What Shopify advertisers need to know about Apple’s iOS 14.5 and Facebook’s AEM
If you’re running an online business you probably heard a lot of buzz around Apple’s release of iOS 14.5 and how it will profoundly impact online advertising. However, the majority of the content available on this topic is either PR driven by the rival Apple and Facebook, trashing each other. In other cases, it’s content written by tech reporters that too often don’t understand the nuances of online advertising, so it doesn’t really convey the happenings correctly.
This is why we’ve chosen to cover this topic from a performance advertising perspective. I’m a web analytics expert, working for a decade now in this field. Together with Etgar Shpivak, with over 15 years of experience in PPC (and one of PPC Hero’s Most Influential PPC Experts in 2020), we’ve sat down to break this down into actionable insights.
A quick recap of iOS 14.5/PCM and AEM
Apple’s release of iOS 14.5 (in public Beta, as of April 18th) is expected to be a major step towards enforcing greater privacy measures on Apple devices. The main change is around the way advertisers can measure and report on their campaigns targeting Apple users.
To enforce this, Apple introduced a new protocol called SKAdNetwork. This protocol enforces several limitations that restrict the reporting on campaigns by detaching any identifiable information from the user that interacted with the campaign (clicked an ad, installed an app or converted). Unlike the current situation, where advertising networks (e.g. Facebook or Google Ads) can trace a specific conversion to an ad view or click, the reporting is now handled by Apple, in a way that removes this data and reports only on the campaign that drove a certain user action (install, purchase, etc.). The data sent is also controlled by the permission to track prompt.
This protocol is relevant for apps only and does not apply to websites. For websites, Apple has introduced a similar protocol, PCM (Private Click Measurement), that will apply to Safari users on iOS 14.5. This protocol, though similar to the SKAdNetwork one, differs in several ways.
First, it won’t be enforced on the advertising platforms, i.e. all pixels will continue to work uninterrupted. Second, since we navigate many more websites than apps, there will be no prompt loading on every site, but rather a browser setting in Safari
So there’s a new protocol that isn’t enforced, why is everyone freaking out? The ad industry didn’t approve of Apple’s unilateral announcement of these two protocols. While most advertising platforms are still on the fence, Facebook has been vocally opposing this. Their claim has been primarily that small businesses will be damaged by this change, as campaign performance is expected to deteriorate. Besides trashing Apple, Facebook has also proposed its own, equally restrictive, protocol for tracking advertising performance. This protocol, called AEM (Aggregated Event Measurement), applies many of the limitations set by Apple and will be enforced with the official rollout of iOS 14.5.
Why is Facebook rolling out such a restrictive protocol? Glad you asked. Apple’s protocol won’t be relevant if the advertising platforms won’t adopt it. Unlike SKAdNetwork, it cannot technically be enforced. Facebook is quickly rolling out this protocol in hope that such a protocol that comes from the ad industry will be more widely adopted than one coming from Apple. The winner will probably “take it all” and set the standard for years to come.
So what are the key restrictions set by PCM/AEM?
Let’s break down the three key elements of AEM/PCM.
Campaign level reporting
The main element of both protocols is restricting the data reporting to the campaign level only. This means that the tracking will stop at the campaign, and any data on ad set or ad level will be extrapolated using statistical modeling. For example, we can know for a fact that campaign A drove 50 conversions, but the split of conversions between the two ad sets in the campaign will be modeled, i.e. Facebook Ads will assume how many conversions were driven by each ad set.
To prevent manipulation, Apple is also limiting the number of campaigns available for reporting to 256 concurrent campaigns. Facebook hasn’t declared any similar limit but has recently limited the number of active ads per page.
Another element of the protocol is a delay in reporting. This is done to prevent exploitation of the campaign level reporting to tie back to specific users. The delay window ranges between 24-48 hours in Apple’s PCM and 48-72 hours in Facebook’s AEM.
The last element is the limit on event reporting. This happens on two levels. First, the number of events available for each advertiser will be limited. Apple set this limit to 16 events (after originally limiting it to 8 events), while Facebook set the limit to 8 events. This limit applies per domain, not per ad account or pixel.
Second, the event metadata will be trunked, so that it cannot be used for identifying users. This means those event parameters will not pass, for example, the item purchased or its price.
How this will impact advertisers
In absence of user data, the reporting breakdowns in Facebook will become significantly limited. Limited data, such as location, will appear on the campaign level only, but most demographic data will be lost. This means that analyzing which user segments convert better will be a difficult task.
With user data gone, Facebook won’t be able to report on view-through conversions. Because of this, the conversion windows will now be capped at 7-day click or 1-day view (modeled). This is bad news for anyone with a long conversion funnel (e.g. expensive product or a long consideration cycle). It also damages the ability to analyze cross-platform conversions, for example, a user viewing an ad on Facebook and converting later via Google Search.Preview Changes(opens in a new tab)
Since events won’t pass metadata, the value of each purchase will be lost. This will mess up ROAS reporting and will retain only CPA (Cost per Acquisition).
For advertisers that want to keep using this functionality, Facebook asks that 4 out of the 8 events are assigned just for the Purchase event. Though no formal statement has been made, we understand that the Purchase events will be split into 4 groups of value (e.g. $1-24, $25-49, etc.) and every purchase made on the site will be assigned into a matching group.
Since the data is reported on the campaign level and extrapolated down to the ad set level only, ad level reporting is expected to be very minimal.
This means that optimizing our creatives will become difficult. While metrics that occur within the platform, e.g. clicks/link clicks, etc., are expected to report accurately, any action done outside of Facebook will probably not be associated with a specific ad.
Moreover, any change done to an ad (copy or visual), will only reflect in reporting after 48-72 hours, making every optimization cycle painfully slow.
In this post, we covered the main changes happening with Apple’s iOS 14.5 and Facebook’s AEM. These will both have a great impact on how Shopify merchants can continue to drive revenue and scale. In the next post in this series, we’ll touch on the different actions you can take to continue having visibility for optimizing your campaigns.
This article was brought to you by: Elad Levy from Fixel
Fixel is a fully automated solution that helps Shopify advertisers target high value visitors. By analyzing anonymous visitors, Fixel can identify which ones are closer to conversion – so you can focus your advertising efforts on these. Remarket efficiently to engaged visitors or scale to new audiences with high value lookalike seeds. Try free for 14 days.
Get Fixel Shopify app with 15% OFF discount, just use our code: ZAGA15