Keynote + Tools and best practices to improve game quality (GDC 2019)


Learn about the latest tools, services, and best practices to help distribute mobile games on Google Play. We’ll cover how to create high quality games using Android vitals, improve player conversion with Google Play Instant, and build effective pre-launch strategies to maximize the lifetime value of your players.

00:00 – Keynote
13:00 – Discovery & distribution on Google Play
32:51 – Using Android vitals to identify performance issues and improve app quality
45:32 – Building the hype – effective pre launch strategies
59:10 – Lifetime valuation of games

Mobile Developer Day at GDC 2019 Playlist →
Subscribe →
good morning everyone I'm Frank and I'm 10 and we lead android games engineering and play product management at Google first welcome to GDC we we know you have your choice of 10 a.m. sessions there are only about a thousand others right now so we appreciate you being here with us today so like many of you we have spent our careers and most of our lives playing and building games there's anyone pop-quiz does anyone here remember strike commander or Wing Commander 3 or Wing Commander in general ok whew that was Frank yeah that's thank you to show you just how hardcore he was this was a license plate wait you were in like high school in it well it wasn't high school unfortunately I was quite a bit older but yeah this was in Chicago and before I even started in the games business I was may have been a little bit too large of a Wing Commander fan so no issue with that for my part I helped build the Nintendo GameCube the Xbox 360 and the PlayStation 4 and to prove my nerd cred here's an old picture kind of showing my bad hair choices late at night as we brought up the high performance getting an animation code for the Gamecube that was generally used in all GameCube games and Wii games for quite some time I spent 17 years working on Xbox I worked on all three of the Xbox's what I consider to be my crowning achievement for that was I was the development manager for XNA game studio so if any of you remember XNA and the XNA framework and all that stuff that unity is based on today you can blame me for that too I joined Google because I see the potential here of Google's reach and really the expansive capabilities that Google has that can really take your game businesses to the next level now Frank and I also care deeply about diversity in the games business and we mean that in every sense of the word we live in a world where large driven Triple A games like God of War can live side by side with mobile Indy premium title like Florence on Game of the Year lists that's absolutely outstanding and so we just care a lot about fostering both ends of that spectrum everything from loss Triple A to small indie and all the business models in between from Premium Subscription ads and so on your businesses are helping to fuel incredible growth in the global mobile games market which are forecasted by new zoo to grow yearly by double digits for the next decade by 2021 mobile games revenue will be 60% of the global games revenue and we're seeing some really interesting trends in the industry like the explosion of gaming in India which isn't you know not traditionally a huge gaming market and this is in part due to the amazing success of pubsey mobile this game has become a cultural phenomenon there the Indian Prime Minister has made references to it and couples have made the news by putting out pub G themed engagement photos but success is not only for the large developers every year we see inspiring stories like small giant games who launched empires and puzzles way back in March 2017 it became a top grossing title on play last year and they just very recently got acquired by Zynga they started out this tiny team so huge congratulations to them however as people that have been around games for quite some time we know that Google hasn't quite always lived up to its potential we know we can improve the experience for game developers and we want to share some of the things we're working on to make it better for you you'll hear from various Google teams throughout this week but for now we'd like to focus on Android and play and things that we're doing there to improve the ecosystem as the engineering lead for android games and graphics my team's primary goal is to help you reach more of the 2 billion Android devices that are there by making it easier for you to build for them we hear your feedback today that it's too hard especially when building high fidelity games my team's focus is to build tools that will help all developers bring mobile games to the next level pushing the boundaries on graphics multiplayer experiences and so much more Android has traditionally been one of the toughest platforms to develop for as though that's like a badge of goodness given the fragmentation of the ecosystem and limited availability of game development tools to deal with this complexity when you think about all of the SOC vendors that are out there and all of the different OEMs using those SOC s it's very very difficult to know what to target to hit the kinds of things that you're going to be after in terms of what your game needs to do in terms of performance or look and feel at the same time mobile games are becoming much more complex with more titles moving from console and PC to mobile as developers you know these games need advanced debugging and development tools and those are the kinds of problems that my team is going to be focused on solving to do so we're making big investments to improve the development experience on Android we're expanding our scope and ramping up teams to unblock game developers and problematic areas and also working with the platform teams to build solutions at scale there are a lot of hard problems to solve and if you're up for a challenge we are hiring hey you know I used to optimize games for the Gamecube right no I didn't know that all has it been since you coded 7 or 8 years maybe oh well maybe you should keep your day job you don't just say we're also investing in tools to help give you better insights into what's going on we know that the CPU and GPU profiling stories on Android today are not up to what they need to be we need again better debugging workflows and we're investing in wider Vulcan availability as well so that you can tap the graphics potential that's in a lot of these mobile GPUs today we've also heard from developers that it's hard to find information and resources on mobile game development industry lucien so to help today we've launched a new site as the central location to make great games and maximize your reach keep an eye on this as we're going to continue to add new resources overtime these are just a few places we're working on to make things less complex for you to build games on Android and we look forward to sharing more with you throughout the year in fact we'll have much more to say at Google i/o in a brand new track dedicated entirely to games this is the first time ever that games have been featured at i/o and this speaks to the commitment across Google but because we really want to be the full lifecycle partner for you as you're building these titles if you can't make the event please make sure to follow along on the live stream there'll be a lot of information we know that developing your game doesn't mean the work stops when you publish this is where you shift focus to improving the games performance in the marketplace and that's where Play comes in so over the past year I've spent a lot of time speaking with the game developer community around the world and a couple things really stand out that we want to focus on first discovery it's hard and we want to help your game succeed on the Play Store every day more games are appearing but we only have the same number of pixels and we want to make it easier for everyone to get discovered now we know that it's especially hard for newer and smaller game developers and we want to do a lot more to support you so if you visit our booth starting on Wednesday in center hall you can learn about some of the programs that we run specifically for Indies that help amplify your great work such as the indie games showcase so to learn more please go to this site so while we're