Addressable advertising is the default in the digital display world—and it relies on being able to identify users in order to serve them the right message at the right time. This eMarketer Report explains three significant disruptions developing this year in how marketers identify and target users in digital environments and discusses tactical and strategic approaches for the future. IDENTITY 2020 Changes to Cookies, Ad IDs, and Regulations Take Aim at Tracking presented by
Transcript
Addressable advertising is the default in the digital display world—and it relies on being able to identify users in order to serve them the right message at the right time. This eMarketer Report explains three significant disruptions developing this year in how marketers identify and target users in digital environments and discusses tactical and strategic approaches for the future.
IDENTITY 2020Changes to Cookies, Ad IDs, and Regulations Take Aim at Tracking
eMarketer is pleased to make this report, Identity 2020: Changes to Cookies, Ad IDs, and Regulations Take Aim at Tracking, available to our readers.
This report is a great example of eMarketer data and insights, including analysis of three significant disruptions developing this year in how marketers identify and target users in digital environments, as well as tactical and strategic approaches for the future.
We invite you to learn more about eMarketer’s approach to research andwhy we are considered the industry standard by the world’s leading brands,media companies and agencies.
We thank you for your interest in our report and Verizon Media for making it possible to offer it to you today. Best Regards,
Nancy Taffera-SantosNancy Taffera-SantosSVP, Media Solutions & Strategy, eMarketer
eMarketer, Inc.11 Times Square, Floor 14New York, NY 10036
IDENTITY 2020: CHANGES TO COOKIES, AD IDS, AND REGULATIONS TAKE AIM AT TRACKING
Addressable advertising is the default in the digital display world—and it relies on being able to identify users in
order to serve them the right message at the right time. But the identifiers that marketers and their partners use to
do this are coming under threat as platforms and regulators work to improve data privacy and protection practices
for consumers.
What are the major disruptions this year to digital advertisers’ identity practices?
At the beginning of the year, Chrome announced it was killing the third-party tracking cookie. Halfway through the year, Apple announced it was essentially doing the same to its Identifier for Advertisers (IDFA), a device-specific ad ID. And that’s on top of the California Consumer Privacy Act (CCPA) going into effect and requiring marketers to let California residents opt out of having their data sold.
How are marketers and their partners responding to identity challenges?
They’re going to have to do so both tactically and strategically. Many are looking to ways to “replace the cookie”—which could be a doomed proposition if consumers don’t want to consent to being tracked and targeted by advertisers. More far-reaching changes include moves toward other forms of measurement and attribution, like panels.
How important is identity for advertisers today?
Very. US advertisers spent billions last year on data activation solutions and are investing across a range of tools that all rely (at least in theory) on being able to tell who a user is across devices and platforms. Research suggests many are spending more than 20% of their marketing budgets on identity solutions.
WHAT’S IN THIS REPORT? This report explains three significant disruptions developing this year in how marketers identify and target users in digital environments and discusses tactical and strategic approaches for the future.
% of respondents, May 2020
How Will the Change in Support for Third-PartyAudience Cookies Affect the Use of Data for DataProfessionals in North America?
Increase spending/emphasis on use of first-party data60.4%
Increase interest in third-party identity resolution solutions37.7%
Engage with third-party industry groups to build identityresolution solutions
35.9%
Increase efforts to build second-party data relationships35.9%
Increase focus on developing custom/in-house identityresolution solutions
35.9%
Increase spending/emphasis on contextual advertising34.0%
Increase testing/investment in new solutions32.1%
Source: Winterberry Group, "The State of Data 2020" in partnership withInteractive Advertising Bureau (IAB), July 22, 2020257477 www.eMarketer.com
KEY STAT: The industry is definitely heading toward a cookieless future, but fewer than one-third of data professionals in North America were actively testing and investing in new solutions as of May, per Winterberry Group and the Interactive Advertising Bureau (IAB).
CONTENTS2 Identity 2020: Changes to Cookies, Ad IDs, and Regulations
Take Aim at Tracking
3 Marketers Make Big Efforts to Identify Audiences
Over the past decade, marketers have widely adopted
addressable advertising to target audiences interested
in their messages. That means they need to be able
to find—that is, identify—users of various devices, so
they can make use of information about those users
to send the right message to the right person at the
right time. That identification process has long been
a challenge, as marketers and their partners have had
to contend with a growing set of connected devices
that were never set up to share a universal identifier.
