The benefits of universal identifiers on connected TV and mobile
Universal identifiers gained traction in the ad tech collective consciousness when Google decided to phase out third-party cookies from their ubiquitous Chrome web browser.
Following that decision, ad platforms, publishers, and advertisers began searching for solutions to establish user identity on the web without third-party cookies — and companies with universal ID solutions were quick to hawk their wares.
Different flavors of universal identifiers exist, but the industry seems to be leaning towards solutions that allow publishers and advertisers to securely and privately match first-party data on PII (personally identifiable information) provided by users. The most common piece of PII to perform a match on is an email address, but it is also possible to use other signals like a phone number.
Universal identifier solutions would prevent the need for pervasive and passive tracking on the web via third-party cookie collection and syncing. These solutions can help establish a framework of personalized advertising on the back of opted-in data collection and reward publishers with scaled first-party data sets. They could also lessen the mistrust between users and ad-supported companies.
But do universal identifiers offer any value outside a web environment? Yes — their benefits extend to connected TV (CTV) and mobile in-app environments.
It is understandable if there is an intrinsic mental link between universal identifiers and third-party cookie deprecation on desktop and mobile web. Google's decision to remove third-party cookies from Chrome was the impetus that sent the industry scrambling for solutions — but some of these solutions are extensible beyond the web.
Let's examine how universal IDs can fit into advertising systems on connected TV (CTV) and mobile in-app — places not threatened by the looming peril of third-party cookie deprecation.
Universal Identifiers on Connected TV
If a connected TV application requires users to log in, then a publisher has the opportunity to start converting user email addresses into universal IDs. They can then use the created identifier to match their data to advertiser data without ever exchanging or revealing any personally identifiable information.
But why would they want to? There are several reasons for connected TV publishers to consider integrating a universal ID solution.
Fewer and fragmented device IDs
As a quick reminder, device IDs are persistent alphanumeric strings that device manufacturers tie to a device to allow advertising systems to identify individual users. Ad platforms can match a device ID to advertiser data or third-party data collected elsewhere to deliver personalized ads or measure attribution (tying ad impressions to purchases or conversions).
Many TV panel manufacturers have only started creating persistent identifiers tied directly to a device in the last few years as they leaned into establishing advertising revenue streams for their own businesses. Not every connected TV device has an advertising identifier (device ID) available, and CTV device IDs have not been floating around the digital ether as long as mobile device IDs like Apple's IDFA on iOS or Google's GAID on Android.
Users don't typically browse the web or make purchases with their TV — they confine these activities to desktop and mobile devices. Consequently, advertisers or data brokers do not have the same limitless opportunities to build fine-tuned data sets based on CTV device IDs as they do with mobile device IDs or cookies.
Nearly every human in the developed world has either an iPhone or an Android phone. Compare this to the fragmented ecosystem of connected TV panels and devices.
Universal identifiers can create an additional opportunity for publishers to increase their user match rate with advertisers on CTV. The limited and fragmented nature of CTV device IDs becomes a non-issue if publishers, advertisers, and data providers can match their independently collected user data (like email addresses). Universal identifiers can span all platforms and devices, so they are immune to device ID fragmentation and can help with scale.
Roku and Amazon command 41% and 18% of total big screen time in North America, according to Conviva. But these two top dogs of CTV may have to act to prevent competitors from chipping away at their lead and perpetuating CTV device fragmentation.
TV manufacturers like Samsung and LG are building more useable and snappier CTV operating systems and have begun to reach feature and UX parity with their plug-in peers — which creates a compelling reason not to buy a Roku or FireTV device.
More usage on varied CTV operating systems not only fragments the device ID ecosystem (lessening their scale and value) but may also have pressured Amazon to manufacture TV panels in-house to avert user loss. Reports point to Roku exploring a similar strategy to manufacture TVs that would extend the reach of their operating system beyond existing licensing agreements with TCL, Hisense, Sanyo, Philips, and Sharp.
Manufacturers could restrict device IDs
Even if a connected TV device manufacturer grants access to an application to use their device ID, they may eventually have an incentive to keep this information to themselves. Vizio, Samsung, LG, Roku, and Amazon all have burgeoning advertising businesses — and they could eventually conclude that restricting device ID access offers a distinct competitive advantage.
These CTV manufacturers with complimentary advertising and data businesses may want to draw advertisers to conduct buys through their proprietary ad platforms. They may desire to restrict access to their device's advertising IDs to hinder third-party ad platforms' targeting and measurement capabilities.
Apple set a precedent for such a strategy by introducing the ability for users to opt-out of tracking on iOS. While protecting user privacy is a noble effort, make no mistake in understanding that Apple had ulterior motives.
Apple advertising and subscription revenue grew 24% YoY in Q1 2022 since restricting access to the iOS device ID (IDFA), and 58% of all iPhone app downloads attributed to advertising came from Apple App Store search ads (up from 17% the year prior). App developers have no choice but to increase spending on app store advertising since Apple crippled their ability to advertise effectively on third-party advertising platforms.
