Six Steps Your Competitors are Taking to Grow Their Games | PGC Digital #1

hi guys thanks Chris this is an awesome opportunity and it's really exciting to be part of it and being the first one to speak no pressure here so I Sandra I am VP partnerships at iron source which means that I'm exposed to a lot of the industry trends that are happening and you know new things that our clients are doing and so I called my talk six six steps your competitors are taking to grow their games um let me start with with iron source so we are a net network for anyone of you who does not know that we also our mediation provider and last week we had some pretty cool news that that I do want to share obviously because we're pretty pretty pumped about that after I ranked us to be the top one network for for casual games and the top three for all beings globally and that is some pretty big news that we were

really excited to hear so I wanted to share that here as well so let's get into the talk um so what I did was I split it into one part that is going to be about monetization and then the second one is going to be about user position so we're gonna speak about benchmarking your ad implementation then we're going to move to automating your waterfall and we're gonna finish the monetization part with segmenting on the user level and then when we move to user position I want to talk a little bit about bidding granularly um how to automate your user acquisition and then I'm gonna finish on cross promoting your games so monetization benchmark your ad implementation when you look at the ad units that you have available you have system initiated X right which are ads that the user doesn't show it doesn't choose to to see interstitials and banners and those improve your art pools so they make you money but they might you know risk retention so that's

the two things that you need to balance against each other on the other side you see the user initiated eight units which is recorded video and off a wall and those also improve your article right they make you money in addition to that they also improve your attention and your so what you do want to mitigate here is the risk of cannibalizing your in-app purchase revenues and we're going to talk about this a little bit more so now let's say you started your game you chose the ad units that you wanted to run and now the point that I'm trying to make is how important it is to make sure that the engagement health your users with Ed's is maximized because what that means is that you're maximizing the available impressions in your game and so what I wanted to do is look at the numbers that your competitors are generating and compare yourself to those guys and see how they implemented ads and all you did and maybe you can change accordingly so this is really important to do and as you do

that you keep measuring the engagement of your users with ads right so the metrics that you would use to look at this is the engagement rate which is the amount of users out of your daily active users that push the recorded video button and the usage rate means how often they do that in a day so this is really important to to look at and any solution that you're using today you do want to make sure that you are you have a way of measuring this more importantly and this is something that we don't speak about so often but I think it's really important is to measure those metrics also across the lifetime of the game so meaning that the user as they move through your game they reach different levels right so let's say now they're in level 20 and watching a video gives them a coin but they need 20 because the economy inflates as they go through the game so you would see a drop and engagement rate over time and this is really important for you to know and

understand because the fix is very easy right you just make sure that the reward increases over the lifetime in the game and you fix that so um so these are the two things to to really pay attention here let's move to the waterfall side of things so right we just looked at the maximizing impressions part right which is about the amount of rewarded video and interstitial that you implementing and now we're gonna make sure that we sell those impressions at the highest price possible which comes down to the waterfall right so what do you want to do here is he want to automate as much as possible and we'll talk a little bit more about that right now so it needs to be in a smart way right because the way that the big systems today the big mediations today they all have bidding right doesn't matter who you work with but you want to have a dashboard that allows you a quick overview and allows you maximum flexibility in how you want to manage

your menu waterful versus your bidders right because you will still have manual networks because they're not all networks or bidding and all mediations and as time goes by we also see that you know for business reasons some of them and some because technically they're just not there yet so that that's gonna be a slow transition so this hybrid model is here to stay for for a while right so you want to use the dashboard that allows you full flexibility to actually actually do that so this is a little bit of the best practices that we see from the from the Triple A studios that we work with right so implement in a bidding that's like the main thing that I wanted to take away from this implementing in a bidding will do two things and I'm gonna start with the one that we actually talk about all the time and that's improving the performance right and here I want to talk to you a little bit about expectations that you

have from the interest that you get from adding in a bidding because the more optimized your manual waterful was before you turn on the inner bidding that you know you will still see an app lift but it's not gonna be thirty percent because he already had kind of a semi bidding environment before you started and and actually I had a conversation with someone from a really big big network that is a bitter that confront that as well so there is like if you already ran with a wonderful that is really optimized the increase will be there but it's not going to be you know mine throwing the big and the less implement the less optimize your wonderful is the more uplift your generate and the second thing and that's the thing that we don't really speak about as much as we should in my opinion is of course the saving you know in resources and manual optimization that you need to do right if you have a person who looks at the waterfall every day or every second day now as they are suddenly you know available for other things so that's awesome so this is