But industry developments in 2020 are set to make it
even more difficult.
A Note on Terms
The identity of a user is all of the information about that user, bundled together into a profile.
Identity is a distinct concept from identifiers—such as cookies, hashed email addresses, website logins, device IDs like the IDFA and similar identifiers in connected TV (CTV) apps, and so on.
Identity resolution is the process of connecting those identifiers when they refer to the same person and putting together all the data associated with them. Identity resolution is made possible by identity graphs.
In the research for Winterberry Group and the IAB’s “State of Data 2020” report, conducted in May, they asked several questions about what issues would be top of mind for respondents this year:
■ About one-third (34.0%) of data professionals surveyed in North America said compliance with data regulation would be taking up their attention this year, and three in 10 (30.2%) said they would be working on the development of a custom identity resolution solution. About a quarter (26.4%) were spending time on cross-device identity capabilities, and the same share said they were looking at the development of an industry-standard ID resolution solution.
■ About one-third (34.0%) said the decline of cookies and the change in availability of third-party data were among the obstacles and functional challenges that would occupy their attention and resources in 2020. About one in five (18.9%) also cited government and data privacy regulations.
■ More than two in five (42.0%) said identity resolution/cross-channel identification was a first-party data use case that would occupy their time, attention, and resources this year. Many of the other first-party data use cases marketers had in mind rely in turn on cross-channel identification and tracking (at least, as normally practiced today).
% of respondents, May 2020
Which First-Party Data Use Cases Do DataProfessionals in North America Expect Will OccupyTheir Time, Attention, and Resources During 2020?
Source: Winterberry Group, "The State of Data 2020" in partnership withInteractive Advertising Bureau (IAB), July 22, 2020257476 www.eMarketer.com
Marketers are spending significant amounts of money on their identity problems. Research published in July 2020 by Winterberry Group and IAB found that US marketers spent $7.8 billion on audience data activation solutions last year. That was up from $7.2 billion in 2018 and $6.2 billion in 2017.
According to November 2019 research from performance agency Merkle, about half (49%) of US and UK marketers reported spending more than 20% of their budget on identity solutions.
% of respondents
Percentage of Their Company's Marketing SpendingAllocated to Identity Solutions According to UK/USMarketers, Nov 2019
0%-5% 6%
6%-10% 7%
11%-15% 11%
16%-20% 26%
21%-25% 29%
>25% 20%
Unsure/don't know3%
Note: numbers may not add up to 100% due to roundingSource: Merkle, "Q1 2020 Customer Engagement Report," Feb 18, 2020253220 www.eMarketer.com
These budgets are going to a variety of services and solutions. The Merkle research found that at least half of US respondents were planning to invest in each of five customer-identity-related areas in 2020, including data management platforms (DMPs), identity resolution and identity graphs, and databases or data lakes.
% of respondents in each group
Investments Specific to Identifying Customers thatSenior- and Executive-Level Marketers Worldwide WillMake, by Region, Nov 2019
Consumer data—better consumer understanding and basis for targeting and/or analytics
Data management platforms for digital identity management and activation
Customer data platforms for digital identity management and activation
Identity resolution/graph—mastering persistent identity of customers and/or consumer population and for matching and onboarding for targeting and analytics
Database or data lake—a secure environment for data collection, integration, management and analytics
None of the above
Asia-Pacific
65%
57%
55%
57%
55%
49%
0%
Europe,Middle East
& Africa
64%
68%
60%
62%
50%
56%
2%
US
69%
63%
59%
43%
53%
51%
2%
Total
66%
63%
58%
54%
53%
52%
1%
Note: in the next year; numbers may not add up to 100% due to roundingSource: Merkle, "Hyper-Personalization and the Connected CustomerExperience," Jan 15, 2020252494 www.eMarketer.com
It’s no surprise that marketers are devoting a lot of attention and resources to identity when it underpins much of what they do with digital advertising today. Two-thirds of ad tech professionals worldwide surveyed in February 2019 by ID5—builders of a shared identity infrastructure for ad tech—said user identification was “extremely important,” with another 22% saying it was “very important.” Respondents were split on the biggest limitation of poor user ID, with similar shares pointing to difficulties with retargeting users, reconciling data from other sources, and accessing third-party data.