improving the storefront we continue to offer new and effective ways to help you distribute your games to users last year at GDC we announced the launch of pre-reg which allows you to start marketing and building up the pre-launch hype and demand for your upcoming release this is a tried-and-true best practice across the games industry and it helps build that hype with players you're gonna hear some big news later today about pre-reg and some great new features so stay tuned for our session called building the hype there's a lot more that we're thinking about when it comes to games discovery your games have so much creativity and so much innovation and then we tell you to condense it into one icon on the homepage of the store so we want to do more to show the soul of your game to users before they even click through to the store listing page so what you see here on the left is one of our new auto-playing treatments it should be going out into experiment this week and in general you should be expecting to see more and more of your videos and screenshots show up on the home page over the coming quarters on the right you will see our live ops events page and this can help drive re-engagement and discovery for your live ops events and games major updates can also be shown here and so in general we just want to be there to help amplify those big beats in your business we also launched Google Play Instant for games at GDC last year players can hit the try Now button on your store listing page and go straight to a demo experience of your game within a matter of seconds with no installation required you know I remember when we had to cut discs for demos and we the Front's of magazines if anyone remembered gluing you know the going get the discs off the front of the magazine and then rushing home and having fun anyone missed that you know okay well this is the best part is that now you can download and play a match of something like arenas free fire in about ten second which is a very very different world to be living in great times we live in we also enabled instant for premium titles like hyper beards dr. MEAP an incredibly cute take on the match three genre we also want to make help you make money we want developers to have more options on how they monetize in general in-app purchases of course are super important but we want to support other methods like subscriptions premium games or rewarded ads and we're constantly trying out new approaches here we see huge benefits in helping developers diversify their revenue streams and later today you'll hear more about the success of developer developers like pledge Andheri and congregate stay tuned we're investing in even more options and we hope to give you even more choice in the future so if it isn't already clear Google is super serious about helping games throughout their lifecycle and if you want to learn more come to our booth in the West hall or visit our new game Google for games site before we kick off today I wanted to share some other events we have lined up join us tomorrow at 10 a.m. for the big reveal yes then on Wednesday there's a full day of sessions from Google cloud and finally Wednesday through Friday there were over 40 Google sessions and talks throughout the conference full agendas and live stream details can be found at gqo slash GDC 19 here's a quick overview of today we have four main sessions full of updates and tips to help you succeed on Android and play the first three sessions will cover tools monetization strategies and new opportunities then make sure to come back at 2:40 for a more technical session focused on building games on Android we appreciate everyone being here today thank you for choosing Android and play to create and scale your games and we look forward to seeing what you do next with that will hand over to Ben Frankel who will share more about distribution discovery thank you all thank you all right hello everybody I'm Ben Frankel I'm a product manager on play and today I'm going to go a little deeper into some of the things Tian talked about earlier about how we're improving discovery on the store and joining me later on in a little bit is my esteemed colleague Kobi Glick who will be talking about distribution on play but before I get rolling I wanted to kick things off with some audience participation get you guys energized all right and so you know just to provide some context for this about a month or so ago I was looking at the latest results of the GDC stay to the game industry survey and among other things that asked this question you know which discovery methods you the game developer find most effective and the top five responses in no particular order were word of mouth promotion on the digital storefronts that would be the Play Store for example social media very popular YouTube video which apparently is its own category and traditional press and bloggers so take a second to let this soak in all right look at them closely keep them in your mind because they're going to be coming in handy in a second so now what I want you to do is to take a moment and guess which discovery method most developers selected and which you believe the fewest so take a guess what's the yellow bar and what's the blue bar all right I'll give you a second to think about it all right which is number one and which is the last place one okay all right so survey says that imagine a drumroll the number one choice is social media I'm very exciting I know and and last place last place survey says it was promotion on the digital storefronts and for completeness this is how everything else shook out so raise your hand if you successfully guess that social media would be ranked first who thought social media would be ranked first I think there's more of you who really thought that but now raise your hand if you successfully guess that the storefront would be ranked last not that many of you like me I was a bit surprised too now raise your hand if you just thought this whole thing was contrived to make the Playstore look good all right so there's at least some cynics in the audience and I appreciate that so when I saw these results I wondered how is this possible with the billions of people that are visiting the store and the fact that all discovery roads end in one app store in the other where is this disconnect coming from the developers believe that most users have made up their mind before they come to the store I was wondering the same thing well I did some digging and in fact a majority of store sessions are discovery focused with people searching for things like free games or scrolling through the homepage so my takeaway was that we as play have a tremendous opportunity to improve the effectiveness of game discovery and here is how we think about that opportunity so for years game discovery started with the tried and trusted game icon lots of icons all of the game icons and we know that you put a lot of effort into carefully crafting these things but there are physical limits to how much information these could possibly convey to users the game icon is really just a narrow keyhole view into the store listing and the store listing has a lot more information there are video store screenshots but the store listing is still just another slightly less narrow keyhole view into the layer below the actual game and each layer sure provides more information than the last but playing the game is really the first time the user can tell if they found a match for what they're looking for so we asked ourselves how can we elevate what's inside the game the soul or essence of the game to the front line of discovery how do we move things from this bottom layer up to the top further and further up and over the last few years we've impacted progress towards this goal some of this TN covered earlier moving content like trailers and screenshots from the store listing to the top layer we've recently expanded to auto-playing elements like this which cha mentioned earlier and there's also a ton of activity inside of games in that bottom layers the stuff that's happening inside the game daily events competitions and promotions and that's where live events come in and why we have the events tab and we believe the next step in the store's evolution is to bring more of the essence