about benefiting regarding the non bidding networks that again they will still be there because we're talking about this hybrid model for a while you want to make sure that you create a semi bidding environment this is how I called it through just adding a lot of instances or line items for those non bidding networks at different price points to kind of make sure you simulate like a bidding Network a bidding environment right and at the end of the day you want to optimize this based on a CPM and impressions to make sure that you're accounting you know for the fill rate as you go cool so now we're gonna move to segmenting on the user level and here before we dive into the segment's first let's let's talk about why we want a segment so obviously you know every user has different nets that are served to them different and engagement some actually push the video button a lot some don't they all get it differently

CPM right with the bidding coming like every user is actually gonna get a different price which is mind-blowing right they also engage within a purchase in a different way some pay some don't so there are so many things to account for and that's why segmentation is actually something that we see more and more of the big guys going after but before segmentation what I want you to do is measure the full value of each of your users and what I mean by that is that you already know obviously the inter purchased value that your users had right but what you don't know or you know so you might not know what's the end revenue that each of your users bring to you and that's something that is so valuable in terms of the data so when you use a mediation and a lot of the big guys out there support this already you will get the data from all events that you're running in your mediation and it gives you the full flexibility of the value of the user I also really believe in this like

we were actually the first ones to to develop this feature and like a year ago and then everyone followed so we're we really think this this makes a lot of sense and right now we are also working on allowing an impression level breakdown for the revenue data so not just per user but also per each impression of each user so that's gonna add an additional granularity that you don't have today just to make sure everyone understands how this is done how do how do we measure ad revenue or how it is measured um so let's say user a comes in to your insurer game yesterday and that's the impressions that he saw Facebook $80 AdMob $70 and iron stores $50 so out of this very simple math right I can tell you that they made you yesterday this user made you $20 almost $20 which is a really you know an important thing that you didn't know until today now how do you break this into segments right so segments first of all it start with your paying users which you kind of want to

you know fence fence off a little bit those guys you probably don't want to show them ads or if you want to show them ads you want to show them rewarded videos maybe that don't compete with your in-app purchase you know currency and that's how you prevent cannibalization so that's kind of done then the second group and this is where it gets interesting when you look at the user ID revenue is when you want to also make sure that your ad whales so guys that watch a lot of compressions and get very high CPMs by the networks are staying engaged with this egg unit so you want to show them premium recorded video content you want to increase the rewards and really make sure they stay engaged the group c is engaged users the guys that push the button you know they get like decent eCPMs and you want them to keep doing that obviously and then group D and this is something that we also see in the industry in the last like six months I would say Triple A studios are starting to think about what do I do with the users who are not payers and never click on a button

forward video or a fool so you have system initiated editing that's right latest issues mostly and maybe even banners and you can consider showing those stripes there's nothing wrong with that like you can be more aggressive less apparent so if you can do it after they are at the game for 30 days and didn't make you any money or seven days like right but this is really important as a concept to think about and in general segmentation you can take this really fun right you can be really granular here and we see that a lot of the guys are making sure that they kind of tailor the ed experience to the users and their value and the setting and app part is where usually you know get a little bit of like you know confusion but I want to show you like this is this is iron sources therefore what you do is you just kind of set the groups that you defined you define the behavior that you want to have for each of the groups in terms of the edge right which II D which

reward how many and so on and so forth and then as the user comes in the game in the SDK gets imitates you just pass us the value we've run this and that's all you do so it's really simple also in terms of the setup right cool let's move to the user acquisition site and here I'm going to start with bidding on the most current in the level so there are two parts to this the first part brings me back to the user ID revenue that we just spoke to right because you can use that value on the monetization side but you can also and maybe that's even more interesting use it on the user acquisition site because when you look at the value of the user that you purchased from it from a source you want to make sure that you understand the full value that he made you from in a purchase and in revenue and that changes your ltd curves and it's something to really pay attention to because what it means at the end of the day is that you can probably pay a little bit more and get more scale for a bunch of users that

