The release of iOS 16 for iPhone users has sparked a concerning trend in the world of ad fraud. Several iOS updates over the past two years have significantly limited the tracking of mobile IDs and associated data, requiring user consent for tracking. As a result, data companies have been at risk of losing revenue, leading to the emergence of fraudulent practices to resell old data as new.
Below is a breakdown of the impact of recent iOS updates on ad fraud. We quantify the scale of the issue, explore the concept of invalid “events” as they relate to ad fraud, discuss the implications for marketers with a decrease in the number of addressable mobile devices, and highlight the need for evolving identity resolution techniques.
A Breakdown of the Ad Fraud Issue
There have been several iOS changes over the past two years that have led to increases in ad fraud across the board. With the implementation of the iOS 16 update on iPhones around the world, data vendors were facing declining volumes of mobile data due to limited tracking capabilities. As a result, many of them began generating invalid data for the purposes of reselling it to advertisers as new.
Without proactively checking or evaluating the data being sold by third-parties directly to advertisers, the scale of ad fraud in digital advertising is now staggering. Recent investigations reveal that up to 80% of the identity data being sold as new to major ad platforms is actually invalid. This alarming figure highlights the extent of ad fraud resulting from the recycling of outdated data.
Marketers find themselves largely reliant on ad platforms to combat this problem. Unfortunately, there is little marketers can do directly to detect and prevent the entry of bad data into their decision-making processes. Ad platforms that ingest invalid data put their identity graph at risk of corruption, leading to advertisers reaching the wrong audience. This misalignment can have a negative impact on their return on advertising spend (ROAS) due to the decreased performance of their campaigns.
Manipulation of “Events”
Invalid “events” in the context of selling advertising data refer to the recycling of data. For instance, a data vendor may take an event such as a user logging into an app like Zillow, which is valuable for advertisers interested in potential homebuyers, and present it as new data. However, if that “event” actually occurred a year ago but the data vendors are selling it as new, advertisers believe they are targeting someone in the market for a home, when in reality they have probably already purchased one, rendering the ad essentially useless to the audience. This is what has happened at scale with the new privacy changes that came with iOS16. When billions of data signals across multiple verticals undergo this recycling process, billions of ad dollars are wasted.
Due to the nature of recycled data, fewer and fewer mobile devices become “addressable.” For example, if recycled data is linked to a mobile ID that no longer exists because the consumer got a new phone, that ID becomes unaddressable. In other words, advertisers believe they are targeting someone who is no longer using a given device. Additionally, Apple’s efforts to protect consumer privacy by generating temporary mobile IDs for users who opt out of tracking further contribute to the decrease in addressable mobile devices. These temporary IDs, although indistinguishable from valid ones, are essentially invalid since they cannot be reached in the ad bid stream.
While the prevalence of bots on the internet remains a separate issue, recycled mobile IDs cannot consume ads like bots do. However, if ad platforms ingest these invalid IDs and link them to consumers in their identity graphs, the consequences can lead to an overlinked identity graph, upon which programmatic advertisers build and deploy their ad campaigns. This inflationary effect on audience size negatively affects campaign targeting, reach calculations, and overall ROAS.
The burden of all of this ad fraud lands on advertisers and their campaign budgets, underscoring the need for identity resolution techniques to adapt.
Need for Evolving Identity Resolution Techniques
Ad platforms and identity graph providers cannot rely on the same simplistic metrics used in the past. They must reevaluate their processes, recognizing that linked IDs do not necessarily imply addressability. Building an identity graph becomes more complex, requiring more data, diverse events, and authentication mechanisms to distinguish between addressable and non-addressable IDs.
Marketers now find themselves at the mercy of ad platforms and face the risk of reaching the wrong audiences and poor performing campaigns, while the decrease in addressable mobile devices further compounds these challenges. To combat these issues, ad platforms and identity graph providers must employ new identity resolution techniques that incorporate more comprehensive data and authentication mechanisms. By adapting to the modern realities of ad fraud, advertisers can safeguard their campaigns and make more informed targeting decisions.