Identity challenges are significant in cross-screen video advertising. The lack of a universal cross-screen identifier is a factor in several of the problems US agency and marketing professionals were concerned about in November 2019, including managing audience deduplication, managing ad frequency capping, and executing cross-screen data-targeted buys, per research firm Advertiser Perceptions.
% of respondents
Leading Challenges to Cross-Screen Video AdvertisingAccording to US Agency/Marketing Professionals, Nov 2019
Lack of standard measurement57%
Managing audience deduplication55%
Managing ad frequency capping49%
Managing creative executions, format standardization46%
Inability to execute a cross-screen data-targeted buy43%
Inability to plan media buys seamlessly40%
Note: figures reflect the sum of respondents who listed each challengeamong the top threeSource: Advertiser Perceptions, "Video Advertising Convergence 2H 2019,"Feb 27, 2020253306 www.eMarketer.com
For more on the difficulties of executing and measuring video advertising across screens, read our May 2020 report, “TV and Digital Video Measurement 2020: Addressing Cross-Platform Challenges.”
Marketers use a wide variety of identifiers and other information to help figure out who their customers are across devices and touchpoints. The most widely used piece of information among US companies is a customer’s email address, according to October 2019 Forrester Consulting research commissioned by Epsilon-Conversant (which provides a range of marketing data solutions). Email addresses act as a popular identifier for several reasons, including that they are widespread and reliably refer to the same individual over long periods of time. Phone numbers, used by just under half of respondents, have similar advantages (and with the rise of mobile telephony, are now likely to be unique individual identifiers, as opposed to being shared by a household).
% of respondents, Oct 2019
What Do US Companies Use to Help Identify TheirCustomer Across Devices and Touchpoints?
Note: n=205 decision-makers who influence marketing and/or datatechnology decisionsSource: Forrester Consulting, "Is Your Identity Program Built on a House ofCards?" commissioned by Epsilon-Conversant, Oct 2, 2019250318 www.eMarketer.com
Many marketers are also using identifiers that are being disrupted, as outlined in this report. More than two in five respondents said they used cookies to identify customers. The Forrester/Epsilon-Conversant research didn’t ask about IDFA usage specifically, but it did find that 33% of US companies said they used ad IDs such as the Google Advertising ID (GAID) and IDFA to identify customers across devices and touchpoints.
Additionally, according to July 2019 research from video ad platform LoopMe and Sapio Research, almost half (47%) of US and UK marketers were using cookies.
One of the key takeaways from the Forrester/Epsilon-Conversant research is that companies today must use many different pieces of information to figure out user identity. That’s because of fragmentation—one of the market factors more than one-third of US marketers in the Merkle survey said were holding them back from managing customer identity in-house. Disparate identity spaces and other market issues, like the prevalence of walled gardens, were the most commonly cited barrier to in-house identification.
% of respondents in each group
Barriers to Managing Customer Identity In-HouseAccording to Senior- and Executive-Level MarketersWorldwide, by Region, Nov 2019
Note: numbers may not add up to 100% due to roundingSource: Merkle, "Hyper-Personalization and the Connected CustomerExperience," Jan 15, 2020252493 www.eMarketer.com
Advertisers aren’t the only ones trying to identify users, either. This capability is important for publishers as well, in part to help them monetize their inventory more effectively by allowing for targeted advertising, but also for their own marketing and content development.
Kerel Cooper, senior vice president of global marketing at programmatic email advertising firm LiveIntent, explained that many of the company’s publisher clients are mining insights from email newsletters—which are sent to identified, consenting users. “They’re looking at how they can create more of a holistic approach that includes content, growing subscribers, and effective ad placement,” he said. “They know that the end of the cookie is coming, and publishers need to figure out what their first-party data strategy looks like.”
IDENTITY CRISIS 2020
The ways in which advertisers and their partners
can identify and track users across devices have
been changing for years. But 2020 has brought two
major developments sending shockwaves through
the digital advertising industry around the world:
Chrome’s deprecation of support for third-party
cookies and major changes to the IDFA set to take
effect in iOS 14. Another major disruptor occurred
on July 1 this year, when the California Consumer
Privacy Act (CCPA)—the first major US data privacy
and protection regulation to tackle consent in digital
advertising—began enforcement.
