham train saw more playtime so players are really liking this so we basically pretty quickly said ok we we give it this emergent behavior is actually a good thing when we switched our approach instead of trying to stop the champ train we asked how do we make them as fun as possible for everyone ultimately what that meant was we relocated many of the champ trains away from starter zones who sort of moved them around the map to keep them away from people who were just starting out in the game because that wasn't a fantastic experience for them we added boss timers to the worlds for better coordination in the game we adjusted many of the event loops to remove the incentive to fail events because previously some of those bosses that the champ train wanted to get could only happen if they failed an event which was not great and would disrupt the story for people who were trying to consume the story and we made a lot of

adjustments to loot and loops and all of that great stuff one thing this didn't solve was champ trains still made boss battles pretty trivial but we actually learned from this as well eventually and we eventually actually used this to design really cool challenging open-world bosses that were meant to be played by like a group of 100 players at once this is - quaddle which is one of those first really large-scale bosses we added after the game was already live it's meant to be played by group I think the optimal group is 80 coordinating together to take him down um and so this was actually a big achievement for us in listening to player feedback and not reacting and trying to force them into the mold we thought they should be in but actually going along with their emergent behavior and what they really enjoyed to aid but was really important is that none of this would be possible if it hadn't been easy for us to move things around mess

with boss timing change balance rewards without breaking the structure of everything in the math because we knew before we launched the game that we would probably want to make changes to it after a lot of players were playing it it was possible for us to make these changes and I'm not gonna say it was trivial but it was definitely doable and we didn't have to worry about breaking the whole game because we were gonna move around some quest so the point of that just is that content is the first thing you want to think about and while that does require some planning before you to the live game in order to make it possible for you to iterate you want to be focused at this point in your live off journey on that iteration improve the game's performance extend its life respond to feedback that is gonna require that you also build some feedback management processes and also some live incident processes so you're gonna want to figure out how you want to take in that feedback um you know you can start by just you know reading forums or reddit that's probably gonna

give you a bit of a skewed perspective on things so you might want to think a little more deeply about how you're gonna process feedback as it comes in are you gonna set up other channels for players to deliver that feedback to you how are you gonna handle the feedback coming from inside the team as well and then of course live incident processes when inevitably something's gonna go wrong right as you're iterating the game you're gonna make mistakes the games are incredibly complicated and messy and it's pretty much impossible to avoid so what are you gonna do when that happens hopefully not panic and hopefully respond in a way that's gonna be healthy for your community and finally rounding this out is business intelligence tracking I put this on here even though it might not seem like a design area of focus because in my experience teams that are really good at this phase of live ops do you have design take a larger role in that area with all that innovation and feedback response going on really does need to be on top of how those changes are affecting the community and the business to avoid conflicts where design wants to do one thing and business doesn't want to do that so you really need to keep those

guys close and working in concert in order to make this a more successful iterative process for your game so awesome LiveOps basics you're iterating your game if you can get good at this if you can in fact iterate on your game without you know having to destroy the whole thing and you're making incremental improvements and you get to this point where you start saying things like man I wish that we had a little bit more insight into which changes we want to make also maybe some of the changes we want to make are good for somerville players but not another you're ready to move on to the second phase of LiveOps maturity which is live ops optimizing with a focus on personalizing your players experience so this is all about serving content and experiences to players based on who they are and how they behave instead of just making blunt overall changes to the game this is what

most people think you know when they talk about gather and have tons of player behavior data right they think like yeah we're gonna like do a bunch of segmentation and like and analyze it and then it's gonna tell us all kinds of cool stuff that we should do to the game to make it better I do think a lot of teams try to do this step first but I would suggest that it is very difficult to make this step work for you if you don't have the basics down because data is only valuable if you can actually do something with it so even if you collect lots of fantastic data it can be really really hard to action on it if you don't actually have the means to make those iterative changes into the game so we're optimizing what does this really mean to a development team functionally a lot of it is about segmentation segmentation whether we're talking about segmenting by acquisition timing like these players