We'll touch on Apple later in this post, but publishers would be wise to look for alternative solutions (like universal identifiers) even if they can use a device ID today.
Universal IDs can link CTV to everything
Universal identifiers are inherently "universal," meaning they will work across all platforms — this is in contrast to connected TV device IDs that are only relevant within the context of a CTV environment. Establishing a link between a user's connected TV device and the device they use to browse the web or complete purchases complicates targeting and measurement for ad platforms.
Let's say "[email protected]" signs into a streaming service — and that service converts this email into a universal identifier. The streaming service then shows an ad for a Peloton bike, and the user later checks out on Peloton's website with "[email protected]" as their contact email.
Peloton has a clear path to connect the purchase to the ad exposure — either through their DSP or by working with the publisher directly.
The DSP can collect inbound universal identifiers in bid requests, and publishers can retain ad exposure data by universal ID. The DSP or publisher can allow Peloton to upload a list of universal IDs representing "checked out" customers and compare the overlap of users who checked out against those who viewed an ad to attribute impressions to purchases.
If we run through the same scenario, but the streaming service does not convert the user's email address to a universal ID but instead relies on a connected TV device ID, things get a little trickier. It's not as easy to directly attribute a purchase to an ad exposure on connected TV as on the web or in-app.
The disassociation between the device where the ad impression occurred and the device where the purchase occurred requires ad tech providers to build or license a device graph to determine if a user owns both connected TV device 1234 that displayed an ad and mobile device 5678 that made a purchase.
There are two problems with device graphs. The first is that they are often probabilistic — meaning companies selling device graph solutions make educated guesses. The second problem is that they are somewhat user-hostile. Probabilistic (or deterministic) device graphs may link users' devices together in a database without explicit permission.
There is another way to tie connected TV to other devices, but it may not be the most user-friendly solution.
IP Address is not the answer
IP address solves the attribution use case described above. Ad platforms can use IP addresses as the connective tissue between the connected TV device where an ad displays and the mobile or desktop device used to make a purchase. If a user connects these devices to the same router in their household, both devices will emit the same IP address to the outside world.
An IP address can serve as a "universal" identifier of sorts since an ad request from any device in a household will share the same IP address. Advertisers and data brokers can collect purchase or browsing IP address data from a mobile or desktop device and then use that data to target a user on their connected TV.
Data brokers have many opportunities to collect IP address data since the identifier is inherently cross-platform. Many advertisers cannot resist the sheer scale and cross-platform capabilities that third-party audience data sets based on IP address offer — but the days of using an IP address as an identifier could soon come to an end.
I wrote an entire post about IP address data targeting if you want to learn more about the practice. But I also wrote another post about the future of IP address as an advertising identifier that examines the headwinds facing the usage of IP addresses in digital advertising along with current and upcoming efforts to restrict its use.
In that article, I detail why users and regulators may negatively view using an IP address as an identifier since it offers no privacy controls. Users can reset a device advertising identifier, clear their cookies, or opt-out of universal IDs — effectively breaking the link between collected data and the identifier/cookie. But users have no out-of-the-box capability to reset their IP address.
IP addresses are also less durable than universal identifiers - meaning they are not persistent. Internet Service Providers rotate residential IP addresses constantly.
When a household's IP address rotates, it becomes useless until it expires from a data source. It can even cause an advertiser to target or measure the wrong person or household in the meantime. Universal identifiers do not face this problem as long as a user continues to create accounts with the same email address.
Speaking of households, another knock against IPs is that they can cause comically bad mistargeting since it is a household-level identifier. A family with all their devices connected to the same home Wi-Fi network will emit the same IP address to data collectors — so if Dad is super into crypto, then the whole family might be subject to weird crypto.com ads starring Matt Damon.
These problems could happen on CTV with universal IDs since families typically use one streaming account (tied to one email) with multiple profiles for the whole family. Streaming services could mitigate this risk by requiring an email for each profile on an account.
Companies can also surreptitiously collect IP addresses without the user's consent. Universal identifiers can offer a transparent and private opt-in model if implemented conscientiously, as publishers and advertisers can explain to users how they use their personal information.
Some universal identifier solutions can ensure that no personally identifiable information will exchange between third parties (only obfuscated identifiers). Additionally, any universal identifier solution worth implementing should offer clear opt-out mechanisms for users.
It is an understandable tradeoff for users to recognize that ad-supported services can use collected data for marketing purposes as long as they anonymize and protect it. Reputable publishers or advertisers should explain this during email address collection.
Companies will continue to use IP address as an identifier for targeting and measurement so long as it remains readily available. IP address gains even more value in environments like Free ad-supported TV (FAST) apps (Pluto, LG Channels, Samsung TV Plus, etc.) that allow users to access linear programming on a connected TV without requiring a login. Without a user account email, there is no opportunity to create a universal identifier.
Universal Identifiers on Mobile in-app
We learned that universal identifiers could solve audience targeting and measurement on CTV — so they can probably do the same for mobile app environments, right? Technically, they could, but Apple and Google's mobile operating system duopoly might not allow universal identifiers to achieve their maximum benefit potential.
