A Single Source of Truth for Measurement

right coming up now we've got an interesting conversation with um yael altool from appsflyer about a single source of truth for measurement something that is very very close to my heart i really love this type of um discussion so i'm really really excited to meet you yell and find out a little bit more about single sort of source of truth measurements how are you doing today it's really great to see you thank you very much it's uh great to see you too awesome well look if you've got any questions for yelp please don't hesitate to put them into the q a box below you can also talk about it hashtag pg connects you can go over onto um pocketgamer and we're here all afternoon so please keep your feedback please keep your fun coming in over to you yell i'll be back in a few minutes so hello everyone my name is el alcul

and i'm a product manager in apps file scan group and in today's session i want to walk you through some of the main challenges presented by ios 14 and how apps flyer created a single source of truth that really helped advertisers to maximize user value measurements so let's just start so the mobile marketing industry used to have a unique identifier which was the gold standard for user level matching between ad engagement installs and then in-app events advertisers have also had a real-time data that helped them to optimize and get decisions on the spot they also had unlimited ltv tracking and one source of truth where they could optimize campaigns and get decisions with the introduction of ios 14 and sk network in wwdc 2020 the mobile

marketing world was changed forever ios 14 gave users back control over their privacy require them to consent to share their idfa but on the other hand marketeers have now lost access to most of the idfas in their database so in order to fill that gap apple required marketeers to adopt s-cad network which is a deterministic solution that help to attribute all users regardless of att consent but this solution comes with the cost of user level granularity so now advertisers are facing two main challenges the first one is how to understand the actual picture of marketing performance and the second one is how to optimize their mid-spend and channel mix and in abstract we see those trends very clearly the adoption of ios 14.5 is almost 80

percent usually the shroud lift starts when apple starts pushing for the adduction and here we can see that the sharp lift starts a little bit in delay since apple delayed enforcement which also include att with high correlation to this one we see that we have a continuous decrease in the rate of idfa over time and if we will focus on gaming we can see that for hyper casual gaming the percentage are a little bit higher compared to other gaming categories by the way if you'll exclude the 30 of devices that are with older versions the numbers here will be much lower around 11 15 max so let's start just breaking down each challenge how can marketeers really

understand the actual picture of their marketing performance marketeer lives were switched from a unified reality to one that is silos across different dimensions we have sk networks we have apple searches we have the mmps but which one of them is true how we can really understand the actual picture of our marketing performance so one can say there is no problem let's just use the relevant model for the relevant use case so unfortunately we can't sql network by design is anonymous that means that we cannot really know for sure if a specific conversion was also attributed to a different model so we cannot really understand the actual picture since the same install can be attributed only to scan only to a different model to both of

them or to none of them and again apple by design developed sql network with features like random delay and privacy threshold in order to prevent from user level matching and that's what pushed mmps to to develop different apis different dashboards in order to represent those different realities and we all know that when we have different dashboards different apis different realities we cannot get the actual picture we cannot create a single source of truth on top of that apple by design limits advertisers ability to measure post install activity like retention conversion engagement and so on after talking with hundreds of our clients from different regions here and verticals one thing is very clear to us mastering

the conversion value mapping is something that is super challenging even for the most heaviest one first of all we have only one prospect with one conversion value that should represent any post install activity also all the work of the decoding and the encoding itself is something that is very challenging and it's also very time consuming requiring our marketers to invest more time in conversion value optimization rather on campaign optimization so how can marketeers now keep growing their business with all of those kinds rules and limitations that are a huge burden on them as the industry continues to acclimate to this new era it is critical that data accuracy and user privacy will continue to coexist and here at appsflyer it's our mission

to empower advertisers through these changes creating a true picture of marketing performance is something that is mission critical for us and a single source of truth will ensure that advertisers will be able to optimize their ad spend and grow their business so with that in mind i am more than excited to introduce you the new conversion studio the new conversion studio is one of its kind solutions super flexible that help advertisers to really maximize user value measurement in this new reality using this solution advertisers will spend less time on manual work and more time in optimization they will be able to focus on what they would like to measure without writing a single line of code since it's a fully automated fully sas control solution so without further ado let's just jump right into a demo

