In-app advertising is a growing source of revenue for app companies, with estimates predicting a 60% increase in the sheer number of apps monetizing through in-app advertising this year alone. This form of advertising, used most by apps in the Gaming vertical, allows marketers to get the most out of non-payers, generating revenue for impressions, clicks and installs, and other activities when users view and engage with their ads inside apps.
But to net the so-called ad whales, you have to know who they are. To help you in the hunt we partnered with ironSource, an in-app advertising and mediation platform, to launch User-Level Ad Revenue earlier this year. In this Adjust webinar, Josh Jones, Product Education Specialist at Adjust joins with with Yevgeny Peres, VP of Growth at ironSource, for a deep-dive into what User-Level Ad Revenue is and how it impacts the ad spend measurement status quo. They also detail the benefits and new capabilities it unlocks for mobile marketers. Read on for our key highlights.
The blind spots before User-Level Ad Revenue
When it comes to measuring ad revenue, marketers were previously limited to tracking performance in an aggregated way on the dashboard or through APIs—how much revenue was generated over a certain time period at a tracker, network or creative level. As Josh explains, you were blind as to the precise revenue specific users generated, which channels they came from and other detailed behavior information. The revenue figures were limited to in-app purchases (IAPs), which, in the case of gaming apps in particular, meant you were missing out on an important piece of the puzzle—knowing the average revenue per user (ARPU).
This state of play made it difficult for advertisers to make strategic decisions from a UA and monetization perspective, because they were missing a complete picture of a user’s ARPU and lifetime value (LTV). This partially-obscured view meant advertisers were limited to optimizing UA campaigns based off this high-level view of revenue figures and in-app purchases, without access to insights about which users drove the most ROI.
How User-Level Ad Revenue works
As a mediation platform, ironSource mediates between the user and channels bidding different values for each ad. This role as mediator, Yevgeny explains, gives the company unique access to valuable, granular information about how users engage with different ads, ad placements and ad products—with a bottom line of how much revenue a specific engagement generated. When we combine this information with user acquisition data from Adjust you have a full picture of the value of users like never before.
Josh and Yevgeny take a look at how this works in real time, starting with a phone, an app, and ironSource’s SDK within the app. When the user opens the app and an ad is requested, the mediation platform will serve a network’s ad which best fits the publisher’s criteria, the ad is shown in the app, an ad impression is generated. IronSource records the advertising ID of the device, associated revenue, ad placement information and other relevant details. Thanks to our integration with ironSource’s API, we can provide this data to advertisers in real-time, in addition to the important statistics we already provide, such as attribution data, in-app purchases and KPIs for multiple marketing channels. Armed with this data, advertisers can compare the total income from a device to the cost of acquiring a user and thus have a clear ROAS.
Unlocking full measurement capabilities
With User-Level Ad Revenue, Josh explains, advertisers can “confidently retarget their valued users, invest in new campaigns, gauge user journey and cross promote their other apps.” Knowing this granular level of detail “ensures that advertisers can spend on ad campaigns that generate a positive return on investment...so you can optimize and deliver the best possible experience to the most users.”
Yevgeny points out that these new insights into user behavior can arm your UA team with the information they need to “decide on bidding strategy and take action based on this full view user level measurement.” Your UA team can then “take action using automatic bid optimizers to meet those return on ad spend goals.” This can have a positive impact on your team on an operational level as it frees up time that was previously spent on the manual bidding process.
Yevgeny goes on to give some examples of how User-Level Ad Revenue “unlocks that capability of accelerating that growth loop through a full measurement.” One of these capabilities is optimizing through segmentation, because you can “actually understand how the different segments behave, based on where they came from.” He looks at an example of organic users; if you can see that organic users didn’t engage well with banners—through segmentation and deep linking—you can stop showing these banners to these particular users.
Another new capability is seeing the correlation between ad revenue and in-app purchase per user, by testing a rewarded video ad placement and monitoring the impact it has on your in-app purchases—if there’s a positive impact on revenue after this placement or not.
To learn more about User-Level Ad Revenue, watch the webinar or read the full transcript below. You can also check out our mobile marketing glossary for definitions of terms used in the webinar.
User-Level Ad Revenue is an additional feature available as a separate package. To find out more and learn how it can benefit you, get in touch with the Adjust Sales Team.
