Apple’s deprecation of the IDFA mobile ID and Google’s deprecation of third party cookies in Chrome have rightfully caused the advertising industry to question assumptions. For years, advertisers have relied on these device identifiers for almost every targeting and measurement methodology, including behavioral, attitudinal, location and demographic.
Beginning with the launch of Apple’s iOS 14.5 in late April, the number of ad opportunities that will include targetable device IDs could decline sharply. Estimates for the loss of scale vary, but we generally expect opt-in rates for user tracking on apps and sites to be on the low side of these estimates, perhaps 20-30%. Addressable programmatic scale could decrease by more than 25% in just a matter of weeks.1 As iOS adoption increases and Google’s changes take effect, we could start to see a dwindling minority of programmatic inventory that includes the device IDs required for traditional targeting and deterministic match-based measurement.
Emodo provides a unique suite of non-ID based solutions. Our portfolio of capabilities combines the best industry ID-based solutions with an array of our own innovative options. We offer a portfolio of solutions because our goal is to enable the same levels of scale and efficiency marketers have come to expect from digital advertising. Given the scope of change, that’s a high bar and there is no single solution that achieves that goal.
Inevitable Change Across the Board
Changes to identity will affect all players, from walled gardens, to open web programmatic platforms, to publishers of all types, to marketers. Even if walled gardens have high registration and relatively high opt-in rates, a low rate of opt-in on network inventory will impact their ability to monetize inventory outside of their owned and operated media. When marketers run campaigns on that owned and operated inventory, they should expect to run into measurement challenges when users don’t opt in to tracking on the marketer’s site or app. Publisher opt-in rates will vary, but the universal tracking scale of cookies and IDFAs will no longer be available.
Much of the industry chatter has focused on replacing the third party cookie and IDFA with alternative ways to deterministically track users and match them to databases, such as the UID 2.0 initiative that has garnered a wide range of adoption already. These new identity structures will play an important role. Emodo is actively working to align with the alternative identity platforms that present the greatest opportunity for scale. That said, it’s critical for marketers to keep in mind that all of these Identity structures are by definition opt-in. That requirement is fundamentally different from the opt-out identity structures on which the industry has become dependent. No matter how creatively or aggressively publishers entice their users to opt-in, there will always be a substantially lower percentage of user opt-in, and thus less programmatic inventory that’s addressable and measurable based on an identifier.
Deterministic / First-Party Data
Even if marketers’ or platforms’ deterministic databases remain sizable, no amount of data about any user is actionable for advertising unless that user has opted into tracking on the particular app or site. Databases will no longer be measured and valued simply by their scale. The addressability of the database will suddenly be paramount. That’s a serious shift that has remained unrecognized or undiscussed across the industry.
Supply and Demand
The sharp decline in addressable inventory comes at a particularly challenging time. As consumers emerge from the COVID-19 pandemic, many marketers are looking to ramp up their advertising spend. Basic supply & demand economics suggest that prices for addressable inventory will increase substantially. Not only will scale decrease, but the cost of doing “business as usual” will go up.
What Can Marketers Do?
Marketers have a lot to consider. Effectively targeting audiences and measuring campaigns in a world without device IDs or third party cookies is a significant, immediate and complex challenge. And, there is no single solution.
Emodo’s Portfolio Approach
Many vendors are trying to recreate a legacy system based on deterministic data matching, albeit opt-in. We learned from GDPR that a system that requires users to opt-in has great merit, but it has no chance of obtaining ubiquity. That’s why Emodo is taking a “portfolio” approach – offering a more encompassing set of solutions that enable effectiveness, scale and consumer privacy. Our portfolio includes four primary solutions:
- Implementation of the new alternative identity structures that earn wide adoption and meaningful scale, such as UID 2.0
- Redefined Contextual Targeting beyond just words on a page
- Predictive Audiences and other machine learning-based approaches to predict audience attributions in real time
- A new world of immersive consumer experiences such as Augmented Reality
Not Your Parents’ Contextual Targeting
You may have heard that “contextual targeting is making a comeback”. When you think about old-school contextual targeting, it kind of makes sense. You don’t need an ID to deliver an ad based on the words on a certain page. Today, however, context takes on a broader definition. It can include geography, weather, publisher and category, inventory source, device OS version, device movement patterns and much more. Strong contextual targeting is dependent on having the industry’s most accurate truth sets – including data from mobile networks.