CHROME ANNOUNCES THE DEATH OF THE THIRD-PARTY TRACKING COOKIE In January 2020, the team that works on Google’s Chrome browser announced that it would “phase out support for third-party cookies in Chrome” within the next two years. Other browsers have already made this change, including Apple’s Safari, with Intelligent Tracking Prevention (ITP), and Mozilla’s Firefox, with Enhanced Tracking Protection (ETP).
The cookie’s days had long been numbered; consumers have been spending increasing amounts of time in cookieless environments for years. We estimate that mobile apps have accounted for a majority of smartphone time since at least 2008, the earliest date in our forecast. As smartphone adoption increased and mobile began to account for a significant share of internet time, so too have apps. We estimate that this year, the average US adult will spend almost 48% of their digital media time with mobile apps—more than a quarter of total media time. The rise of CTV has meant only more digital time spent in cookieless environments.
But it’s been clear to any website visitor bombarded with cookie opt-in requests since the EU’s consumer data protection regulatory framework—the General Data Protection Regulation (GDPR)—took effect in May 2018, that advertisers, publishers, and their partners have remained partially reliant on cookies to track users across the web, serve them targeted advertising, and attribute the results of that advertising.
Even after ITP made it impossible to track Safari users’ web browsing via third-party cookies, advertisers didn’t do much to find an alternative; reports of dramatically falling prices for Safari impressions have been widespread. We estimate 45.6% of smartphone users in the US are iPhone users, though they may use browsers other than Safari.
“Earlier in the year, I had reached out to our sales team and asked them to go back to the buy side and ask, ‘What are you seeing? Are you concerned about this?’” said Premesh Purayil, chief technology officer at programmatic yield partner Freestar. “We saw a huge drop in performance on iOS in the past year. The somewhat surprising response, fairly unanimously, was that [those on the buy side] weren’t too worried, because from their point of view, their KPIs [key performance indicators] were still hitting. It’s almost like they’re waiting for this problem to potentially hit their doorstep or just get figured out for them.”
Chrome’s announcement may have made the problem impossible to ignore. At the very least, it created a new sense of urgency, especially within the programmatic advertising ecosystem. Soon after, IAB announced its Project Rearc initiative to develop technical standards for collecting and using whatever identifier ends up being tapped for addressable advertising in the absence of third-party tracking cookies.
Other organizations are working on developing what those identifiers might be and implementing them—more discussion of potential solutions is in the next section.
CCPA IS ENFORCEABLE, AND CPRA IS UP FOR REFERENDUM In July 2020, CCPA became enforceable, with some clarifications to the final language of the regulation released in August. The law placed new requirements on companies to allow consumers to opt out of sharing their information, including a requirement to respect a global opt-out mechanism—like a browser-based “do not track” setting.
In a March 2020 poll by Forrester Consulting for promotions eligibility verification provider SheerID, about a third of US business-to-consumer (B2C) marketers reported being “fully compliant” with privacy regulations including the CCPA and the EU’s GDPR. The bulk were still working on the privacy compliance—and of course, as of March, no one knew what the final CCPA rules would be.
% of respondents, March 2020
How Prepared Are US B2C Marketers to Meet theRequirements of Privacy Regulations?
Fully compliant 31%
Currently rolling out privacy plan 31%
Developed, but not rolled out 34%
Not at all prepared1%
Don't know2%
Note: n=210; includes CCPA and GDPR; numbers do not add up to 100%due to roundingSource: Forrester Consulting, "Put Identity Marketing In the Center of YourPersonalization Initiatives" commissioned by SheerID, May 28, 2020255818 www.eMarketer.com
There are many differences between CCPA and GDPR, but we can look to Europe for some indicators of the effects of this type of data privacy and protection regulation. According to May 2019 research by privacy compliance software firm TrustArc and Ipsos MORI, 25% of UK internet users surveyed reported having opted out of a request to install cookies under GDPR. Almost half of those polled said they had exercised their rights under the regulation in some form or fashion.
In our July 2019 report, “Digital Marketing in Today’s Privacy-Conscious World: What Companies Need to Know About GDPR, CCPA and Other Industry Changes in the Next 12 Months,” we outlined several significant changes in digital advertising practices due to GDPR, including:
■ Companies were placing greater emphasis on their data infrastructure.