Apple / iOS
Let's cut to the chase — if a user opts out of tracking via the App Tracking Transparency (ATT) Framework, app developers cannot use a universal identifier in place of an IDFA.
If you don't know, Apple ATT is that fun prompt that asks users if they want an app to "track your activity" (it's a wonder that even 25% of users opt-in after such a foreboding question). App developers cannot access the iOS device ID (IDFA) if the user declines.
Like Thanos snapping his fingers, Apple erased 75% of the iOS user base from personalized advertising delivered via device IDs by implementing ATT. This little prompt eviscerated $230 billion from Meta's market cap in a single day. It also caused significant market value damage to many other companies with a vested interest in delivering personalized advertising or measuring performance with attribution.
These companies wanted to implement workarounds for this newly introduced problem, and universal identifiers could have offered a reasonably straightforward solution — but Apple responded with a resounding "NOPE."
Not only will Apple not allow access to an IDFA if a user opts out via ATT, but they also do not allow apps to use workarounds (hello universal IDs!) that could link users to third-party data sets.
If there was any doubt, here is the language on their official page detailing their rules on user privacy and data use:
You need to receive the user’s permission through the AppTrackingTransparency framework in order to track them or access their device’s advertising identifier. Tracking refers to the act of linking user or device data collected from your app with user or device data collected from other companies’ apps, websites, or offline properties for targeted advertising or advertising measurement purposes. Tracking also refers to sharing user or device data with data brokers.
The whole point of a universal identifier is to link user data collected from an app (email) with user data collected from other companies for targeted advertising or measurement purposes. So this section can be short and sweet because there's not much else to say.
The only silver lining for universal identifiers on iOS is if the user allows tracking. Ad-supported app developers could access a device IDFA and create a universal identifier without violating Apple's policy. Sending multiple identifiers in an ad request can increase the odds of potential advertisers finding user matches on audience-based buys and successfully measuring attribution.
So universal identifiers do not solve device ID loss resulting from ATT opt-outs, but they could make an app's opted-in user base even more valuable.
Google / Android
So what does Google think of universal identifiers on Android? Unclear.
We know that Google announced the Privacy Sandbox on Android — a set of features meant to limit the need for sharing user data or device identifiers to maintain advanced advertising features. From the announcement:
Today, we’re announcing a multi-year initiative to build the Privacy Sandbox on Android, with the goal of introducing new, more private advertising solutions. Specifically, these solutions will limit sharing of user data with third parties and operate without cross-app identifiers, including advertising ID.
Well, they did say cross-app identifiers and didn't specifically say universal IDs — but it does not sound great. It does make clear that the days of readily accessible Android device IDs are numbered:
we plan to support existing ads platform features for at least two years, and we intend to provide substantial notice ahead of any future changes.
It's unclear from this announcement if Google will introduce a similar prompt like Apple ATT or completely deprecate Android advertising IDs (typically referred to as AAID or GAID). Google also did not divulge where they stand specifically on universal identifiers.
Google has the unique challenge of balancing antitrust concerns with privacy concerns. This challenge is the same predicament they find themselves in with deprecating cookies and may be why they keep delaying that decision. I wrote this tweet to summarize that situation:
Google is damned if they do and damned if they don’t deprecate cookies.
Do: Google is unfairly hurting all ad tech companies with their monopolistic browser power.
Don’t: Google hates users and doesn’t care about their privacy. https://t.co/7NqPpuP64g— Ad Tech Explained (@AdTechExplained) August 1, 2022
Google earns over 25% of all digital ad revenue in the US and controls the most used web browser (Chrome) and mobile operating system (Android) worldwide (but notably not in the US). So it has an enormous competitive advantage over everyone, and any decision they make will surely draw the attention of regulators trying to maintain a level playing field.
Google banning the usage of universal identifiers as an Android policy could limit competition in digital advertising. Google's massive logged-in user base and ownership of the two major vectors we use to enter our digital lives (browser and mobile OS) can allow them to flourish if they remove device IDs and prohibit universal identifiers.
The rest of the digital advertising industry won't stand much of a fighting chance if Google bans universal identifiers, deprecates third-party cookies, and eliminates Android device IDs.
The universal question
Universal identifiers offer a clear set of advantages on CTV for:
Publishers who want to juice the value of their audience
Advertisers who desire to target and measure their ad campaigns more precisely
Publishers, advertisers, and ad platforms who wish to future-proof their businesses
Although the incentives aren't as clear cut on mobile in-app, embracing universal identifiers on iOS or Android could, at the very least, increase the value of opted-in users.
Universal identifiers will offer immense value on the web if Google ever gets around to finally deprecating third-party cookies in Chrome — but they provide immediate value in browsers like Firefox and Safari that have already pulled the plug on third-party cookies.
Any company running a sizeable digital advertising business must now answer a universal question: Should we adopt a universal identifier solution?
Even with Google perpetually delaying its decision to remove third-party cookies from Chrome, it might make the question a little easier to answer if decision-makers understand that the potential benefits of universal IDs extend beyond web environments.
Reply