The full transcript
Hello, and welcome everyone. We’re very happy to host today’s Adjust webinar in partnership with ironSource. I'm Josh Jones, Product Education Specialist with Adjust. We’re going to be looking into user level ad revenue and how you can benefit from this granularity of measurements as a mobile marketer.
Joining me today is Yevgeny Peres, VP of Growth at ironSource. Yevgeny, would you mind introducing yourself.
Hi guys. Welcome to this important webinar. My name is Yevgeny, the VP of Growth at ironSource, spending most of my time with ad developers on growth strategies, which tools, processes and products we can help to fuel their growth. I'm very happy to talk about this in the context of this product in a minute. How it affects our product strategy and how it's implemented within Adjust.
So we’re going to dive into a broad definition of what is exactly a user level ad revenue, the current status of measurement for the purpose of returning that spend optimisation. The problem this specific product solves and obviously the benefits of measuring that data.
Josh is going to cover how mobile marketers can get started on the Adjust platform and then we’re going to summarise with which new capabilities are unlocked and key learnings from this product so far. And we’re going to finish with some of the company overviews.
So what is exactly user level ad revenue? Essentially, we’re talking here about a product that is giving app developers the measurement capability to accurately understand how much each user generates when it comes to ad engagements. And then, in order to make this actionable, the challenge which is supported by Adjust in terms of solution is how do you make that data actionable when it comes to understanding where did these users come from? Which channels, which campaigns, which creatives, to make this part of your growth strategy.
One thing that I think is critical for this webinar is to understand where exactly this product fits within your app and growth strategy. When you think about your app, once it's ready, you’ve built it, you have your product in place and you're trying to ramp it up and start growing things, one of the first things you're starting with is acquiring users, to have people engage with the products, very basically, bring people from the market inside your app, either through pages or acquisition, or you have organic users step into the app.
The second step when you think about this loop would be, how do you monetise those users? So whatever strategy you're using, when it comes to the monetisation, what are its ad monetisation or IAP. IAP, that’s where people engage with the product and they move on, hopefully after generating revenue.
The next step would be, after seeing how people engage within your app, is measuring their engagement and then analysing what exactly can be done in order to optimise monetisation for the purpose of maximising LTV. Then to close that loop, and become smart and acquire more users profitably tomorrow, you need to analyse all this stuff together and understand where exactly where those users came from, how did they interact with the app, how did they monetise, what can I learn better on my UA and retargeting efforts in order to maximum profits.
Where you have the operation split into optimising and maximising LTV within the app and then maximising profit within your growth operation.
When you think about how ad revenue is generated, for each user that comes in into the app and is engaging with ads, whether it’s banners, rewarded videos, video interstitials, labels, off-walls. Whatever that user is engaging with there's a platform behind the scenes that is maximising the revenue for that specific impression. And that is handled through our mediation platform where each impression served would be different per user based on how the different demand channels are trying to bid on that user and compete for that impression.
Today the market is still transitioning from the waterfall era to the programmatic and that’s something that I’ll touch on a bit later. But each user has its own journey when it comes to ad engagement and which ads exactly are shown, how much revenue that generates, and that’s different per user, just like in-app purchase. Where each user either purchases or not and each user engages with that in a different way.
When it comes to ad revenue, performance has been reported over dashboards and APIs historically, but only in an aggregated way. A way you could only know how much revenue was generated on a certain day for certain breakdowns, not giving you any insights into which specific users generated those revenues, how much each user contributed to that and which channels did those users come from.
In this slide we can see an example of anonymised ad revenue data for a mutual client of Adjust and ironSource. In this example it’s the month of May where $300,000 were generated for two apps, one Android and one IOS app. And it's very difficult to understand what exactly does this mean when trying to understand what my next step should be when you think about optimisation.
This is basically what this product is about. It's about measuring the ad revenue generated by each user, so you can take the right actions within your growth strategy.
When it comes to the current state of measurement for the purpose of a return on spend. So the ad spend is covered when it comes to cost. But when you think about return, there's a big gap in the market that is around ads. So, specifically for ad based apps when you think about user acquisition, because there was no way, until a year ago when we launched this product, to understand when exactly that ad engagement occurred for each user and how much revenue did that generate, it comes to the point where people had to be optimising acquiring users based on retention. Assuming that a certain retention will generate a certain amount of revenue. ROAS optimisation was not really possible for ad based apps.
