When IDFA Is Gone, What Should Take Its Place?

Two options have emerged as the main alternatives for mobile ad targeting.

Written by David Finkelstein
Published on Sep. 23, 2020
When IDFA Is Gone, What Should Take Its Place?
Brand Studio Logo

When Apple announced in June that its identifier for advertisers (IDFA) would be an opt-in feature for consumers in its iOS 14 update, the adtech industry went into a frenzy. For something that seems so small and technical, many outside the industry were probably wondering: What’s the big deal?

An IDFA is a type of mobile advertising ID, or MAID, used by marketers to gain anonymous insights into user activity. These insights are invaluable for evaluating individual user’s needs and preferences to serve them ads that meet their interests. As the IDFA is the most widely used MAID, it has been a tremendous tool in providing cost-effective advertisements and helping end the dark age of “spray and pray” marketing, when ads would be blasted out to consumers without consideration of whether these messages would be relevant to them.

So when Apple decided to make the IDFA an opt-in feature for consumers, it felt like a seismic shock for the $80 billion ad industry. Without IDFAs, marketers would need an alternative means for effective mobile ad targeting.

While Apple has decided to postpone the IDFA changes until 2021, both new options and old alternatives to IDFA have already come into focus. However, the two biggest contenders are Apple’s newly developed SkAdNetwork and the tried-and-true hashed email (MD5s).


Apple’s SkAdNetwork

With the IDFA gone, the SkAdNetwork is the most robust option in Apple’s ecosystem. And with more than 100 million current iPhone users, it represents a strong alternative for those who rely on mobile app targeting.

This ad network allows advertisers to measure the success of their campaigns within a privacy-forward platform. Simply put, advertisers register their ads with Apple to be displayed to users, and if a user engages with the campaign, the advertiser gets notified that their campaign was successful, but this doesn’t include user- or device-identifying information.

The advantage of this network is that it delivers the most stringent privacy capability while still allowing some form of ad-targeting. The enhanced privacy practices are a major positive to this option, especially in today’s increasingly privacy-conscious consumer environment. However, it also makes targeting much more difficult: Advertisers will have to depend on contextual marketing in order to determine which campaigns are successful, thereby decreasing an advertiser’s ability to optimize and budget for future campaigns.

In addition, view-through attribution, or impression tracking, isn’t supported, and neither is click-through attribution for ads displayed in non-Apple media platforms. For these reasons, advertisers lose a lot of valuable information on how their campaigns are performing.

The SkAdNetwork also misses out on one of the biggest value-adds for targeting: real-time data. Since notifications of attribution are sent between 24 to 48 hours after the fact, advertisers won’t be able to target consumers at the moment they are in-market and won’t be able to tie app activity to a particular time, which hinders the usefulness of the data itself.

With all these confinements, the SkAdNetwork also does not allow for the targeting of ads with third party data, very much limiting the targeting ability and thus the potential return on ad spend (ROAS).

Finally, the SkAdNetwork is also a closed system, owned and operated by Apple. This means that advertisers wanting to utilize Apple’s user base would have to register with Apple’s network, and Apple would own all data gleaned from this process. As a result, more restricted access to raw data could limit the Adtech industry’s ability to use these insights to gain a deeper understanding of consumer behavior and create new innovative marketing strategies.


Email Hash — MD5s

The MD5 has been popular for nearly a decade and is widely used to connect with consumers across devices through an anonymous, secure identifier obtained through converting an email address into a 32-character string. While cookies and MAIDs were identifiers of choice for years, as they are phased out, MD5s will be brought into more focus.

Because MD5s are hexadecimal strings transformed from an email address that went through a hashing algorithm, the entire system processes sensitive consumer information down a one-way street that cannot be tied back to the individual. This makes it a privacy-focused identifier that can securely link data to create anonymous user profiles, but is still able to target consumers with highly personalized ads.

One negative to MD5s is that they are not as private as the SkAdNetwork. Since consumers generally maintain the same primary email address for several years, MD5s have a large map of digital behavior and activity. But since the identifier cannot be tied back to the actual consumer, the identifier still maintains user privacy.

Additionally, this open-use identifier benefits companies on an individual level, without having to rely on tech giants like Apple or Google. Any website, app or platform that has a registered user base will be able to benefit from stronger data, advertising relationships and monetization.

Having considered the pros and cons between MD5s and SkAdNetwork, the former especially when combined with IP address information emerges as the most effective and time-tested solution in a future without MAIDs.

Protecting consumer identity and sensitive information is essential, but advertisers also need to be able to cater to users’ needs and serve information that is relevant to them. With Apple’s postponement of the IDFA changes, the industry should use this time to align on a new gold standard for gathering consumer data. Until a new, more effective open system is created, MD5s should play that role.

Read This NextThe Secret to Better Email Marketing

Hiring Now
John Deere
Artificial Intelligence • Cloud • Internet of Things • Machine Learning • Analytics • Industrial