It’s important to note that the addition of new capabilities less reliant on device identity doesn’t have to mean reduced efficiency or revised campaign objectives. Many KPI optimization models don’t need ID to relentlessly drive efficiency. Engagement, clicks, VCR (Video Completion Rate), LTR (Listen-Thru Rate), time spent, IVT, and others don’t require an ID to optimize.
Even inventory that has device IDs can be fine-tuned with the application of similar machine learning models. The best designed and responsibly built models will be a key differentiator that can help maintain marketing cost and efficiency as the cost of ID inventory rises.
Most importantly, Emodo Predictive Audiences offer an effective scale enhancer that enables Emodo clients to continue executing highly targeted campaigns at the scale to which they’re accustomed. With the increasing price premium for ID-based inventory, predictive targeting can drive efficiency at scale on a cost-adjusted basis by leveraging hundreds of real time data points to predict the likely audience designations for a given programmatic ad opportunity. Emodo Predictive Audiences are built using machine learning models trained with data that is privacy-friendly, representative, accurate, and thorough – Emodo’s core expertise.
Machine learning-based approaches have other key benefits to Emodo’s clients. Training datasets don’t have near the scale or usage rights requirements necessary to build a legacy consumer database for deterministic matching, so they can be much more privacy friendly. Machine learning also enables marketers to focus on the quality of data rather than chasing the waning benefits of data quantity. Massive data inaccuracies and data fraud will become lesser issues as the industry leans more toward responsibly-built training data sets.
In some ways, opt-out based identity has driven marketers to be overly focused on audience characteristics, and lose sight of the importance of consumer experience that has driven advertising on broadcast media for generations. The timing is great for this shift in identity because of the iPhone 12 launch bringing 5G into its prime. 5G makes an entirely new world of immersive consumer experiences possible at unprecedented scale, most notably through the use of Augmented Reality (AR). Emodo’s proprietary research shows 70% of 5G phone owners are actively asking for more AR advertising, report that AR ads reflect well on a brand, and view AR ads as more interesting and engaging than normal ads. Because of Emodo is a subsidiary of Ericsson, the infrastructure company powering 5G for every US carrier, we’re uniquely positioned to build successful, immersive AR and 5G ad campaigns in ways not possible elsewhere, and the best news is the most immersive experiences span Audience segments and traditional digital targeting criteria.
The loss of device identity presents challenges in every corner of the market. You can’t continue targeting only devices that have IDs without significantly reducing your reach and efficiency. Heavy investments in first-party data will produce diminishing returns as a growing portion of those devices become unmatchable in the bid stream. The Walled Gardens will be impacted like the rest of the market – with rising costs and reduced addressable scale.
Emodo will continue to provide the level of efficiency and scale that’s become synonymous with digital marketing. We’ll continue to do that by offering a unique, powerful array of both ID-based and non-ID solutions. Our non-ID portfolio will include support for the most widely adopted alternative ID structures, modern contextual targeting techniques, immersive creative opportunities and predictive audiences – our proprietary machine learning models trained by the highest quality, privacy-friendly data sets.
When you work with Emodo, we put a broad portfolio of non-ID options to work, so you can continue to reach the audience you’re looking for at scale, with or without device identifiers.
1 Estimates based on historical adoption patterns: iOS 12 and 13 reached 50% penetration amongst existing iOS users in about a month. Emodo platform data shows that iOS devices represent slightly more than half of programmatic inventory.