So what people were doing was bidding based on average LTV. So assuming a certain LTV in a certain market and basically bidding the same CPIs across all channels per market, which was crazy when you think of it. The most common approach was the Peanut Butter Approach, where there was an assumption in place of, if I'm generating 80% of my (recording scrambled 0:07:51) from IAP and 20% from ads, then I should apply a 25% Peanut Butter on top of all of my ROAS IAP only strategy.
A lot of people just gave up on ad revenue and just treated it as a nice bonus. And a few companies tried to hack it themselves within their own BI stack, requiring some basic methodologies to assume how much revenue each user generated.
So a very big gap between IAP versus ads. The best analogy that I'm easily thinking of is, think about not knowing how much each user generates from in-app purchases and the only thing you can rely on is your IQs Connector Cab or your Google Developers Console, seeing your overall revenue from yesterday or last month and trying to make it actionable. Which is crazy.
So most of the data was aggregated when it comes to ads, and how exactly is that transition to the actionable level of how much each user generates, that would be the challenge.
Making it actionable is impossible without connecting it to the attribution data. And when you think about a platform like Adjust that supports the optimisation decisions and measurement decisions, that data needs to be connected in order to be actionable. And we’re going to cover that a bit later. Attributing each user to where it came from and understanding how much revenue is attributed to that user, gives you that complete picture of true ROAS.
The way it looked like until recently, so within the app you would have your attribution SDK, in this case we’re talking about the Adjust perspective, so there's a lot of user engagements that are passed back to the attribution platform, for in-app event, (recording scrambled 0:09:48) any events, any level completes, (inaudible 0:09:50) registrations, the sessions and essentially the revenue would be passed through the in-app purchase events. And then you can easily funnel this back to your BI stack or just use the dashboard for understanding your data.
And then those measurement insights are funnelled through the UA team or they can assume that data and decide how to optimise on all the different user acquisition platforms.
The monetisation, when you think about it, the ad monetisation is covered by a different entity within the same app. So there is a different entity that knows exactly which user engages with which impression, from which channel and how much revenue that engagement generates. But that, historically, was just aggregated in a similar report that I showed a couple of slides ago.
And that’s the challenge that’s where we’re trying to tackle. We started working on this almost two years ago and, when it comes to our thought process, we were very focused on making something that is seamless when it comes to integration. We really wanted to avoid having tens of thousands of developers reintegrate our SDK and go through the troubleshooting process instead of them investing in other product (roadmap 0:11:12) features.
The zero discrepancies point is critical when you think about data, the second people judge the data based on accuracy, it closes all the trust and, even though data’s fine it's not being trusted to really make things work. And since the way the data is flowing through these pipes and, in many cases, there is some revenue changes coming in from the demand sources where some revenues change after the fact, some revenues coming in after the fact, so it was very critical for us that we look at the total number of revenue generated on that platform and you summarise your column of data from revenue generated yesterday, it ends up being the same number. And we wanted to make it actionable, without being trustworthy it's not actionable.
Optimising cost of data’s critical. When you think about ad based apps, ad based and heavy apps, mostly games, the amount of ad engagement can reach 50, 100 engagements per user, sometimes even per session, and that’s a lot of data to deal with. So we focused on optimising that part as well, which we assumed could be a big roboteer.
And when you have things actionable connected to the attribution sources, you want to close the loop on user acquisition automation. So if you know how much each user generated, where it came from, how do you automate that loop to make things move faster and grow your business.
When you think how much revenue each user generates from ads, each user has different ads served to them, different networks, different DSPs bid on it programmatic, which some of our beta clients can see these days or through the combination of inner bidding, programmatic and waterfall. You will see that today and that’s something that hopefully will be shifted to a new paradigm shift pretty soon.
But the fact that, today, is that most of the market is monetising through what is called multiple instances. Where each network is ordered with a specific CPM price point, where that network decides whether or not they would like to bid that specific value for that specific impression for that specific user, or pass to the next available down the line.
