SKAdNetwork plus everything else: Building a single source of mobile marketing truth
How do you build a single source of truth for mobile marketing data that includes a major new set of attribution and conversion data from Apple SKAdNetwork?
Mobile marketers who wanted a simple life might be forgiven for wondering lately if they should dust off their buy-a-house-in-the-country-and-make-goat-cheese plans. After all, it is 2020. And Apple’s planned IDFA opt-in changes are still coming. However, Singular’s Director of Product Marketing, Saadi Muslu, has some good news: Singular’s got your back.
And yes, it can be simple. No farm, house in the country, or goat cheese required.
Listen to the interview behind this story:
The core questions most mobile marketers have when they’re thinking about SKAdNetwork are the obvious ones: how will I measure marketing on iOS? How will I set up my conversion protocols? Do I need to change my app behavior to get better signals? Is SKAdNetwork data guaranteed accurate and fraud-free because it’s cryptographically signed by Apple? If I ask, will people opt-in to sharing their IDFA?
But there’s a whole other set of questions to think about too:
- How do I build a single source of truth with SKAdNetwork data included?
- How can I aggregate all of Apple’s postbacks?
- How do I get a unified view of attribution when I have device-level data via GAID (the Google ad ID) and IDFA (from people who opt-in), plus aggregate data from SKAdNetwork, plus, potentially, data from web-to-app marketing flows that preserve both privacy and granularity?
Granted, this is a lot of questions. But don’t run off to the goat farm just yet: there is a simple solution, Muslu says. Here’s a brief, shortened version of our conversation, which you can hear in full at the top of this blog.
John Koetsier: Before we even talk about SKAdNetwork or iOS 14 or anything like that, why is centralized marketing data so critical?
Saadi Muslu: Historically — and today more so — marketing data has been really fragmented. So whether that is upper funnel data like ad spend and creatives that marketers need to collect and standardize across channels and ad networks, or conversion data that typically comes from attribution.
So data is essentially scattered all over the place. And it can be time-consuming to collect all that data, transform it, make sense of it, and really get a clear view of which ads and campaigns are driving the most growth for marketers.
Koetsier: It’s also challenging to see your ROI in that case, right? Or what all your costs are or where you’re getting the best LTV … all those sorts of things, correct?
Muslu: In particular upper funnel data — like cost, impressions, and creatives that you get from publishers and ad networks like Facebook or ironSource — need to be tied to attribution, the bottom funnel data, in order to understand what type of business impact you are driving. Marketing inputs and associated business outcomes.
Understanding ROI and LTV is definitely challenging.
Koetsier: So we’re in a scenario where a lot of marketers have fragmented data: data in different systems. How has that been impacted by iOS 14 and SKAdNetwork?
Muslu: We’ve always seen a lot of fragmentation in the upper funnel data. Particularly as marketers are scaling, they’re probably using different channels and scaling the number of ad partners that they’re working with to get incremental reach.
And each ad partner reports with their own nomenclature … and they might have different kinds of ways for you to collect that data from them. Attribution for the most part historically — mobile attribution in particular — was centralized with your mobile attribution provider. And now there’s going to be a few different types of data sets for attribution that you really need to make sense of.
So it’s definitely a lot more fragmented in iOS 14 and beyond.
Koetsier: If you can’t bring that together … if you can’t, as a marketer, understand both that top of funnel data that you’re talking about and that lower-funnel attribution data where SKAdNetwork comes into play … what are the implications of that?
Muslu: The biggest problem that we see would be partial and incomplete views of marketing performance. It’s all about accurate, granular, and timely insights. For that, you need the ability to combine different data sets.
As we mentioned, upper-funnel data is reported in aggregate. IDFA-based attribution data is user-level, and now the SKAdNetwork attribution data is also aggregated. So you need a way not only to standardize amongst each of the data sets, but then to interpolate and combine in order to understand what’s working and not. And if you’re not doing those interpolations correctly, or if the schema that you’re using to standardize isn’t really optimized, then you’ve got problems.
First, you’re not going to be able to get accurate data. And that’s just like navigating a boat with a broken compass … not being certain is a problem, especially when you’re scaling.
Second, the other issue would be speed to insights.
We are dealing with a lot of tough times with the global pandemic and economies are up and down and fluctuating … things change quickly! You need to be able to have the speed to insights or what we call near-real-time reporting available at your fingertips. You need to quickly react to shifts in performance and make sure you’re not missing out on any optimization opportunities to see better performance.
Or to squeeze more growth out of your growth lemon, as well as to know when to cut your losses.
Koetsier: Once data is centralized, what can you do with it?
Muslu: There are a few things to think about here.