all joined the game in the same month or these players all join the game from the same marketing campaign or in-game behavior right these players only ever do PvP and these players only ever do story content or you know player motivation right these players really will do anything that gets their progression to to increase or these players will only do things that are socially motivated all of these are segmentations that you can make to your players and look at them differently and it can be an incredibly powerful tool in driving more targeted changes into your game in phase one teams tend to rely pretty much solely on just you know the design team's expertise the things you already know that you should do in direct player feedback right to drive iteration but where that can start to fall down is when you have changes that you want to make that are really only for one group of players and that group

learns may actually be quite different than you expected or different than the group of players that's spending a lot of time giving you feedback tutorials and first-time user experiences are notoriously difficult to optimize without using a data approach because designers and engage players don't usually give great feedback on being a new player because they don't really remember what it was like to be a new player and it's difficult to give that mindset you know just to drive this home this is some retention and data from a game that I worked on and you can see that it gets better over time but what's really important is what we did to make this improvement happen we started by just making basic updates to the game we asked the design team hey could you improve the first-time user experience and they went in and they looked at well these things are not not that great these things are not that great we'll change them and they didn't have an improvement when we made those updates it did actually help but it didn't help that much and we were sort of running into this problem over and over again where design was like well I don't really know what's gonna make a

difference and so then we added a/b testing an a/b testing is just making a small change and having one year of of players that would be the a group experience that change and another group of players the B group not experienced that or experience something different and then see which one is better and then whichever one performs better make that the final change and so you can do this iteratively over time with three small things and that can start to add up to big changes that you can see in this graph we really started to unlock and this wasn't just doing exactly what the AV testing said it was actually a combination of a bee testing and design because as the a/b test results came back that helped the designers go oh this is really what's making a big difference I'm gonna focus on making great designs around that instead so the next thing we can talk about in this optimizing phase is the idea of sort of machine learning this is very popular right now and a lot of times I talk to teams and they're like yeah I want to do

a lot of multivariate testing I really want to build a bunch of machine learning that's gonna make those changes for me and that is really cool and if you work for a really big team with a lot of fantastic analytics available to you you can do that but I do want to say that you don't need that um you can actually get a lot done with pretty basic information as an example there's this example from maplestory so maplestory this is quite a long time ago before we really had a lot of fantasy machine learning at all so this is actually one of my favorite examples it's a system we introduced in maplestory where when we saw players exhibit behaviors that correlated with churn and this is pretty basic stuff like I'll get in less frequently leveling up less frequently not talking to their friends as much we would look at that behavior and we say oh you're doing these things you get a little flag on you and then there's a chance that when you log out you'll get this special message which is basically hey if you

log back in tomorrow or in the next three days you're gonna get a free gift and we didn't really worry about making term prediction model perfectly accurate we were really just trying to see if this would work with basic information about chores really obvious stuff and in fact we actually eventually expanded this to allow the gift to trigger on accounts that didn't meet the term prediction criteria because the effect it was having was so positive it was a really really successful campaign players were actually really into this idea they didn't really know why did we get the gift and there's a lot of conspiracy theories about what would get you the gift and it just sort of generally increased retention across the board and even though that's not what we think of when we think of like super tight focused machine learning algorithms that you know optimize everything this was still a big benefit to the game and still really worthwhile to do so I would say even if you don't have access to that kind of sophisticated machine learning you can

still try some things that can have a really positive effect it's just important if you don't have that kind of sophisticated analytics to make sure you do look and make sure it did have a positive effect not a negative effect and then react to those changes accordingly okay so we have the basics and then we've added in this optimization where we've really focused on personalizing the experience and targeting offers and content to our players like reduce risk and optimize individual player experience so if you're doing both of these things and you're really having a lot of success already in getting these things done we can move on to this final step where we add in the most sophisticated LiveOps tactics which is sort of called laughs mastering and specifically when I think about this step they're really the most advanced tactics for live ops I think about expanding your games engagement beyond the time players spend playing