So each user engages with different ads, different ad placements, different ad products, and there's different channels bidding different values for each. And the entity that is mediating all of those engagements is the entity that really knows how much revenue that specific engagement generated. When you can combine that with user acquisition data through Adjust, or any other attribution platform, the value that you see here is that you see organic users behave differently with specific ads compared to other users that were acquired from other sources. Think about how people engage with rewarded products and if they're educating about these products and the value exchange we put on them or not.
And that tells you a completely different story when you start looking at your user acquisition data and start making (actions 0:14:38).
So what we launched is the capability to provide that data, to the attribution platforms or your BI stack, and can be consumed in a way that can truly make it actionable because of the lack of discrepancy, the optimisation of cost data. No required SDK updates, it's seamless, you flip a switch and you see this, as I'm going to show in the next couple of slides. And that’s the beauty of this product, that it fully paints the true average revenue per user picture when you think about ad optimisation and when you look at the raw data you can truly see, just like in-app purchase, how much each user generated, which exact sources did they engage with, for how many impressions, which placements, or which segments, and also take that into account when you're trying to optimise and maximise your ad monetisation.
This is a snapshot from the Adjust platform that is anonymised for another mutual client of ours, and you can see that this provides you with a full picture of how many users came from which source. All their top of the file engagements, impressions, click-through rate, clicks and many other columns that can be custom added to this dashboard.
And then, as part of your revenue measurement, you can see the amount of revenue events happening, the revenue generated by that cohort but also the ad impressions. How many impressions were generated by these 135,000 users? How much ad revenue was that generating and what is the ad revenue per mille? That is generated here.
And already in this snapshot you can see that users that are coming from different sources, organic, different UA channels are behaving completely differently when it comes to how exactly the demand sources are bidding on those users.
And when you download a snapshot of that in this specific example from the previous line, we’re talking about a game developer that has 55% of his revenue coming in from in-app purchase and 45% is coming from ads. So, theoretically, if they apply the Peanut Butter Approach they can just double their bids.
But when you look at the sources here you can see that there's a completely different picture with how those users are engaging with in-app purchases and with ads. Some channels are heavier in-app purchase, like Channel A, which looks very close to organic users. And the further you go down you can see that there's completely different examples here. Look at the lines H, I and J where – and K would be the most extreme one – where most of the revenue is actually coming from ads. Which would be 5x in these cases, 4x and 11x the revenue here.
And that is pretty critical when you think about growing profitably and making the right decisions. And this is something that is eye-opening. Across the board it's really interesting to see that organic users are actually behaving in a completely different way with ads. Less engaging with ads, as a much more different appeal when it comes to those users and, when you think about it, it makes sense. If someone is coming in from a rewarded video they will most likely convert through a rewarded video or an offer wall, when you think of it. And this is a great example that tells that story.
Going to pass the mike to Josh.
Thanks a lot Yevgeny. So that was very interesting to see some of those differences between in-app purchases and ad revenue and see those comparisons. So we've seen a bit about what user level ad revenue is and the current state of ROAS measurement.
So I wanted to talk a bit now about what exactly marketers can do with it and what are the benefits.
So knowing ad revenue at the device level is key for advertisers to understand their return on ad spend, so they can accurately measure the efficacy of their mobile advertising campaigns. Advertisers already had their in-app events counted, with the corresponding revenue on the Adjust platform. However, there was still a need to correlate those individual purchase events with user level revenue generated by advertising.
So user level ad revenue from ironSource gives advertisers that important other stream of income from their app, so they can confidently retarget their valued users, invest in new campaigns, gauge user journey and cross promote their other apps.
So, specifically, knowing user level ad revenue ensures that advertisers can spend on ad campaigns that generate a positive return on investment. So prior to this solution, as Yevgeny said, BI and marketing analysis for advertisers left aggregated, averaged ad revenue mixed for all their users. Advertisers can know how much they spent on a campaign by visiting the ironSource or the Adjust dashboard and they could know, in total, how much revenue was being generated in return.
However, they wouldn’t know which fraction of those users were generating the vast majority of ad revenue. So just because a user acquisition campaign isn’t running a long term deficit, of course doesn’t make it the best strategy.
So having that granular level of detail massively changes the picture. So you can optimise and deliver the best possible experience to the most users. We’ll see some examples of this in just a few minutes.