if you don't look at that driven you you don't even know about this so this is really important to keep in mind and the second part in the bidding granular section is to bid per source right and I know a lot of the guys you know are doing this already but not and it's you know it doesn't make sense not to do it and I want to show you why um so this is a simple comparison of not bidding per source to bidding per source and let's say I have two sources a and B I my boss gave me a margin goal of 30% and my LTV averages at 50 cents right so now it means that I can bid 35 cents that's you know that gives me a scale of 10 K it's pretty pretty basic right now if you look at this per source and you break it down per source what you see is suddenly wow maybe the average is 50 cents but source a only makes me 30 cents but I'm bidding 35 I'm overpaid shouldn't be doing that doesn't make any sense right so you want to reduce the bit here and what you see and source B and that's even more interesting is that you are you could be

paying more and what happens when you pay more is again you get more installs and that's a really good source for you so what happens when you do this is you're increasing your LTV you're increasing your installs and your margin stays the same so it's it's a pretty much no brainer and in the cotton the concept level right but when you think about it you have thousands of sources so how do you do that right how do you do each and every source in each and every country there are so many breakdowns so you want to automate and this is number fence automate user acquisition and not just the bidding per source but also the fact that your game performance can change you know you're adding a feature and something changes or the source quality changes right like let's say the top source that you are buying from today changes their user acquisition strategy boom your your numbers don't make sense anymore you also have new sources that enter the market all the time and you want to account for that so what you want is you

want to use a Ross optimizer which basically it's a lot of networks offer this so doesn't matter who you work with but make sure that you're automating this because basically what you do is you give the system a Ross call and say my goal is 30% and then it's gonna change the bids the CPI is per each source that you have and that's thousands of sources and in addition to that it's also doing it based on historical data so it looks at in a purchase and ad revenue because you need both to make a right decision and then it's just gonna adjust the bids as it goes this gets more and more granular right like our source is adding a lot more breakdowns as we go if I so as system you can be really really really granular here and it just adds to the to the quality so we have a just a small case study done by rescue cut ITI where they edit you know they started running the optimizer and they saw sixteen percent increase in installs and twenty percent more rows so

that was a pretty nice success I'm gonna finish on number six which is cross promote Irving's and obviously this is relevant for the guys here that have more than one game right and what we see in the industry is that more and more Triple A studios and even medium guys are starting to think about this cross promotion you know idea because the solution has been out for a while right the market solutions today are CPM based usually which is kind of like dumb systems basically you give it a bit it runs it takes impression but you don't really have any machine learning behind it and so um what we built was a dedicated network that is kind of like your own cross promotion network that you manage the UA side and you also manage the monetization side so you decide which positions in the waterfall

at which prices you are ready to give up to other networks for your cross promotion and our algorithm sits behind that and will predict if this is valuable for you it's gonna take the impression and if it's not it's not a pass on it and what this allows you is first of all for transparency so you can see both sides and you can actually see what works and what doesn't and it allows you also to run this unlike with an economic idea behind it because right now again CPM solutions are not not smart at all right this also is very I'm gonna come as a bitter so you don't even need to do the manual setup of the instances in the waterfall and it's gonna save you a lot more time this is a case study by Emma notes medic Charles 3 where we started running running the cross-promotion product and look at the blue bars right that's the install so basically once this started the installs

really scaled and the orange line is the CPI so they got more installs at a lower price so it just made a lot of sense for them to to keep doing this cool if I bring everything that we just said back into this one side benchmark your IDE implementation automate your waterfall that's going to make sure that you have the maximum amount of impressions and you're going to sell it at the maximum price possible segment on the user level make sure you target the ED experience to the exact user then on the you a site bid granularly and automate your user acquisition and if you have more than one game try to find a solution that allows you to cross-promote your game in a smart way the three takeaways that I want you to to say out of this are working with a partner who can close the loop is key and that's because a network that allows you to do monetization and user position just has a 360 look on the industry they can also be very strategic

about you know working with you there are some like deals that you could think about it it just makes a lot of sense to choose partners that allow you to do that the second one is data and segmenting are critical to making strong decisions I think we're talking about data-driven decisions all the time but data can be very daunting right like it's a lot of like bits and bytes and you need to make sure that you draw the right conclusions for you and I think that you need a partner who allows you to kind of like distill this down into what makes sense for you to do and then just try and do it and the third one is that you want to have a partner that has a cutting edge technology to make sure that your allows for maximum automation because I think as we move into bidding and as we move into automating user position you will see that you have um more time of like your staff to do the things that actually you know make sense for them to do because you're automating the things that they don't really need