■ Customer databases and prospecting data sets were smaller.
■ The publisher landscape shrank—and the price of impressions (with consented data) went up.
■ Marketers (and publishers) got smarter and choosier about audience targeting.
■ The customer experience mostly didn’t improve.
Our November 2019 update, “UK Programmatic Digital Display Ad Spending: Growth Underpinned by Social Spend and Programmatic Direct,” noted that GDPR hadn’t been much of a growth inhibitor in terms of programmatic display ad spending, once the initial disruption was past—though it had helped steer more euros to private marketplaces (PMPs) and programmatic direct deals.
But, as that report also noted, by late last year it was clear that European regulators, including the Information Commissioner’s Office (ICO) in the UK, had their sights on real-time bidding (RTB). The ICO published a blog post in December that said, “Some of what is happening now appears to be unlawful, based upon the evidence we have seen to date.”
Major sticking points include the fact that bid requests are broadcast far and wide, without controls on who can access and then process the information in those bid requests, which include personal data. Websites can’t ask consumers to consent to uncontrolled data leakage. We predicted in our January 2020 report, “UK Key Digital Trends 2020: Privacy, Brexit and Video Take Center Stage,” that enforcement against the advertising industry, and specifically ad tech, would begin this year. That hasn’t materialized yet—perhaps a pandemic got in the way—but we don’t think this problem is going to just disappear.
What’s the problem with RTB and personal data protection? Listen in to this August 2020 episode of our podcast “The Ad Platform,” where Johnny Ryan discusses GDPR and the principles of data protection regulation. Ryan recently left the ad tech world for two nonprofits and is focusing his attention on the regulators he says have failed to enforce the law and protect consumers.
This November, voters in California will decide whether to pass Proposition 24, the California Privacy Rights Act (CPRA), which would build on the CCPA by adding restrictions on data use and consumer data rights, as well as creating a California Privacy Protection Agency to enforce the law rather than the state attorney general.
A poll conducted in late July 2020 by Goodwin Simon Strategic Research found that 81% of likely voters in the state supported the proposition. The survey then presented respondents with language from the official voter ballot guide that included arguments against the CPRA, and 72% still said they would vote to pass it.
The CPRA puts a spotlight not just on “selling” consumer data, which is what the language of the CCPA refers to, but also on “sharing” it. Marketers must be ready for more and more stringent regulation of these activities in the US as well as elsewhere in the world.
THE IDFA GOES (MOSTLY) AWAY In June 2020, Apple announced at its Worldwide Developer Conference (WWDC) that the iOS 14 update, coming in September, would involve major changes to how IDFA works. (In early September, Apple announced a reprieve until “early next year” before the IDFA- and privacy-related changes will go into effect.)
The identifier would become opt-in on a per-app basis, meaning users would have to consent to share their IDFA with each app they use, through a pop-up that will display when the app opens. The operating system will also dictate the nature of the pop-up and the language it can use. Blog responses from the mobile app community fell on a spectrum from “Apple killed the IDFA” to “IDFA is not dead yet, but it’s definitely on life support.”
According to July 2020 polling among US agencies and brand marketers for Pivotal Research Group’s investor update, there was widespread uncertainty among the advertiser community at the time as to how these changes would affect major digital ad sellers. At least half of respondents said they were unsure how the iOS 14 update would affect the businesses of Snapchat, Twitter, and The Trade Desk, while 45% said the same about Facebook.
% of respondents, July 2020
Do US Agencies and Brand Marketers Perceive thatSelect Advertising Platforms Will Be NegativelyImpacted by IDFA Changes*?
Yes No Not sure
Facebook 30% 25% 45%
Snapchat 9% 36% 55%
Twitter 19% 30% 52%
The Trade Desk - 50% 50%
Note: numbers may not add up to 100% due to rounding; *Apple's IDFA(Identifier for Advertisers) tool will require opt-in consent on a per-appbasis beginning in iOS 14, which will launch in Sep 2020Source: Pivotal Research Group, "Views on Q2 EPS/Q3 Outlook, and DeepDives Into Key Areas of Investor Debate," Aug 12, 2020258220 www.eMarketer.com
Across all four ad sellers, direct-response-focused advertisers were dramatically more likely to be unsure of the effects of the changes than branding-focused advertisers. For example, 67% of direct-response advertisers weren’t sure how the changes would affect Facebook, compared with just 27% of respondents focused on branding. Direct-response advertisers are more likely to be reliant on advanced attribution use cases that rely on an IDFA or similar identifier. Many have likely spent the past two months educating themselves more on this issue—and so have their ad tech partners—so it’s possible these numbers would be different if the survey were fielded today.