or soul the game up to that top player and Google Play instant plays an important role in this area so the sex section is going to provide an update on how the product has matured since we talked about it and announced it at GDC last year so as a quick refresher Google Play inciden Able's users to start games instantly with a single tap to try the game without installing or uninstalling and gives you the ability to launch games from anywhere all over the store and place properties beyond that and since GZ last year developer adoption has been booming over the last six months we've seen 3x growth in instant games and sessions have grown even further so we asked ourselves with that success can we elevate instant games further what else can we do with this format could we shift instant games from being just part of the existing discovery experience to being the discovery experience well the play games app playlists are the first real attempt to do this imagine in the future being able to click play all on the new and updated section on the store and having an experience like this where you can go from demo to demo to demo until you found something you love to play think about how different that experience would be when you are looking for a game and to play games app in the store are going to continue to invest and iterate and experiment with concepts like this and our success has also emboldened us to elevate instant games from the store listing to the higher up of the you know the higher layers here on the homepage like in this ads example one of the new try now treatments that we're testing out so expect to see more tried out buttons on the homepage in the near future and all these new experiences were really made possible because of the incredible support of our partners who build instant games so the question is why were they building them what's behind this growth so there are three driving forces the first Google Play instant is delivering business impact to you the game developer second we've made it easier and faster for you to build instant games and lastly we've cultivated an ecosystem of implementation and technology partners that can help you build instant games so diving into business uplift the key thing I want you to take away is that Google Play instant can work for any game any of your games and here are some here's a sample of recent successful games to help illustrate a few key points so number one they've all achieved improvements for the developer in terms of number of users that install the game that's a big important thing to emphasize from 6% to 26% but in addition to that it's also important to emphasize that success can be found in any category of game so whether or not you're building a strategy game an arcade game or a casual game you can find success in a second irrespective of how you monetize your game Google Play instant can work for you and that impact can be particularly pronounced for premium games as this example illustrates and lastly you can find success regardless of how you have built your instant game so whether you prefer to use HTML like flip the gone unity or a rodeo stamp unity like rodeo stampeded or a custom engine or app on board you can be successful and speaking of developer tools we've made major strides in this area so in the lead-up to GDC our early access partners were taking months to build an instant game and with our investment in unity and Koko's we got it down to weeks then we got it down to days but going forward we aspire to do two things enable developers to build instant games in a matter of hours to cut that down even more and enable every sing one of you in the audience to build an instant game and for those of you who may be skeptical we have a video demonstration that shows that this is possible courtesy of a pond board the demo was of their new studio product it was announced last week that takes a codeless approach to building instant games and just a warning the demo runs at 18 times normal speed so I will be talking quickly all right so you create a project and after you've created the video and audio assets you can actually import them into studio where you can drag and drop them to build out your user experience you can set up your transitions and link them to gestures like clicking and swiping you can quickly preview experience and iterate through that process and as you can see in this time-lapse here in under 10 minutes you could have your first iteration of an instant game and this approach to build instant games could unlock a whole new range of use cases imagine in the future having different instant games for every user segment for every content update for every live or vent or promotion that's the future we're building towards and another important signal of instance maturation is the implementation partner ecosystem that's grown up around the product and they help solve critical problems in adoption for us and they have been critical to our success as a result so to learn more about these partners and everything else they talked about today check out our GDC medium post here and come to our session at the Google booth this Wednesday to learn how to build instant games with unity and with that I'll hand things over to Cobie to talk about the power of play distribution [Applause] thanks Ben hi everyone Ben showed you how plays content first approach and instant games help you with discovery I want to tell you about how play helps you with distribution I'll cover three things and if there's one word I want you to take away from this talk it's dynamic I'll talk about Android our bundles and Google Play dynamic delivery I'll talk about preparing for 64-bit and how you can keep your users on the latest version of your game so let's start with the bundle we launched the Android app bundle at Google i/o last year and more than 60,000 apps and games they've already switched to publishing Android our bundles in production on Google Play you can build them with Android studio you can build with unity and recently also with Coco's creator and absent games using the app bundle becomes smaller because plays dynamic delivery only sends the device what it needs our data shows the tether as a result game switching to the bundle see more than 3% increase in their install success rate large games benefit even more with more than 4% install uplift Gameloft's minion rush Despicable Me switch to the app bundle reduce the size of the game by 20% which is obviously great for users and as a result so more than 3% increase in their install success rate and there are other benefits too it's not just more installs we see that game switching to bundle get more updates which makes sense since the update patches are smaller too in addition we've learned that the size of the game is a major factor in the uninstall decision we're looking into this area and we plan to share more numbers later this year but generally speaking the larger your game is the more likely it'll be uninstalled by users will limited storage a few months ago we with you that by August we want you to provide 64-bit versions of your games this provides tremendous performance improvements and sets you to a world with 64-bit only hardware some concern we've heard over the years is that adding 64-bit support makes your game bigger because you need to ship both the 64-bit and the 32-bit libraries to the device with the Android app bundle this is no longer the case Google Play is dynamic delivery will only ship the required device the required libraries based on the device architecture and the great thing here you don't need to do multi apk to achieve that you can do this with a single publishing artefact to learn more about preparing to 64-bit please join this session we're co-hosting on Wednesday with arm all the details right here and please also read our blog post for more information you all build great games and you keep improving them with time and for your users to take advantage of your latest and greatest features they need to be on the latest version of your game and you're probably familiar with plays auto-updates which reach tons of users but given plays enormous scale there are always users who will not be on the latest version of your game so recently we expanded our dynamic framework and