and this phase you should be thinking of your game as more than a collection of systems and content right discussing the game making fan art reading tie-in comic books meeting guild mates in real life all of these are actually a part of your game and they should be something that you care about designing and that experience becomes vitally important as you move through this live ops journey because if you want to get to the place where you're strengthening your brand and your loyalty and you actually start seeing organic growth and sustainment ie not always having to pay for all your acquisition you're gonna have to focus on more than just what's happening inside of the game while a lot of this work happens quite literally outside of the game supportive raishin with game systems is really what separates the best from the rest in this area it really behooves designers to think a lot about what these experiences are for their players even outside of time spent in the game many of the examples for this phase we have to talk about systems in the game and not just

outside community events and things like that many of these examples are things that lots of games actually do but very few do really well an example I like to use as matchmaking it's something that we've gotten really good at as an industry if we define it as predicting who will win competitive matches but we've struggled to really create matchmaking that predicts who will have fun together and so we actually ended up using matchmaking for something a lot more social in Guild Wars 2 that I think is a really interesting example of this kind of thinking of engagement beyond just game systems so I'm going to talk about mega servers which is a whole program that we developed in Guild Wars 2 after the game had launched but when the game launched the way that Gilmer's 2 worked was players would log in and then they would tell us what map they wanted to go to right in this case silver waste which is a map and go worst to you and then we would go okay you want to go to silver waste well let's see if there's already

a silver waste server up and running oh there is here it is and there's an empty slot in it so awesome you will be moved into this server and that was sort of the way it was supposed to work right players would be sorted in the server until it was full and then when it was full and another player showed up they come along and say hey I want to go to silver waste and we'd say well there's no room in silver waste so we'll make a new version of silver waste for you and you can go there and then you know that will eventually fill up too and this all sounds pretty good on paper but in reality when the game was live what happened was we had a bunch of servers for maps and then naturally players would empty out of those servers in a very organic unorganized way and eventually we would end up with sad lonely players on an empty server with really no way to get more players in because they aren't Nemo recent server and so when someone new

comes to join they would actually go to this server that was most recently spun up instead of to the empty lonely person above worse we didn't really do anything to ensure that when a player join your server you were happy about it in some fact sometimes players would not be happy about it they wouldn't like the person that they were randomly paired with in their server and there you know enjoyment of the game would go down so this led to a lot of inefficiencies right you can see that this is also from a services component incredibly and efficient and so we decided we wanted to take a totally different approach to this a lot of it was spurned by this idea of can we make this more efficient but it quickly expanded to asking can we make the game more enjoyable so we changed everything about how we did this now when you logged in we would say where do you want to go but before we put you there we'd ask you a whole lot

of things about yourself are you in a guild have you played in a party recently what kind of content do you like to play what's kind of playtime you usually have what's your level what's your class what's the last map that you were in and we've learned all these things about you and different weightings and we'd say okay now we know a little bit more about you you're not just a random person you're a specific person and you want to go to silver waste which is great we actually have multiple silver waste servers open which one are you going to be the happiest on well this is the one that we think you're going to fit the best with because of your play style so you should go there and this was took some learning right like you know the waiting we didn't we had to make some guesses at first and then we had to optimize that over time we also made some fairly large mistakes early on for instance the first version of this sorting algorithm just straight up didn't ask what language you spoke and it turns out that sorting people into servers with a bunch of people who don't speak their language makes them pretty unhappy that was

fairly easy to fix right we started we added the language option and pretty quickly and that was very heavily weighted because even if you would get along really well with a server full people if none of them speaks Spanish you're probably actually not going to have that great a baton so we might actually find out that maybe you do match the first server here a little better but we're going to put you up here because that's where people are going to be speaking your language and that makes a big difference this effort which was driven by engineering and design ultimately just really improved the social experience in the game in general and we had we had general uplift in player engagement and retention and and you know I don't know it might be overstating a little bit but as our be matured I think this system really helped the game grow and thrive after it got over that initial excitement of launch and really became more about a continual community such as one example