Part of the uncertainty relates to a lack of understanding of Apple’s new app tracking policy. Some of the language currently available in Apple’s marketing materials suggests that developers must get opt-in consent for not only the IDFA, but also any type of user tracking across apps.
There’s also a lack of understanding of the SKAdNetwork, the Apple-run application program interface (API) intended to help advertisers measure the success of app install campaigns. SKAdNetwork has existed since 2018 but has gotten dramatically more attention now that it “could become the de facto way to attribute mobile app installs on Apple devices,” as AdExchanger’s Allison Schiff explained in July.
And there’s uncertainty about how many users will choose to opt in to IDFA tracking—but most of the experts interviewed for this report believe the opted-in share of users will be low, around 20% or less. For the rest, there will be workarounds (more on that later) and the SKAdNetwork.
The SKAdNetwork is not set up to be a replacement for current attribution practices, however. It doesn’t allow for sophisticated reporting of multiple post-app-install events, and it’s not based on user-level tracking and reporting.
“There are still many open questions about how SKAdNetwork will work and how the changes will affect app developers and the mobile ecosystem,” said Brian Quinn, US president and general manager at mobile analytics and attribution provider AppsFlyer. “At its core, the SKAdNetwork presents new functional challenges for the advertiser, including no real return on investment/lifetime value, a lack of granularity (campaign-level only, limited to 100 campaigns), postback delay (at least 24 hours), data ownership and reporting is done by networks, data manipulation (which leads to ad fraud risk), and no reengagement attribution support. Additional structural changes are also present, i.e., postbacks are sent only to the attributed ad network; therefore the advertisers, or anyone processing data on their behalf, are blind to it.”
In late August 2020, a Facebook announcement may have brought more clarity to the question of how its business would be affected. The company explained in a blog post that it would not be asking for the IDFA in its own apps and warned that it expected the change could “severely impact” the ability of publishers participating in the Facebook Audience Network (FAN) to monetize their inventory. Earlier this year, Facebook sunset FAN for the mobile web, which the industry understood to be a response to the ongoing deprecation of third-party tracking cookies. The August blog post telegraphed that the same fate may await FAN generally on iOS devices.
Facebook reported that it tested mobile app install ads without “personalized ranking” and found they can lead to a more than 50% decrease in publisher revenues. Those ads were served to users based on targeting criteria like demographics—so that result still involved identifying users and targeting them. In other words, no ability to target ads based on user identity could lead to even worse results for publishers.
iOS already has an option to Limit Ad Tracking (LAT) in apps, and inventory with LAT turned on is much cheaper, according to our sources.
The advertisers surveyed by Pivotal Research Group were split on whether Apple would eventually do away with the IDFA altogether—which some in the industry thought may happen this year. Direct-response-focused advertisers were only slightly more likely to think this than brand advertisers.
% of respondents
US Agencies and Brand Marketers Who Believe thatApple Will Eventually Eliminate the IDFA*, July 2020
Branding48% 52%
Direct response52% 48%
Total50% 50%
Yes No
Note: *Identifier for Advertisers tool on iOS that allows advertisers to trackand identify a deviceSource: Pivotal Research Group, "Views on Q2 EPS/Q3 Outlook, and DeepDives Into Key Areas of Investor Debate," Aug 12, 2020258221 www.eMarketer.com
But given the expected low opt-in rates for the IDFA going forward, perhaps the better question at this point is whether (or when) Google will follow suit. Several experts interviewed for this report pointed out that Google is a “follower” in this arena, with Chrome announcing the deprecation of third-party cookies within a couple years of other browsers making the same move. The GAID’s days may be numbered—but Google is likely to provide a more advertiser-centric solution than the SKAdNetwork if so.
IDENTIFYING A PATH FORWARD
One of the biggest effects of the slow demise of
the cookie has been an increase in the reliance on
first-party data. Other factors have driven this trend as
well, including the rise of data privacy and protection
regulation, the recognition that third-party data is
often not that accurate or up to date, and the desire
to be more reliant on owned assets.