so in this new ui of the conversion studio the advertiser can choose what he would like to measure on the right side you can see that we have a full transparency for the conversion value capacity and mapping we will be able to see how many values were used and how many values we have left and also how each event is being mapped to a specific conversion value in the end we will also be able to download the csv file with all of this mapping and save it for later so until now advertisers were able to measure only one metric at the same time they had to choose between revenue engagement and conversion and if they wanted to measure more than one metric they had to code all the solutions themselves they think for example require them to

upload a new version each time so in this new solution we can measure all of those metrics at the same time for example funnel is the most efficient way to measure sequential event that means that one event must happen before the other so for example for if you want to measure level uh complete one two and three we will use the funnel in order to measure all the conversions and if for example level two happened we will report on both a and one and two and if level three happen we will report on both one two and three as you can see on the right side since the events are sequential we will use only three values on top of the install if you will use the the in-app event component in order to measure them separately you can see that i am using seven

additional values on top of the install and that is because we need to calculate all the different combinations because each event is independent so funnel is just one example of how we can utilize even better now all of the conversion values and when we save conversion values we can measure more with the same number of conversion values another example for measurement is revenue for revenue we can measure either overall revenue or in-app event revenue for in-app event revenue we can choose for example to measure a specific conversion meaning if the in-app event happened or didn't happen or to measure engagements using the ranges we can reach now up to one thousand one one two thousand how

how many engagements that we would like to so we are not limited to a 63 engagement or to 63 dollars anymore we can also use the revenue component in order to distinguish between ad monetization and in-app event revenue you can see now that i'm measuring the revenue that is being associated to ad monetization and also the revenue that is being associated to af purchase and if you will look at the right side you can see that i can know how each value is being used in order to represent my matrix for example the number three represent an ad revenue purchase event of

one dollar and an and an af purchase revenue event of one dollar and an activity window of 24. so this conversion mapping happens in real time and give us full transparency for our conversion values mapping another capacity another really cool capability that we have is the way to extend our post install activity measurement to up to 72 hours post install assuming that the user is still acting because if apple timer stopped because the user was not active we cannot update the conversion value anymore so now you can measure up to 72 hours post install and you can split this window into 12 or 24 hours interval the intervals help us first of all uh to calculate the install time in a more precise way because we can know

when the user was last active for example if we will indicate that the user was last active between 36 to 40 hours post install we can use this indication in order to calculate the install time in a more precise way we can also use this indication in order to build kind of a retention report so as you can see in like two three seconds i can configure my conversion value mapping in an easy ui and then apps flyer will do all the encoding and all the decoding and all that hard work uh providing you reports dashboards apis with all of the analysis so let's continue if you will summarize the conversion value capability in order to answer the market needs for better um user value measurement we provided the conversion studio

and the conversion studio solution is first of all a solution that provides us an ability to measure more than one metric at the same time the measurement itself is super flexible we can measure any event that we would like to in any range we can extend the post install activity window to up to up to 72 hours and we can really fine tune our strategy now so we talked about the metrics that we would like to measure but how can we now understand the actual picture of our marketing performance in the conversion studio we provide an option to allocate one bit in order to indicate if a specific install was also attributed to apps flyer or not then in the aggregation itself we will take all of app's flyer attributed users and a tweet all of scan prospects for which the found attribution flag is

false that way we will enjoy the advantages of both words and overcome some of the advantages of each model we will be able to provide a single source of truth where each user is being represented only once so using the conversion studio we will be able to allocate one bit we will be able to indicate uh if a user is a duplicate user or not we will be able to create a single source of truth and then advertisers will be able to really get actionable marketing in fact but wait don't we harm the end user privacy absolutely not all of s-cad network protection layers are still there in order to make sure that the end user privacy is not being harmed and that the user is still anonymous

the only thing that this solution is doing is aggregate those two different models scan and the mmp in order to create one reality where each user is being represented only once by the way we will send all the postbacks to the network regardless if we found attribution or or didn't find attribution that way the networks can keep optimize the conversion value of sk network so the conversion studio is just one piece of the puzzle after working and investing more than 50 and the last 15 months in educating ourselves our client and the entire industry about ios 14 we focused our efforts in order to develop a cutting edge fully privacy compliant product that can really help advertisers

in this new reality and as you can see the fact that we have only six bits does not mean that we cannot really innovate just the opposite another example for innovation is our predict sk products the predict sk product leverage early post-install engagement measurements in order to provide long-term predictive insights for campaign performance so we can use those six bits in order to predict the user behavior later on rather rather on um measuring the actual performance in the first 72 hours and this is another way to how we can use conversion value in order to innovate and understand user behavior in this new reality so we are just one percent done we have a lot of things in our roadmap