of using in game systems to to drive the kind of social behavior we want that can extend outside of the game itself another one is to think about community events when we think about community a lot of times designers think about you know forum moderation and fan meet ups but communities really form and thrive because they have something to talk about and that usually actually comes from the game itself and so sometimes we actually want to create content that's about making a community moment and not just about driving playtime or increasing revenue but it's really just about getting players something cool and exciting to do so one of my favorite events we ever ran in maplestory was the random Beauty coupon drop event so we had these random Beauty coupons you can buy in maplestory and they were this consumable item and when you use them they would change your appearance totally randomly you get a random face random hair random skintone random eyes

all these would be random this was not a very popular item because it was truly random and there was hundreds of appearance options so you can end up with something pretty funky was a pretty significant roll of the dice so they weren't that popular you didn't sell that many of them but we decided it would be fun to run this event where we dropped them commonly off of all the monsters in the game so now when the players were they would defeat an enemy and they would get sometimes multiple random beauty coupons in their inventory and so they had access to all these random Beauty ones and tons of them and they started consuming them all the time and we started seeing these crazy combos we'd never see in their appearance and people were posting screenshots on the forums but look at this weird you know green skin red hair thing that I got and it really drove a ton of fun engagement it was just light and fun and players were really enjoying it it actually ultimately did end up having a positive effect on our revenue he's actually

expose people to cosmetic options they didn't know about before and they actually wouldn't bought this permanently but mostly it was done to drive this kind of interest and get people excited about logging in it was a pretty generous event right we were giving away something that technically you know we sold and had you know monetary value but it really was incredibly worth it and a lot of fun okay so this is the journey I wanted to talk about today right starting with these basics moving into optimization and then mastering every game the CSIs a lot seeks to do LiveOps has to tackle these coal areas core areas right we have to think about content we have to think about players very we have to think about community engagement all of these things are really core to live ops some teams are gonna focus on different portions for different amounts of time some are actually going to you know move pieces of this and different lines in the journey but what's really important I think for me to drive home to you is it really is a process where you want to

add things on slowly don't try to take it on all at once my advice is start with content then moved appliers then into community that's the method that has really been successful for me and my teams in my career there's some play to move some stuff around and you'll get examples from teams you know that really focused on a lot of the stuff that I have in mastery early on and that may have been really good for them but in especially if you're starting out from scratch now we're having done it before this could be a good road map to help you get your feet wet and really get in there and start working and I would also say resist the urge to move on and add more complicated techniques until you really get the basic ones down really get to the place where you can iterate before you start trying to do fancy things and segmentation really get to the place where you feel like you can personalize and focus before you start trying to you know drive a lot of viral excitement about your game and then I think you're gonna have a much easier time getting success and help keep your players

happier with a more predictable game that's really focused on giving them what they need yeah what's this all about so this is a lot of this talk just came from me feeling like design games in general but design really needs more than just services it's great then we have a service mentality now it's great that we have access to these new ways of thinking about business strategy and new ways of thinking about games but design does need more than that and I think it's important for designers to push for our view of live games to be equally as important as business news view you know to design our games our culture right that's what we make we create these communities of people that play together and have a shared interest it's more than consume media it's about deep intrinsic motivations continual engagement it's really about thinking about our games through this lens you know sometimes at Xbox we talk

about games as community the game itself is a community and that really is the core of it we also have about games us culture and the effect that games have outside of just the the client and the thing that players play with and I think that that is getting a little closer to the heart of how design sees live gigs okay so I leave you here with this journey in that declaration the design needs more than just games as a service thinking it's important to us and I would like to encourage you all to keep pushing to make designs voice heard and use these live ops techniques to make your live games better now in this virtual version of UTC I can't take questions right now which is unfortunate but my email address is right here if you questions or you just want to talk about LiveOps please don't hesitate to drop me a line and let me know what you think thank you so much