According to the May 2020 research from Winterberry Group and the IAB, 60.4% of data professionals surveyed in North America said they had increased their emphasis on first-party data because of the change in support for third-party cookies. The changes were also driving respondents to look more toward a variety of other solutions, including third-party ID resolution solutions, custom ID resolution solutions, other types of data relationships, and also contextual targeting.
% of respondents, May 2020
How Will the Change in Support for Third-PartyAudience Cookies Affect the Use of Data for DataProfessionals in North America?
Increase spending/emphasis on use of first-party data60.4%
Increase interest in third-party identity resolution solutions37.7%
Engage with third-party industry groups to build identityresolution solutions
35.9%
Increase efforts to build second-party data relationships35.9%
Increase focus on developing custom/in-house identityresolution solutions
35.9%
Increase spending/emphasis on contextual advertising34.0%
Increase testing/investment in new solutions32.1%
Source: Winterberry Group, "The State of Data 2020" in partnership withInteractive Advertising Bureau (IAB), July 22, 2020257477 www.eMarketer.com
In March 2020, Advertiser Perceptions research conducted for Foursquare found that most US marketers polled had increased transparency around data use and created new internal standards for its collection and use as a result of privacy regulations. Many had gone further, including by limiting the information they collect, reducing the amount of consumer tracking they engage in, and even reducing the amount of audience targeting they’re doing.
Ways in Which Privacy Regulations Have ImpactedTheir Company's Data Strategy According to USMarketers, March 2020
Increased transparency around data use57%
Created new internal standards for data collection and use54%
Considered new advertising channels and/or platforms41%
Limited information collected or included in customer profiles39%
Reduced the amount of consumer tracking39%
Changed or reduced audience targeting32%
Ask consumers for less data31%
Lowered social media advertising spend29%
Appointed chief data officer27%
Note: n=150Source: Foursquare, "Location Data in a Marketing Lifecycle" conducted byAdvertiser Perceptions, June 23, 2020256557 www.eMarketer.com
Marketers and their partners are responding to these privacy- and regulatory-related changes in two dimensions: tactically and strategically. Neither can be ignored.
“From a tactical perspective, it’s imperative for all advertisers and their partners to understand that how they do things today is dependent upon a set of measurement standards,” said Philip Smolin, chief strategy and revenue officer at demand-side platform (DSP) Amobee. “Those measurement standards are changing. So therefore, by definition, how you’re doing things today is going to have to change. On the assumption that what you should do is keep doing the same things you’re doing today, you have to begin leaning in toward trying alternative methodologies.”
For example, marketers who don’t know the level at which their current programmatic advertising practices are dependent on probabilistic graphing based on cookies should get ready to rely more on deterministic IDs other than cookies.
“But the bigger picture is there is a sea change that is going on right now,” Smolin said. “A lot of people, like us in the industry, love to focus on framing things as technology problems. ‘The cookie is going away—therefore the obvious question is, how do you replace the cookie?’ But the really important question is, ‘Why is the cookie going away?’” The reason for that goes back to consumers and their ongoing discomfort with digital advertising and tracking practices.
“You have two groups of digital marketers,” said Maor Sadra, CEO and co-founder of incrementality testing platform Incrmntal. “There are the ones who are really trying to figure out how we can continue doing what we have been doing, and then the ones who were like, ‘Oh, cool. There’s a new opportunity. We just need to learn how to adapt to that, and that’s it.’ Eventually, the second group will win because the first group is going to run out of options. If anyone has been counting on fingerprinting to be a solution, fingerprinting will get deprecated.”
Sergio Serra, senior product manager for the supply-side platform at InMobi agreed that fingerprinting, or other ways of using some kind of proxy for the IDFA, would not be an ultimate solution. Serra discussed the critical advertiser use case of frequency capping, which will become impossible for most users in most environments under new iOS rules.
“A use case we will no longer be able to definitively solve is avoiding presenting the same ad multiple times to the user,” he said. “If you don’t rely on a consistent ID—then 5 minutes down the line, you present me with another request from the same app—well, I don’t have any definitive way to tell whether this is the same user or not. And I might end up presenting you with the same ad, which is a lose-lose situation. Now, I would admit there are other proxies, but when someone refers to other means to target a user, that again is going against the Apple philosophy, because what Apple is saying is that you should not target a user anymore.”