to be doing manually thank you I thank you for listening I'll take a few questions now wow that was incredible right um I always try to make notes as you're going to thanks I said like I always watch League on flight 20 minutes not saying very much and you've kind of gone through an entire complex side platformer loads of complex attempts we do have a lot of we do have a lot of questions and we've got play time sets right it's very difficult I know I was going 100 miles an hour because I was trying to I realized the time I thought oh god I've got halfway so I still is getting three but yeah gonna be good so thanks about so I got a lot of questions here's a lot from mr. side praneeth good luck to you as well um so I'm gonna try to ask a few of these and if it's something you can answer them answer it's not we can pass on to the next one so I've got a question here saying how can I track use a level ad revenue with Tenjin attribution is that I don't I don't know that's the thing that that makes sense to you or not is not as far

as I know at least with our users revenue they did not integrate it yet so what you can do though is you can pull the ad revenue from an API from your mediation provider right so it doesn't matter who you work with like we have an API that you just pull the data from and then you just need to integrate the inner purchase data that you get from tianjin we see user ad revenue and you just need to do it in your own bi-state but it's not available intention okay good good answer that's good okay another question is can you share how to serve mediation waterfall and the best practices for admin monetization is saying for indie devs launched from your game especially considering the Kovac 19 situation now seem like a lot of questions that's gone through those so basically setting up a waterfall again like indie guys right they don't have a lot of resources so what you want to do is you want to try to automate as much as possible so the bidding

like the way to though try to turn on bidding and end as many networks as possible that are available with the solution that you're using in the bidders and then the networks that are not bidders just make sure that you have like enough kind of line items or instances and all kinds of price points basically right what you want to do is you want to add let's say I don't know five ten instances per each of the non bidders in your menu or waterfall and then just kind of start running like this and adjust as you go in the beginning and then at some point you should be kind of good to just let it run the situation I guess maybe this goes in the direction of is eCPM RDBMS changing right now is like a lot more people playing right and we don't have very conclusive like things to say right now it's still kind of its p.m. for now are going out but we don't expect this to to last so it's it's I don't want to answer on that right now because I don't want to make fun yeah it's very early nobody really knows I'm fascinated if

anybody's sharing date at this point because thing first things anecdotally reported but what that means in terms of revenues is very unclear I've got another question here from Corey Oaks Forres which is how do you measure LTV to split in source bid if LTV means you need days weeks to get data does that make sense again it's alive in my head I feel like letting something just video back a few times to come again it's a good question you you do see you know a trend and out to the after like day seven days three depending on the game a little bit right but you are kind of sure you just okay I just lost my screen but I think that's fine alright sorry I just I just answer the questions yeah so so I think um you you you do you can make predictions and you can build your LTP curve as you go

you're right that if your game is like you know a really you know highly engaged thing where you're gonna keep players for like up to half a year longer it a little bit harder to do but you can predict LTB and and there are ways to do it so I don't want to expand on that too much right now cuz it's very technical ways for me I'm sure the people I'm sure you can connect with with Sandra now source team I'll get two more questions and we'll maybe one loss of God against the next week this is great I've got a question from Peter Kwang I don't know again the answer to any of these was I wouldn't ask you I still can use the confirmation feature is it ready question mark I don't know it's a bad question to ask but when you use when you use answers mediation which you do because it's kind of network inside the mediation right and you ran use our position with us it's basically just kick off a button like it should be pretty easy to do so

so maybe just connect with me and we can try to figure it out exactly so you can connect with standard Burmese mask I've one last question because sigh I'm afraid you've asked about 50 questions you probably more answered you should meet with with with Sandra team Athens so sound a great talk this early in the morning we're getting a lot of good feedback actually so well done first one um they said looking nice about mine but yours every low so that's good um okay can you explain a little more about segmenting on the usual user level what would you advise as a basic starting point okay so what I did was that with it four groups this is how I would advise you to start thinking about it right I kind of look at your game and try to figure out like who falls in which broken at what point in time can you predict if they are in that group or not right and then just start at that basic level with those four four groups and from there just you know you can go more granular so so I think it's about just

kind of starting to define it in a very broad way and then set it up and as you go you can you can add more okay fantastic and that is just about we've got time for this Noreen I know you I think you're back on later yeah you're gonna do okay fantastic all right well I'm gonna I'm gonna let my my co-host we'll send you back into attendees you can feel free to keep watching but a round of calls for me I know you don't else if everything gives love in there as a hands up if you thought was great talker finger we're still working out from these tools but thank you so much Sandra