introduced an early access program for in up updates with this API you can update your users to the latest version of your game without actually sending them to the Play Store you can see here an example of how the chrome app is using this API and I want to show you two flows that we support the first flow is the immediate update flow this is for cases when the user must be on the latest version of the game to start playing today some multiplayer games already do this kind of blocking flow where they send users to the Play Store to get the update and ask them to come back afterwards the great thing about this new API is that it allows you to do this all in line you can keep the user in your game while they download and install the update the second flow is the flexible update flow this is for cases where you want to upsell the latest features of your game to your users you can have the user download the update in the background while they keep playing your game we are ready to expand the early access program to more developers so if you're interested please register with this URL right here and we're also planning to launch this to all developers publicly soon now I want to tell you a little bit about our future plans in the distribution space mobile devices keep getting better every year more computation power more memory and storage better graphics as you know better than anyone else this creates huge opportunities to deliver exciting mobile gaming experiences we realize that to do that some games will need to be large because they have high fidelity graphics or have a lot of content we also recognize that some of the tools that Play has provided for large games over the years like Oh babies were not always the most optimal ones so the good news is were investing in this space in 2019 before I tell you about our future plans I want to make sure you're all aware of how things actually work now because we made some changes in this space lately and this is about how play validates the size of your game in the Play console historically we've only looked at the upload size of your apk this doesn't make much sense in a world of up bundles and play dynamic delivery anymore so recently we changed size validations across the board in the play console and we're now only looking at the compressed download size of your apk or app bundle I started this talk by telling you about the immediate benefits of switching to the app bundle like more installs and more updates I want to make sure you're aware that bundles are also the few format for games on google play later this year we plan to provide you new tools to help you move away from OB B's this will be based on bundles we're going to provide you new tools to deliver and optimize the delivery of game assets this will be also based on up bundles so switching to the bundle not only provides immediate benefits it sets you up to benefit from our future investment in this area but while it's really exciting to talk about the future we all live in the present so I want to mention a couple of things that are happening right now you're probably all familiar with the large download warning we show to users in the Play Store and over the years we've heard some developer feedback about how this large warning negatively impacts the conversion rate for large games on play so we're making a couple of changes here first of all we are going to be more transparent with users and tell them the exact download size as a result we will no longer trigger this warning when the user is on Wi-Fi connection and in addition we will make the option to proceed with the download more explicit and one more thing we will no longer trigger this warning at a hundred megabyte starting today we're triggering this warning at a hundred and fifty megabytes the size limit of the Android app bundle will be increased to 150 megabyte to make sure you could take advantage of this we are currently keeping the size limit for apks on a hundred but you can still use apk and OBB until we completely revise this face later this year so today I talk to you about our dynamic framework I spoke about the immediate benefits of switching to our bundles like more installs and more updates I showed you how it better sets you for a 64-bit and about our latest API allowing you to keep your users on the latest version of your game finally I told you about some of our plans to better support large games on Google Play and about some changes that are happening today like the new hundred and 50 megabyte limit threshold I hope you enjoyed this talk thank you for joining now I'd like to call Fergus to talk about Android vitals everyone I'm Fergus Harley product manager goo Android violence is a program by Google to help improve the quality of games and apps on Android goo system and we've been running this program for about two years now I'm gonna give you overview of the program and they're gonna have felipe from taps games come to stage and talk you through how they're using android violence and a bunch of other tools to help improve the performance of the many games that they are building so to start off hands up who wants more 5 star ratings and less 1 star ratings okay if you don't have your hand up you might be working for a competitor so you might want to change you up but when we look at the data for reviews on Google Play what we find is that the vast majority of people talking about negative things in one star reviews are talking about bugs and stability and other performance issues that are having it causing problems on their device and so 42% are just talking about stability and bugs alone when we look at the five star reviews on the Play Store we see that 70 plus percent of the time people are talking about speed design and usability positive performance signals so what is the Android Wireless tool in the play console it is a place where you can go and see 15 performance metrics about your applications these are broken out into five performance areas battery stability rendering startup time and permission denials for each of these metrics we provide you detailed reporting with breakdowns by apk OS version and device these help you be able to understand where should you invest your time in improving the technique for 'men serve your application and these metrics are powered by real users out there the hundreds of millions of Android users who've opted into sharing their data with Google and you are game partners and these metrics are becoming more and more important to the business success of your application because we can see that these impact retention engagement and monetization of games maps as a lot of information in Android vials so to help you prioritize where you invest your time at the top of the Android violence section we provide core vials these are the four metrics that we think are most important to your users and we have data to show that they do impact your performance and so the four metrics are in the stability area crash rate in a noir rate in the battery area we have stock wakelocks and excessive wake-ups for each of these bad behaviors we provide for each of these core vials we have a bad behavior threshold and the bad behavior threshold enables you to be able to know if you're one of the worst performing games for that metric for the other violence metrics we want you to look at them and be able to see how you're performing relative to others and I'll talk about that in a second some of the vitals metrics we have right now as such as the rendering metrics just capture UI toolkit rendered frames and so they don't capture opengl and so there are no bad behavior thresholds for the rendering metrics are the other non core Voyles it's an area that we are working to improve in terms of having more insights for your games that are using OpenGL one of the things that Google Play is uniquely positioned to offer is benchmarking across different games because we have the data across the platform right now we offer kateri level benchmarks so for each vital metric you can go in and select which category of the play store that you're in and across all the games categories that we currently support and see if you're doing worse or better than the competitors in that market and so you can see then that games that are in the