This is why your Facebook data is not accurate anymore

Your Ads Manager shows you a ROAS of 0.5, even though your shop is running great, and you hardly have any other marketing campaigns aside from Facebook? Since the iOS update and the increased rejection of cookies, it is becoming more challenging to track ads' performance accurately and optimize them accordingly.  

Why are we not seeing reliable data since iOS 14.5 anymore

Since iOS 14.5 the App Tracking Transparency (ATT) forces apps on Apple devices to ask your permission before tracking your activity across apps and websites.

If a user opts out, Meta loses access to its IDFA (identifier for advertisers). A unique device code that is used to track users’ activity across iOS apps and websites. It’s also used by the Meta Pixel. From this point on, Meta cannot track actions made by those opted-out users on your website since Apple is blocking them to do so with their ​​Private Click Measurement (PCM) tool. 

Because of this, Meta adapted Aggregated Event Measurement (AEM) to comply with Apple’s requirements. It “is meant to limit the amount of personal data used to facilitate conversion reporting and ads optimization, but is designed to solve for key advertiser use cases not addressed by Apple’s Private Click Measurement tool.” As it turned out, AEM is not tracking the highest prioritized tracking event (usually Purchase) for many opt-outs as it was supposed to. That’s why Meta additionally introduced multiple types of statistical modeling to report conversion events. 

Meta explains how conversion events such as Adds to Cart, ROAS, and Purchases are calculated in Ads Manager's numeric footnotes. Essentially past data from several sources is used to make up for partial or missing tracking data to “account for some results”. On past usage data, Meta calculates based on a statistical probability that a user COULD HAVE led to a conversion. It’s further explained that results and other conversions have some differences in how they're calculated. For example, the modeling for iOS devices differs from the modeling used for users who are no longer permitted to be tracked due to privacy regulations. So not only is the data poorly tracked, but it is not comparable due to different modeling types.

So what can a company currently analyze and optimize?

Currently, companies, and especially media buyers, do not make their decisions based on actual data but calculated data. Campaign management is based on many years of experience and sometimes very complex but still inaccurate approximate calculations (from different sources). Based on the still reliable platform data such as CTRs, at least user engagement can be measured with ads. However, well-clicking ads that sell poorly and poorly-clicking ads that sell very well are not uncommon. Conversions - the most crucial metric for ROI driven companies - can only be displayed very roughly on the platform.

Summary

The data presented in the Meta Business Manager is no longer suitable for making data-driven decisions for optimizations due to the incorrect and inaccurate allocations. A solution to the problem is not yet in sight on behalf of the Meta Group. 

The biggest issue in the ad accounts here at Tracify is the following: Conversions are attributed to the wrong Ad Sets and Ads by Facebook, which results in scaling losers and cutting winners. This is costing ROI driven businesses significant profits. Especially when only working with blended metrics.

Sources:

• https://www.facebook.com/business/help/1329822420714248?id=428636648170202
• https://www.facebook.com/business/help/311705270326952
• https://www.facebook.com/business/help/721422165168355?id=1877298665783613 
• https://webkit.org/blog/11529/introducing-private-click-measurement-pcm/

‍