How Modelled Online Conversions Work Around Apple’s App Tracking Transparency

Updated: January 7, 2022
How Modelled Online Conversions Work Around Apple’s App Tracking Transparency

When Apple rolled out its iOS 14.5 updates in early 2021, it included a raft of new features for users. The most significant of these features by far was the App Tracking Transparency addition, hailed by many users but panned by giant platforms like Google and Facebook.

The controversial feature was initially announced for the iOS 14.0 update released in June 2020. However, it was delayed so that app developers could have more time to adapt to this significant change.

Before now, developers had wide-ranging abilities to track people’s usage, location, and behaviour using their apps. This meant you could effectively be profiled so that companies could better market their products and services to you.

Users always had the ability to opt-out of this tracking, but the option was hidden in menus, and many people didn’t bother looking for the setting. The new option means users are presented with the choice to turn off tracking front and centre when they load an app. This means many more people will voluntarily choose not to be tracked.

While this is regarded as a win for user privacy, it has placed an enormous challenge for companies like Facebook and Google. It mainly impacted their ability to place relevant and timely advertising in front of consumers.

How iOS 14.5 impacts digital advertising

Straight out of the gate, any apps that do not include the App Tracking Transpancey prompt are being blocked from Apple’s App Store.

Apple customers have widely used the prompt. However, the number of people sharing their unique Identifier for Advertisers (IDFA) plummeted from 70 per cent down to as low as five per cent because they are not allowing apps to track their behaviours.

This impacts all platforms that have relied on this IDFA data to generate paid advertising targeted toward a relevant audience. Google and Facebook are the biggest names affected, although other platforms like Bing and LinkedIn have also relied on this IDFA data in the past.

What is Modelled Online Conversions, and how is it helping advertisers?

With no more access to IDFA data, platforms developed other methods to predict consumer behaviour.

But iOS changes aren’t the only ones that have impacted advertisers. Consumer behaviour has changed too, making the customer journey much harder to measure. For example, the modern consumer will often look at items that intend on purchasing across multiple devices and even use different browsers on the same device. Plus, cookie restrictions across websites has also thrown up a barrier for targeted advertising.

The solution to these challenges has come in artificial intelligence (AI) and machine learning (ML). For example, conversion modelling is being adopted with holes punched through the customer journey because of iOS 14, consumers using multiple devices and no cookies on websites. This uses AI and ML to take what data is available and feed it into algorithms, which also use historical trends to provide accurate measurements that can be used in paid advertising.

Conversion modelling is not a new technology

Fortunately, conversion modelling software has been used for years, developing and refining it. Previously, it was used in situations where customer journeys were missing data because of online and offline behaviours. However, years of generating reliable metrics on consumer behaviour across digital and physical mediums mean this technology is now highly refined.

Modern conversion modelling software uses various signals across various devices, browsers, the remaining app and cookie data and other sources. In addition, it uses advanced automation that removes uncertainty and creates reliable metrics that can be used for advertising.

How Facebook is using modelled conversions

In direct response to the changes made by Apple that prohibit data collection unless app users consent, Facebook has adopted a twin-pronged approach to conversion modelling.

With companies spending almost USD 12 billion in 2020 alone on Facebook advertising, the social media giant needed to find a way to get around the changes to data collection. Facebook’s conversion modelling uses these two pillars:

  1. Measurement and reporting tools: Modelling has been incorporated into these essential functions to provide better and more accurate conversion reporting. Machine learning is used to estimate conversions that can’t be directly observed because of the new data restrictions.
  2. Optimisation and performance tools: One of the biggest guns Facebook has to fire in the advertising stakes is the platform’s ability to optimise ads using powerful AI and data. Modelled conversions are essential in this optimisation process in 2021 to ensure advertisers still get this optimisation in an era where only partial data is available. Conversion modelling ensures advertising still performs because it if didn’t, Facebook would lose its most significant appeal for advertisers.

Additionally, Facebook has provided a configuration tool to apply modelling solutions for web events for people who use iOS 14.5 and a Conversions API that can be added to Facebook pixel for better visibility and data sharing on user websites.

Google’s modelling response to the loss of data from cookies

While the iOS update impacts Google’s paid advertising, it has been the requirements for consent to collect data using cookies that have affected the search engine giant the most.

Google rolled out Consent Mode for when users do not consent to cookies, and conversion modelling has been added to that function to help fill in the gaps. Using machine learning, available data, historical trends, and other information are used to fill in the missing attribution paths.

Consent Mode has been highly successful, recovering more than 70 per cent of ad click to conversion journeys lost because users had opted out of accepting cookies.

Consent Mode data is shown concurrently with observed conversions in Google Ads so advertisers can see the conversion data split across audiences that have consented to cookies and those who have not.

Advertisers can use smart bidding strategies in Consent Mode and still reach the majority of consumers relevant to them, even if analytics tools like cookies are no longer recording their data.

It’s going to be an interesting future in terms of privacy, digital marketing and Facebook advertising – that’s for sure.


Mira -

Head of Paid Media

1300 558 659 -

Related Articles

Let's Chat

We’re an Digital Agency that has achieved outstanding real results for businesses throughout Australia.

Let's get in touch

We're ready for any enquiries you have.

8:30am - 5:00pm