racing car games area usually have a slower start up time thing other games and so this enables you to be able to understand how are you doing versus your competitors and be able to understand if you should be investing more in that area than others this is an area that we are continually investing in and you should expect to see a lot more advancements in this area over the next couple of months another major piece of feedback we got on the Android violence project is that people want to be alerted when there was a major change in any of the metrics so we have anomaly alert emails and and notifications in the console these show you when there's a spike in any of your crash rating or clusters are in any of the core vitals you can be able to go in and sign up for these anomaly Arctic alert email as yourself I just go to the notification preferences section of the Play console sometimes we hear developers talk about how Android BIOS only applies to bigger developers and we know that developers who are smaller indie developers are getting a lot of value out of the product and if you're not using it already I'd encourage you to join the over 100,000 game and app developers that are using it today now I'll hand over to Felipe from taps games talking about how he's using unreal buttons and other tools for taps games hi everyone my name is Philippe Yashida I oversee marketing analytics for tips games and before anything else I want to talk to you a little bit about temps so your Brazilian cedar in the market since 2012 and we have a little bit over 180 people in your studio in San Paulo we have a little bit over 30 million users but more importantly we've developed ourselves over 400 titles across the years and I want to talk to you a little bit about how we do with this today so the first thing is having 400 titles is complicated so when you're talking about this we split up or following two main areas so have legacy titles which are the ones that we are keeping it 10 they technically updated but we don't really put anything else in the game so we don't get in features the important part though are the games that are life so the games that we have game squads so teams that have with people from different categories so developers and testers and QA etc everyone in the same team and of course several of the teams are not directly related to a single titles such as marketing for example each of the cycles the live ones they go through several stages development right I'm pretty sure you saw several different frameworks for this firstly splitting maul as is five categories so games go all the way from conception we're figuring out the market seeing what actually is happening and seeing what game we should attribute all the way to having the gaming operation where we just keep it in life keeping events and pushes and how that goes across several of these though some of these stages a very important part is deployment right I know some of you have different takes on it and we tend to use it on a very simple three stage step they're going to talk to you about the first one for us is pretty straightforward so just talking about building so getting our game in the store itself we have a specific pipeline because the amount of games that we use to publish every month this is highly automated we use a go CD pipeline where we ship things to the store go through automated and manual testing and get a life the most important part though especially for deployment is the rollout monitoring which is where vital specifically has been very useful for us this has first takes from seven to ten days where you're rolling out across user base who is getting the new version of the game we have both business and technical teams monitoring rollout so while the business team is just using simple tools either through the console itself or through our own analytics which is largely based on firebase by the way these are just checking major metrics so our getting major dropped in any use re getting actual spikes in reviews and how we can actually act inept from a technical perspective though this is where our developer and testing team are the most conscious about what we're doing during this phase we are monitoring both a and R and crash rates which are the the issues that we have the most most of the time and these are being checked constantly and proactive across the whole period after that's done we go into live monitoring which is business as usual for most games so there's not exactly deployment but it's where we shift from a business perspective we shift to a very much more proactive approach to attracting the deeper metrics in the game that we can't really see during deployment and from a technical perspective we go into a more reactive site based largely on alarms so using crashlytics and vitals to having alarms on all the stores and all the games that we have because honestly if you have four hundred titles tracking all of them manually is just not feasible I want to share with you a simple use case that I'm sure several of you have gone through across CES which is was one of our releases last year on bed Wars which is our current bigger biggest title so we were releasing a new version server update several partner updates and of course getting a lot of technical fixes online in this case both our technical and business teams got a sign as soon as we were running out so we got a spike on reviews our games usually have relatively good reviews so getting a spy saying people couldn't access a game or the game is crashing does that be something wrong and lo and behold we were also getting a spike on ANR's of the game investigating this was done as soon as we're reaching 20% rollout so it didn't actually reach the whole user base and pretty much what that was done was just a Android specifically so we were checking we're saying that we were getting a lot of crashes on specific third-party SDK how do you serve that as everyone else does so pull it out stop roll out and actually remove that SDK it was something that would take a little bit longer to test it's very important that you notice that this is something very common right there's no secret sauce here but I'm pretty sure if you have seen it before it and for us being able to catch this before it actually impacted our revenue so it actually affected most of our users it's very important and this is part of what we learned across years so first the to themselves don't have to be that complicated so know that you don't have to necessarily overshoot you don't need a complicated technique oh and how is this seem to make it work you can just have tools that give you what you can react upon second quality of the app differently from what you should do in the past and from what most most people do it's not about just a technical team everyone is impacted and everyone should be paying attention to what we get your net performance and more Indian quality it's about discipline and effectiveness so know that every single time that you're launching something you have to be monitoring actively and once it's life you have to pay attention you can't just expect it to be ready for you so I think that's a little bit about it there when should share with you guys and when should call Fergus back thanks very much Felipe for that we hope that in future we'll be able to share great examples from you about how you've used android vitals to be able to improve the performance of your games we know the technical performance is critical to business success what we've also heard is that sometimes it's hard to make the case for investing in technical performance within companies so I encourage you to use the benchmarks that we provide in the Play console to be make the case of where you and your team should be investing more to be able to compete more effectively I'd also encourage you to focus your attention on the core values if you have limited availability to get a focus on performance and make sure to sign up for the email alerts around anomalies and get the rest of your team that are focused on reacting to technical issues to sign up – we have a session later today about bug hunting on Android by damien who's going to be diving into specific detailed examples that developers have faced as to encourage you to attend