So thanks to the API provided by ironSource, Adjust can now do the heavy lifting of passing so many additional data points to advertisers. So they don’t have to create their own integration. Advertisers are already leveraging Adjust’s raw data offering, so they're receiving ad attribution, the events and many other useful data points, so this is a huge cost saving for an advertiser’s technical infrastructure. They can get all of this data the same way they already get it and there's no additional cost to their infrastructure.
So knowing exactly which users generate positive ad revenue means you can repeat your successes with precision and that’s just an incredible value.
So now we know about some of the benefits and what marketers can do with this user level ad revenue. And let's take a quick look now at how it works, with this slide right here.
So we’re going to start right at the beginning on a user’s phone. The user opens the app and an ad is requested. As the mediation platform, ironSource will serve a network’s ad which best fits the publisher’s criteria. So once the ad is shown inside the app, an ad impression is generated and ironSource records the advertising ID of the device, the associated revenue, the ad placement information and the any other relevant details. And this is possible, of course, because ironSource’s SDK is present inside the app.
So that’s an important first step. But now that ad impression information should be paired with all the other marketing data that advertisers need. So, because of this integration, Adjust can then enquire and record the same ad revenue information. That’s the critical new piece which allows individual ad revenue to be paired with the device’s advertising ID, which is, of course, a privacy sensitive way to identify the device. That is what allows advertisers to achieve that new level of granularity at the user level.
So, once Adjust has this data - so now we’re on the fourth point of the picture there - Adjust has this data from ironSource’s API, we provide it to advertisers in real-time. So Adjust already has a robust and privacy sensitive raw data offering for advertisers, the ad revenue data from ironSource will now be included with the other vital statistics that Adjust provides. So that’s the ad attribution data, in-app purchases, KPIs for multiple marketing channels, etc.
So once app developers have this data, they can correlate the ad revenue per device, along with any in-app events, to know the total income they have from a device. They can compare that income with the cost of acquiring the user and thus they have the return on ad spend. So that’s just the very first step for further analysis.
Okay, so now we know a bit how this integration works with user level ad revenue and I’ll turn it back over to Yevgeny now to see how marketers have already used this feature.
Thank you Josh. Diving in to some of the new capabilities that are really unlocked here – focused on user acquisition, though there are some insights I'm going to share at the end around ad monetisation optimisation as well.
This is probably the biggest paradigm shift that this capability unlocks when you think of it is, that before you couldn’t really build core analysis that includes all your data. You basically had only your IAP revenue, which means only the red part, at the bottom. And now with this capability you can literally build that core analysis that combines both ads and IAP and, for the first time, gives you the full ARPU picture, in this case it's by the 90 and then this will help you build the right LTV prediction curves and allow you to really understand better what is happening here and how can you turn this into the right UA strategy.
When you look deeper, there's a lot of interesting insights that this thing unlocks which, when we look at the two examples on the right, which is a day 30 analysis for certain cohort. You can see how different each app behaves when it comes to how revenue accumulates in-app purchase versus ads. And this is something that helps give a lot of visibility into how users behave and engage and you can just innovate better on monetisation products, better for both IAP and ads.
And, for the first time, it also allows you to build through return on ads and goals without any Peanut Butter. And this is where we took action on really making this scalable. Where, now that you have the capability to really define better return on that spend based on your cash flow and your profit strategy, the challenge here is now really moving from this before state, where you're bidding based on average LTV or trying to optimise retention on your UA side, there's not much you can do to really change bids. Within our user acquisition platform, and also within the Adjust platform, it gives you a full return on ad spend picture per source, per sub-source.
And now, for the first time, you can decide on bidding strategy and take action based on this full view user level measurement. And you can also take action using automatic bid optimisers to meet those return on ad spend goals to really help you operationally within your growth team.
On our user acquisition platform, this is how it looks like, very similar to Adjust, that Adjust perspective. This is specifically for the ironSource user acquisition platform, where it gives you full visibility into how much are you bidding per each user, what the uploader funnel look like, what's the installs per mille, the behaviour, and then down deeper to retention and in-app purchase data. And when you look at this example this is only in-app purchase data. So you have Run Rabbit running user acquisition on all these different games, and Hit the Clown and Talking Cat are generating this much of in-app purchasers, this kind of the before.
