Earlier this week Apple hosted its annual developer conference, WWDC 2022, and introduced iOS 16. Unlike the previous developer conference that shook the mobile marketing industry with many App Store announcements, including Product Page Optimization (PPO), Custom Product Pages (CPP), and In-App Events, this year was significantly calmer.
Many of the sessions were explorations of already announced features such as the ATT framework, and more.
I know that you’re busy, so I wrote this short note to keep you on top of the news and ensure that you don’t miss any important updates for your work.
What is new with SKAdNetwork?
This is perhaps the biggest update from the conference. Apple announced the next version of its attribution solution, SKAdNetwork, with several updates to the framework.
- Web to app attribution
As expected, Apple SKAdNetwork will now include support for web to app attribution, allowing both paid and organic user acquisition to get attribution data for downloads coming in from your website or other web sources.
One question here remains: “How will the privacy thresholds affect this data?” Or in other words, how large does your daily inflow of web traffic need to be for SKAdNetwork to return attribution data? We will continue to report on its feasibility and value as this rolls out.
- Multiple conversions
This is pretty big. With SKAdNetwork, the main limitation was allowing paid UA teams to get only a single conversion value back, which represents one downstream event (be it an in-app purchase, a completed level, registration, etc.)
This made it impossible to collect downstream event data continuously, so the ability to understand how much revenues were generated from a certain campaign was severely limited.
With multiple conversions, SKAdNetwork can now return up to three conversionValues for each download, each with a different attribution window which is 0-2 days, 3-7 days, and 8-35 days, respectively. Each of these conversionValues will be able to include multiple engagements.
Assuming the privacy threshold is being met for a specific source, this means a more accurate value can be associated with a campaign. This still isn’t fully tracking downstream events, which would allow for the accurate calculation of ROAS. But it significantly improves the level of attribution data you will have in order to make decisions, as multiple in-app purchases can be tracked.
This is probably the biggest update to SKAN.
Moreover, SKAN will now return limited data for sources that didn’t meet the privacy threshold called Coarsed conversionValues. So for these campaigns, you will still be able to see limited conversionValue data which will only include one out of three values (low, medium and high). This can be used to get an initial understanding of these low traffic campaigns, instead of no data at all – which is the current reality.
- Hierarchical Conversion Values / Source IDs
Another important update is that instead of just 100 campaign ID values, more data will be returned about the campaign that drives attributed downloads. SKAN will now allow for a four-digit number that is associated with a certain campaign ID to also be returned. This can be used to add information to a campaign around things such as the placement type, the creative type, the campaign value or more.
Effectively, instead of 100 campaign IDs advertisers, you now have 10,000 possible values. This means that insights can be produced in a way that more closely resembles traditional attribution where granular data was available at the ad creative level.
This will also allow certain ad networks to better optimize their campaigns as they’ll be returned information about the value different creatives, placements, or even audience targeting configurations are able to drive. It’s still far from the real-time data they got in the past which they could also associate with an individual user (allowing them to build in-depth user graphs used for targeting), but a big step forward and allows them to offer more quality campaign optimization to advertisers.
This added information will only be available for sources and campaign IDs that crossed the privacy threshold, so there is a benefit for scaling spend on specific campaigns here, from a measurement standpoint.
Performance benchmarks in App Store Connect
Another update: App Store Connect now brings you benchmark data to allow you to stack up your app performance vs. industry benchmarks.
This includes benchmarks on conversion rates on the acquisition side, retention and crash rate benchmarks (which is more similar to the Vitals data you can get on the Android side from the Google Play Console), as well as monetization benchmarks such as average proceedings per paying user.
As with any benchmark, you need to take into account that you don’t have full visibility into its composition, so it’s hard to estimate how comparable the apps and mobile games are that you’re measuring yours against.
For example, looking at Referral conversion rate benchmarks, which represent paid user acquisition traffic for the most part, you have to take into consideration the fact that conversion rates are highly influenced by the ad networks you’re using, as well as the type of ads and targeting,
If you’re seeing that your app’s conversion rate is lower than the benchmark, it doesn’t necessarily mean you’re not doing a good job. It could simply mean that the benchmark includes apps or mobile games with significantly different UA channel mix.
That being said, you can definitely use benchmark data as one more data point that allows you to identify areas for improvement, as long as it’s not the sole data point you rely on.
To learn more about the other features Apple introduced at WWDC, check this piece.
In Conclusion
What do these updates show? That the landscape for measurement of paid user acquisition is still being shaped, and these developments will be warmly accepted by the UA side of mobile marketing teams.
We still need to remember that this plan looks great on paper, but the reality so far has shown us that SKAN is far from a complete framework. To the extent that it was barely usable by teams to actually generate insights that would allow them to optimize UA spend.
In part because of the privacy threshold limitations, and in part because performance was very shaky (with postbacks coming in an unreliable way). But this evolution of SKAN as a framework definitely calls you and your team to go back to the drawing board and experiment with it more to understand exactly its feasibility, and how this added information will benefit you as you make your UA decisions.
As this update will come later this year (not part of iOS 16.0), we will keep you posted on how that roll out pans out, and the potential benefits UA teams are able to get from it.