so with that in mind i encourage you to contact me and the team ask questions and we will be more than happy to help you on board thank you i've just muted myself thank god thank you so much for that um we've got some questions which is uh pretty cool because i thought that was a really good presentation to give us a little bit of an idea about what that landscape's like after the ifda changes so with that in mind have you noticed any differences in the market since then yell since the idfa since the idfa changed so sure first of all um what we can see now is that the rate of idfa among ios app is in a continuous decrease we have less consented users and this and since that the id matching in the attribution model uh also decreased over time

that is why uh we need to use both scan solution that give us a deterministic attribution and the mmp solution that give us a better ltv measurement more precise lte measurement for consented users that's interesting one of the things that i think is interesting about this source of truth particularly is that it provides really good single solutions oftentimes when we do publishing and self-publishing especially we have to go through various different funnels and changes and do stuff when actually this is a really nice just simple dashboard approach i know you've already done sort of one percent of the hard work but you've got to do but there's still a lot here to be able to get your teeth into do you think that these are good solutions for self-publishers or because they're still using independent app stores so they get their analytics from

different places they want to do their different conversions and points and funnels at different kind of places maybe than live ops and publishers do what do you think how does it work for self publishers is it a useful tool so first of all using um appsfire which is an mmp um help you to get a p a full picture from an and by a third party instead of uh going to each uh partner alone and get seeing the picture there you can see it one time in apps fire uh in an unbiased way the second thing is that you can configure your conversion value mapping only one time and then the mmp is doing all the work for you we are doing all the encoding and all the decoding and whole all the hard work without the need of um publishing a new upper version each time

and and we are also working on other products like like predict and protect and other uh products that are that are part of the of this um a p360 uh innovation that can help you um use each part of the mo of the model and anytime you like to test something new without the need to develop it yourself what's the way to get into sort of apps flyer do you have pilot programs or is there some sort of way that people can connect with you on more of a tentative level um do you offer trials can people try what you do apps fire so for sure you can contact me and the team and we will be more than happy to demo you the entire uh product offer you some um a customized solution that will help you

to really fine tune your business strategy and then you can you we have different trial programs that you can sign in and implement our sdk uh user use the tool see if it fits for you because it's that independent thing that you were talking about that made my ears prick up and i sort of thought to myself wow this could be really cool for a lot of people that are self-publishing that don't want to go from sort of dashboard to dashboard to try and figure stuff out so that's a really good opportunity for people to be able to connect with you yell once you've supplied this data how do we apply this to make improvements in the user engagement and you're talking how about how do we reflect the data itself yeah so we have a variety of data outputs each data outputs can fit another different vertical different tier different type of company we have

apis we have aggregate reports and csvs and in json and we have also dashboards that you can use so it depends if if you have a bi team or you don't have a bi team if you want us to do the aggregation for you or if you want to use the raw data yourself there's a really really good question that's just popped up here if we've got apps flyer in your tool do we even need publishers like at all we can just become as effective as they are without you know needing to go there surely that's the whole purpose of what it is that you created here right we are working together with our partners in order to um have a good user value measurement in this new ecosystem so it's a a combined work of apps flyer with the different partners in order to help

advertisers to really understand their marketing performance that's really really good and following that idfa question somebody's asking are there any worries that the google store might follow soon er we can we don't know what are the plans of google and other srms but one thing that i can promise for sure that like we handled the new ios 14 reality and the new sk network reality with a lot of um new products and new innovation and we will handle any uh hiccup along the way oh god i really need some help with hiccups today what do you think is on the landscape in the future vis-a-vis your tool against you know some of these big technology companies what do you see coming and where do you think you fit within that landscape as apps flyer so i think that in general the entire

industry now is facing a lot of new challenges in terms of the new way for attribution yeah and we will we will meet together to adjust to this new reality and we see that id matching is something that we cannot rely on for 100 of our users and we will need to overcome um things like retargeting audience creation and and their attribution interesting really really interesting thank you so much everybody knows how to get in touch with you they can even reach out to you on discord we have a discord channel that's running throughout this week and people if they've missed any call what is that yeah that's presented today excellently can reach out and go and talk to her in that channel thank you so much for joining us today it's really good to

meet you thank you very much it's a really valuable discussion about the future of publishing so thank you very much indeed thank you thank you very much have a nice day