And when you combine in-app purchase and ad revenue, completely different picture is now unlocked and just like in the example a couple of slides ago, this is, for the first time, where you can feel confident that you have a full picture of measurement.
On top of that data, you can either decide on bidding strategies, applying different bids for each of these dimensions. And you can obviously go deeper. Or you can bulk upload a list of all these bids through a (C3 0:28:17) bid manager or you can just apply the ROAS optimise. Which, until we released this product, our optimiser only supported in-app purchase apps and now it's unlocking this capability of really applying your new ROAS goals for your ad based apps and, in order to improve profitability, you would like to define that goal, turn on the optimiser and let this optimise on your behalf. Or the goal is scaling this up while meeting your ROAS goals.
And there's a big, big difference when you try to do this manually, obviously, in terms of capacity, in terms of monitoring this in real-time. And you might as well let the machines do that while you're monitoring this. And in this example we’re seeing one of the few examples, this is something that you can look up on our website, it’s a case study we did with one of our clients, where your ROAS really improves no matter how good you are doing this manually, it just leaves a lot of time in the room to do other stuff while you're monitoring this and not really investing in bidding.
Summarising what we have so far within this product and this specific integration with Adjust and how it's done seamlessly where you just unlock this and you just have it. The first thing that really happens for the first time is measurement really covers everything on the user level. And then, thanks to that, you can make it count with actionable insights when you understand where those users came from within the UA platform or within the Adjust platform. And then this unlocks the capability to really run true return ad spend campaigns, define roles and provide granular bidding.
The interesting impact here is also on the adjunct, when there's things like you can optimise through segmentation, where you can actually understand better how the different segments behave based on where they came from. So if you understand that organic users, for example, do not engage well with banners, this is something that you can decide using segmentation and deep linking to decide not to show banners to these kind of ads… to these kind of users.
You can also understand the impact of ad revenue and in-app purchase per user. For example, you can test how did adding the rewarded video ad placement impact your in-app purchasing from these specific users. So if my in-app purchase revenue went up after adding this placement for those users that engaged with rewarded video maybe there's a positive impact there, or maybe there's not, and then take action. And this is something that unlocks that capability of accelerating that growth loop through a full measurement. And this is a great example of a product that touches all four steps of the growth loop that really unlocks the (recording scrambled 0:31:35).
We’ll jump into a few slides about the companies. Josh, I’ll let you go first with an overview of Adjust.
Okay, thanks a lot. So Adjust is the mobile measurement company. So that means we provide a range of services, we provide app attribution, so we’re analysing that correlation between advertising and then app installs and events. So that way you can know the effectiveness of your advertising campaigns. We’re also an industry leader in anti-fraud, so fraud, fraudulent traffic, advertising fraud, that sort of thing. As well as detecting bots and scripted behaviour inside of your app.
So we are very nearly 380 employees at this point spread across 15 offices around the world. And so we provide tracking for over 25,000 apps and that represents over a billion daily active users, so we are definitely a key industry player there and we just see that increasing in the future. So, Yevgeny.
Josh. So ironSource, essentially, what we do is focus on building the leading growth engine for mobile games, and this is obviously one of the features that is unlocking a critical capability here. These are some of our clients that we’re working with that are utilising our platform and are enjoying these kind of features for their portfolio of apps.
This is a snapshot from App Annie from a couple of weeks ago, of the top 20 games and how many of them are actually utilising the ironSource platform. And when you think about our product offering, it evolves around the growth loop as a growth engine. We covered user ad revenue in this webinar and the ROAS optimiser that is critical for user level ad revenue for in-app purchase apps, and a combination of those as well.
Monetisation A/B testing is something that also works very well with this product we want to… this different (recording scrambled 0:34:00) strategies, just like in the example I gave around, should I show banners or not, and take actions on that. And obviously to fuel our portfolio a very smart cross promotion engine that helps your portfolio grow with an optimisation engine that makes sure that the opportunity cost here is maximised.
Thank you so much. If you have any questions, please feel free to reach out to Josh or myself. Anything else Josh?
No, that’s all for today on our side. So we really hope you enjoyed the content and, like Yevgeny said, please don’t hesitate to reach out to either ironSource or Adjust, we’re very happy to provide you with more details. And any specific use cases as well on anything we covered. So thank you, Yevgeny, for joining us and thanks to everybody watching and see you.