that as well thanks very much everyone for joining my name is emily Pudsey and i am a BD manager on the Google Play games BD team and today I am here with my colleague Jonathan he'll come up to the stage in a couple of minutes to talk to you about some powerful tools that can help you more effectively market your new and existing games in the Play Store as we all know marketing is critical to the success of games at all stages of their life cycle before launch it's important to build excitement and awareness for your new games at launch you want to turn this awareness into installs and also make sure that your store listing is ready and optimized to attract as many prospective players as possible and of course after launch you want to continue acquiring new users and make sure that your store listing stays relevant as you make big updates to your game so today we're going to focus on a few tools that can help you better market your game in the Play Store in each of these three phases first up let's talk about pre-launch marketing and pre-registration today many businesses do pre-launch marketing to build hype and we're nasai had a big upcoming product launches there's a long history of this in tech entertainment and even in gaming particularly in the PC and console gaming worlds where pre-launch marketing and even things like pre-ordering are well established norms well today pre-launch marketing is also increasingly popular for mobile games however until recently it was actually really hard to do pre-launch marketing for mobile games well because there was no real way for developers to tie all the marketing that they were doing across channels directly to their game launch instead what we would tend to see developer do is spend a lot of money prior to launch to build hype and awareness you suboptimal tools that really only allowed you to capture a portion of the demand that you were creating through that marketing and mostly just hope that that buzz that you created would translate into real installs at launch well today as you heard a little bit earlier I'm very excited to announce that Google play's solution to this pre-launch marketing problem for mobile games a feature known as pre-registration is now available to all developers on Google Play for those of you who are not familiar place pre-registration feature is an extremely powerful tool that essentially allows developers to fully capture the demand that you generate through your pre-launch marketing activities and we actually activate that demand for you on your launch day so how does this technically work ad developers essentially tie all of their pre-launch marketing across channels to one page in the Google Play Store your pre-registration store listing and that listing has a very clear call to action for interested users where they can pre-register for your upcoming game and in return for pre-registering what users will get is actually a direct notification from the Google Play Store letting them know that your game is now available allowing them to be some of the first to install and play now pre-registration also has a lot of key benefits for developers at launch as well not only does pre-reg allow you to engage and reward your core fans early but pre-registration notifications also drive a high volume of D 1 and D 7 installs both in general and when compared to pre-registration tools that are available in the market in other words plays pre-registration tool is extremely effective at converting pre registered users into installs in fact across all geographies and all genres we tend to see that there's a 38% conversion rate from pre registered users to installs after launch now this is a much much higher conversion rate that you would see through other pre-registration sources around 2 to 3 X and it's pretty compelling plus place pre-registration campaigns are actually really easy to set up and manage and you can do that directly in the Play console so in the pre-registration Center on the console you can set up and manage your campaigns you can decide which countries you want to can run your campaigns in and when and you can actually view your campaign performance data so for instance before launch you can see the number of users who have signed up and after launch you can see how many of those users convert it into installs and you can do these breakouts across geographies another very cool feature related to pre-registration that is now available in the Play console is related to play instant earlier today from ben and others you heard a little bit about how effective instant games are in terms of driving user conversions well as of today we are very excited to also announce that now you can add an instant experience directly to your pre-registration store listing this will allow you to attract even more users to sign up for your full product launch and drive greater awareness and excitement another pre-registration feature that will be coming soon to the Play console is the ability for developers to grant rewards directly to pre registered users now this is really exciting because what we've seen is that rewards in general can help drive greater consumer interests in pre-registration campaigns so for instance listings that actually market pre-registration rewards can drive as much to 15% higher click-through rate than those who don't and on the backside when it comes to conversions to install pre-registration rewards can also help drive higher conversions because on day one there's value waiting in-game for pre registered users to come claim and consume so as I mentioned you can expect to more about this very very soon lastly I'm excited to share one more product update around pre-registration which is that Google Ads will not offer app campaign support for pre-registration campaigns obviously this is a great new way to reach prospective players but more to come on this in sessions this afternoon finally a few recommendations for developers who are interested in using pre-registration first allow time to build excitement within your audiences before your launch generally we see developers starting pre-registration campaigns anywhere from three to twelve weeks ahead of launch second make sure you're using all channels at your disposal to support your pre-registration campaigns this means social CRM media you name it third make sure you drive your Android traffic to pre-register in the Play Store due to the higher conversion rates we see in the Play Store versus external websites and finally consider using Play instant alongside of your pre-registration campaigns to drive greater consumer interest and awareness for your upcoming launches now I'd like to invite my colleague Jonathan up to the stage to talk to you about some other powerful marketing tools that can help you at launch and beyond thank you thanks Emily let's switch gears and talk about some exciting features around store listings where did your in pre-launch launch or post launch having an effective stall is thinking me critical to the success of your titles for many your store listing is the first thing a user will see about your game if it doesn't immediately grab their attention they're likely to move on in general we see that for game developers your icons videos and screenshots are particularly important on your store listing for user acquisition for instance game devs can see up to 40% difference in conversion performance when testing different screen share icons and data suggests that we see similar variability for videos and screenshots as well we've also seen that the ordering of your marketing assets can impact user acquisition because gamers really want to quickly get a sense of the true gameplay experience we really recommend that it's important to put marketing assets there highlight key gameplay features like your videos and screenshots up front in your store listing in the first position also quick PSA it's important to call out that to be eligible for the new visual promotional clusters on the Play Store you must have a landscape orientation video trailer and a minimum of free landscape screenshots so really make sure you have these and put them up front in your store listings to get the most out of your users now I will say that one of the most common pain points we hear