First, thinking about high-level reporting or the kind of big picture view that you need is really critical: understanding what channels are working for you … what budget allocation decisions you’re going to have to make. Maybe you need to shift some budgets from network A to network B in order to really get the performance that you’re trying to get to.
But also across campaigns: instead of logging into Facebook separately, or your other ad networks’ reporting platforms separately, you want to be able to see that all in one reporting table. And another thing: being able to learn from the different platforms you’re using. Also, app developers have different apps. So they might want to be able to visualize and report on performance across their apps in one given report.
Koetsier: So there’s a lot of dislocated data in a lot of different places that need to be brought together so you can squeeze that growth lemon and get the growth. How does Singular help solve for that?
Muslu: We have a full end-to-end attribution and analytics platform that combines what used to be fragmented or siloed data into one unified view of performance.
One, I’m going to start with cost aggregation. When we kicked off Singular, that was a big challenge that we saw. And it’s really more than cost aggregation … it’s all about connecting that upper funnel data that’s coming from the different ad networks.
So it’s cost, it’s creative, it’s bid data.
Our cost aggregation solution is able to, no matter what source you choose to work with — whether they have an API or they’re able to push data via CSVs even — collect that data quickly, and standardize it. And then the second piece we talked about — the conversion data needs to connected to the cost data. We partnered up with mobile attribution providers to pull in the attribution data and then combine that with your standardized campaign data so you can join costs and revenue to understand marketing ROI.
And what we really realized was that if our customers could control and report on attribution data in such a way that it would be much easier to combine with the upper funnel campaign data, we could actually unlock ROI at very granular levels, which is really important for scaling. So we built our attribution solution with that foresight to fill in the gaps. And that attribution portion of our product has grown over the years. And this year as well with our first-to-market SKAdNetwork attribution support!
Plus of course, if someone decides to provide consent, you’re still going to have that IDFA-based attribution. But there’s also GAID on Android, and SKAdNetwork conversions … and we have a web attribution product and a cross-device attribution product, which, you know, takes mobile attribution data and web data and de-duplicates it.
So you can measure marketing performance across platforms, like web to app measurement to understand those different user journeys.
Koetsier: That is really quite challenging, right? I mean, because on a certain level … attribution for iOS 13 and previous versions had a certain simplicity. You could get the IDFA for, let’s say 70% of devices, you could get the Google Ad ID for almost 95% of Android devices, so you had really, really granular vision to interpret what was happening. Now you have a lot of different data sources that you have to put together in a holistic picture for marketers.
This doesn’t sound like an easy challenge.
Muslu: It’s not easy by any means!
And that has implications for having a solution that’s either partial or in-house. Quite frankly, the implications are going to be misreporting, inaccurate optimizations, speed to insights, and keeping up with those many, many marketing data pipelines and continually managing them all.
Koetsier: So, as a marketer you’re going to lose some granularity when you go to SKAdNetwork. How does this work with what Singular is offering?
Muslu: Going back to our core mission, which is being dedicated to providing really good ROI — marketing ROI analysis — one of the key challenges with SKAdNetwork out of the box is the inability to understand return on ad spend and the inability to look at those bottom-funnel metrics.
So that’s what we tried to solve for first.
And we have built a solution which provides up to seven-day cohorted ROAS. This goes back to our strengths in interpolating and combining and joining those fragmented data sets. We have our iOS 14 compatible SDK as well as our server-to-server solutions, and that’s going to call and collect all of the postbacks across the different ad networks that you’re working with.
That’s going to make the postback collection with your ad networks fully managed and automated. And it goes a step further: validating all the postbacks for fraud.
Koetsier: Saadi, there are huge challenges and complexities here. You are simplifying them for marketers. How can they get that? How do they get started?
Muslu: Well, we tried to make it pretty simple! Navigate to singular.net and check our solutions: there’s a specific page for SKAdNetwork analytics and attribution.
There’s even a demo video there for you to see for yourself what the reporting will look like within Singular’s native reporting. It talks a bit about our ETL functionality too, since folks like to pull that data to make it available in internal warehouses like Redshift and BigQuery as well as internal visualization tools like Looker and Tableau.
You can also schedule a demo … we’re quick to get back to you. Not only are we going to demo the solution for you, but we’re always open to chat about your specific setup and needs, whether or not you’re ready to make a purchase.
And then the one other thing that I will kind of plug shamelessly here is we have a Slack group: the MAP Slack group. It’s an open community for discussing mobile attribution privacy … there are ad networks there, other MMPs, lots of advertisers, and we’re actively surfacing learnings that we’re seeing with SKAdNetwork or with iOS 14 measurement in general. So join that Slack group, ask questions, and we’d love to chat with you.