from developers about store listings is that although they can already localize in two different languages they want the ability to customize their store listings based on where their users live for instance maybe you're making a sports game and you want to market it as football Europe by soccer in the US in order to make your store listing its locally relevant as possible well now you can today we're excited to announce a powerful new feature only available Google Play called country targeted store listings with this feature you can customize your store listings based on the user's country as well as their locale setting' and from the play console you can create up to five different groups of countries and serve a completely different story sting to each group complete with full localization support all parts of your store listing can be customized from your icon your screenshots your promo message your video you name it and you can also look lies everything there as well in order to better appeal to users in your target market even better because you can target different store listings to different countries technically you can target based on the state of your game in different countries as well so for instance and games were running actively pre-registration your listing might want to highlight the rewards users may get for pre registering or display a pre-launch teaser trailer and in countries where actively enrolling for open test your store listing can highlight key gameplay features in order to entice people to install and come part of your test this is a really powerful way to target your marketing message to the right sets of users and country targeted store listings and launch this launching right now to all developers on Google Play now that we've taught about importance of your listing through the lifecycle of your game I think it's also important to discuss one of the best ways to test and manage changes to your store listing in a controlled fashion store listing experiments store listing experiments has one of the most popular tools among top game developers is a tool that allows you to a be test your store listing assets against user traffic so that you can make the most data informed decisions about which of your assets are driving the most installs using play console you can swap out assets on your store listing such as your screenshots and video to certain percentages of users who visit your store listing and compare the performances of each variant before applying them to your full user base as mentioned before testing and optimizing certain assets can yield huge differences in performance running experiments also allows you to roll out these changes in a safe and controlled fashion for doors who've you are already using store experiments we're happy to announce that we've recently rolled out some exciting improvements to how experiment reports our results are reported first off we are introducing two new metrics which were replacing the existing metrics on the Play Store first time installers tracks users who see your listing in a store for the first time and emits reinstalls who are less likely to be influenced by your store listing day one retained users tracks users who install and do not uninstall within one day we believe both of these new metrics more accurately reflect a performance of your store listing and us developers can choose which ones it uses your success metrics in your experiments both metrics are also reported hourly intervals updated every few hours allowing you to see results faster and you can also now opt in to email notifications about when your experiments finish by reporting the results faster and revising multiplications you can run more experiments in terminate bad treatments faster to minimize the store drops all of this is coming right now to the Play console and is available to all developers ok so here are some tips and tricks about store listing experiments have a clear objective about what you want to test and test one thing the time to make sure the experiment results are are great before you apply them test for seven days a minimum in order to get weekday and weekend performance prioritize testing of icons videos and screenshots for the biggest impact and you can also invert your experiments to validate the true performance before applying to all ok so that about wraps up our talk about marketing tools and I want to give you some key takeaways from this session that you can do right now consider using pre-registration for your upcoming launches it's a great way to build interests prelaunch consider building an instant game to run alongside this campaign as well optimize your store listings using SL ease top developers already finding great success using this and finally look into how you can make your store listings more relevant to your users using country targeted store listings for instance using a store listing to run pre-reg and one for launched or test countries Thanks next I'd like to invite natural to stage who's gonna talk about lifetime value for your games hi everyone my name is Ignacio or naturals people call me I am based in London I work with game developers across the Western Europe region and this session is about lifetime value which is a dense topic and it only started for me about three years ago when very often I heard this term mention in my meetings with developers and in fact is a very popular very popular metric to the point that the latest report by merrymakers last year was mentioning this is the number one metric when it comes to artists spending so really really developers they care about it and if you take in consideration the context where we are on on one side we have the mobile gaming industry which is highly competitive and on the other hand we have all these data that we're sitting on and it's becoming more available and easier to handle so this is really leading to developers to to fill the need that they need to understand the value of their players from the very first minute they jump into the game and as such they're using lifetime value metric for many purposes so the first one is like marketing viability which we have spoke already so how far we can push those CPI's and how much we can invest in acquiring a user but it's also we we see a lot of developers thinking about live ops and running some of the offers running some of the promotions within the game as well we think even bolder we can think about product prioritization we can think about new games if we are looking to improve or LTV across the portfolio of games that we have and last but not least if we subtract some of the cost and we think about in net terms of LTV we can think about like the profit margins so as an example gang house is a developer based in Netherland and Spain and they're using lifetime value to do many of the things that we mentioned so they're doing user acquisition they're doing remarketing reengagement retargeting and they're even thinking about live ops for the new games doctors oath so in an ideal world our lifetime value metric would look like in this way so it would be predictive we were looking forward rather than just like historical data would be also it would be a metric that is extremely accurate because we need to take decisions with certain degree of confidence it would be also looking at individual level rather than cohorts and it would be fast from the very first day we will know what is the value that this player is going to bring in its lifetime as well the model that we're using it would be fast and simple meaning that we don't have to spend months in order to train this model but the first thing that you learn when you start researching about lifetime value is that there is no standard way of calculating it which means that these comes with some trade-offs and some challenges we ask clarium Israeli developer the creator of Vikings war of plans and they told us they faced three key challenges the first one is like how to estimate the lifetime value for certain type of players they call it outliers and these are players that they might be highly engaged but they still they are non players as well they have multiple games in their portfolio and it's very difficult for them to use one unique model that really is adjusting to every type of game and last part of this the long window that is needed to validate these models in fact is higher than expected but whenever we're talking about developers they see that these models they all of them they have several common things common areas in this case we could have split lifetime value in three areas and this would be the first one would we talk about what is like them that is generally represented in daily retention and this is a classic way but here many developers they're starting to think about whether the classic retention is like the one that it makes more sense for their own game and they can think about other ways of calculating retention such as for example range or rolling retention also think about retention it means right now opening the app but it's not the same if we're opening the game just to collect the daily reward or to play another bottle the second area would be the value and this is generally represented by purchases or by average averages such as R tau or RP pu and in this case if you are considering averages very important to take into account as well how long are you going backwards let's imagine that for example you run an offer last week and then last but not least is the period so how far are we going into the future you know that estimated lifetime volume and generally from my conversations with developers it all comes down to two things the nature of the game and as well how well-funded are we in terms of like acquiring users for like a period of 180 days or 365 days here we have an example that I just mentioned for example in a case we're taking the RP PU of a user and we are projecting for 180 days so other things that developers take into account when they refer to lifetime value so we have first of all like how do we integrate all the different business models that we are monetizing nowaday so how do we include like in-app purchases together with subscriptions and together with that as well if you are a developer that has multiple games in your portfolio and you are running some sort of in-house art and cross-promotions is the best practice that the some of developers that are thinking about the lifetime value across the whole portfolio last but not least we were talking about as well virality and k-factor effect so this is my not effect the lifetime value but indeed affects the return on adspend and some other areas that developers consider is as I said before metal TV also the frequency when we do these calculations and what are the segments that we are using if we are applying averages so I try to cluster which is very difficult I try to cluster all these models into three areas so the first one if we have a small data we have to take decisions we kind of liken see that these simple predictive models which are looking at averages and we are trying to estimate based on this for example bottoms-up approach or top dense approach we are trying to estimate the lifetime value of a user you have a number of tools available today in the play console and as well in some analytical's tools such as firebase where you can extract this information and start working around these models if you have more data and you have more mature games so you can actually also you have more resources to dedicate you can think about these predictive statistical models which generally they are coming from the offline world and they are backed by academia and you have a couple of examples here Pareto and MIDI BGM BD they good news is that they are open source and you can download the libraries in this case I put live Python library which is open for everyone and you can start playing around the good news is do will have individual predictive lifetime values already and it would be very helpful if you want to run live odds and then last but not least we have the predictive machine learning models way more sophisticated a bit more complex as well but they are extremely helpful if you are trying to understand what are the behaviors that are driving certain actions such as making purchases if you want to start like we have firebase predictions they're going to be a session later on today and you can think about like using these as first step into this type of models obviously if you have more resources you can use tensor flow or o2 ml to build your own algorithms based on random forests or neuronal networks so games house the same developer that I was mentioning before they use historical data combined with machine learning and they're able to estimate a lifetime value within a week for 90% with 90% accuracy so which one is the best this is the most trickiest a question I would say in this case also the answer would be depend it depends on your resources what you want to do with this metric and how much effort you want to put into it if you for example don't have too much data not too much resources some of the simple predictive models would be interesting if you for example have a very clear focus on user acquisition and you have the data and the resources some of these machine learning predictive models could be very helpful in order to understand the likelihood of a new player becoming a payer and then if you have a mature game with lots of data and highly engaged users you can think about the predictive statistical models which they run very well to to calculate live odds I want to spend the last two minutes talking about a case study space ape is a developer from the UK founded in 2012 and is the creator of popular games such as samurai siege and fast lane in this case is space ape they calculate the life time value based on historical data and using a statistical models they have data for about 1 to 6 months and after the day – they are able to estimate what is going to be the life time value with an 80 to 90 percent accuracy the way they do it is they create segments based on tenure based on geography and they multiply the revenues generated of that particular player by a ratio that they estimate based on their models so the LTV is mainly used for user acquisition purposes but also space ape use it for live ops the LTV metric is refresh on a daily basis and they have two interesting additional LTV metrics one it's called LTV Plus which essentially it's adding a small percentage mm to this metric and is used only for user acquisition purposes and this is to take into account the virality or the k-factor then the second one is the Adelle TV and what they're doing is like thanks to net we're like partnerships with networks they are trying to understand what it goes beyond the ad view and even like what it would be the conversion of a particular player when they see a water that for example so in essence for a space a lifetime value is a critical metric and they aim to actually return have a return on ad spend at least hundred percent in the first two months so with these I would like to conclude with four things so the first one is remember we are using lifetime value for user acquisition purposes but you can use it as well for other topics such as or other purposes such as live ops or in this case metal TV as well when possible try to use a predictive model and individual and as well consider other things such as how we're going to use this model how are you we are going to use this metric and if we have for example multiples up in our portfolio or how we're going to incorporate other business models then last but not least predicting the future is obviously is not a hundred percent accurate science so take these more as a tool and as a metric rather than just go

14 thoughts on “Keynote + Tools and best practices to improve game quality (GDC 2019)”

  1. @ 27:40
    He says that in dynamic updates, the user need not go play store… What if developer loads up the update with malicious code?… What about the security… Because the updates are not screened by google!

  2. Here's a simple one: Always make your game playable in both portrait and landscape. If I want to play Golf Clash in landscape, I'm forced to play it on Facebook (not fun)

  3. I have a domain, solid world concept, and I know it will be good. spent the last few years learning to code. about to dive into this…………. Are you saying that I could list an HTML5 game 🙂 I really want to develop the game with lit-html.

Leave a Comment