GETTING TO A UNIVERSAL IDENTIFIER The idea of a universal advertising ID is essentially mythical at this point. Cookies, though they hadn’t been designed for the purpose, were almost a universal ID until smartphones broke web browsers’ near-monopoly on consumer time with digital media, and the industry still hasn’t worked out a substitute. The deprecation of third-party cookies may mean now is finally the time.
Potential candidates include the Unified ID 2.0 (UID 2.0), a collaborative identity effort announced by The Trade Desk with collaboration from major independent ad tech players including Magnite and Index Exchange. Identity resolution firms with their own identity graphs are also working to build these out to incorporate as much of the digital ecosystem as possible without relying on cookies—for example, LiveRamp’s Authenticated Traffic Solution (ATS) expands the capabilities of the company’s IdentityLink ID resolution service. Neustar’s Fabrick is a “unified identity ecosystem” that offers cross-platform measurement without depending on third-party cookies or IDs.
There’s a lot of skepticism about whether solutions like this will take off—mostly because of the amount of cooperation and buy-in involved. The ad tech companies working on UID 2.0, a single sign-on solution, have been open about these challenges. They say it’s one reason why it’s so important to have an open-source solution that’s designed to be interoperable with others, so that publishers, advertisers, and their partners can all select the setups and stacks that work best for them while still being able to facilitate transactions across digital media at scale.
They’re also open about the fact that many in the ecosystem have not always found a way to cooperate in the past. Numerous experts we spoke with in researching for this report had a “this time is different” mindset and a willingness to admit and recognize some failings of the past. That’s not a guarantee of success, but it’s better than the alternative.
PRIVACY SANDBOXES, PUBLISHER FIRST-PARTY DATA, CLEAN ROOMS, AND MORE There are alternatives to simply replacing cookies or IDFAs with a different universal ID, including various methods or proposed methods of allowing advertisers to execute data-driven advertising without getting access to personal information about users.
One example is Google’s Ads Data Hub, known as a “data clean room.” Clean rooms like this allow advertisers to upload their own first-party data, which is then processed along with Google’s information in the cloud, where the advertiser has access to privacy-first aggregated reporting.
Google also has a “Privacy Sandbox”—a project that aims to replace the functionality cookies have served, like allowing for frequency capping and retargeting, but in a privacy-centric way.
One element, Federated Learning of Cohorts (FLoC), would create a cohort-based system of advertising to “flocks” of users based on interests—with all the data remaining in the browser, where the processing happens. Google’s privacy framework, TURTLEDOVE (Two Uncorrelated Requests, Then Locally Executed Decision on Victory) is another proposal to the Worldwide Web Consortium (W3C), which has a working group trying to figure out the cookieless future for advertisers. Programmatic technology provider Criteo has also introduced a framework proposal, SPARROW (Secure Private Advertising Remotely Run on Webserver). But AdExchanger reporting from July suggests progress at the W3C has been slow.
Publishers with the technological wherewithal—and user base—have also come out with new options to continue using targeted ads in a more consumer- and privacy-friendly way. The New York Times, notably, has taken a page from Facebook’s book by gathering first-party data about its logged-in user base that it has processed into audience segments based on things like demographics and interest groups. Advertising clients can target their campaigns based on this proprietary first-party data. But as Times senior vice president of ad innovation Allison Murphy told Axios in May, “This can only work because we have 6 million subscribers and millions more registered users that we can identify and because we have a breadth of content.”
If Attribution Breaks, Does Attribution Fraud Disappear?
If the most dire expectations prove right and the demise of the IDFA does lead to a dramatic reduction in performance advertising in iOS, there could be a small silver lining. As advertisers have relied more on cost-per-action ads, where they pay only after attribution post-install actions (sometimes in very sophisticated setups), attribution fraud has also gone up.
But making attribution more difficult won’t make fraudsters give up on trying to claim it, according to Steven Ahlberg, head of product at cybersecurity firm White Ops.
“The bad guys that we fight each day on the advertising and marketing fraud fronts are notoriously resilient,” Ahlberg said. “We don’t anticipate there being a decrease in advertising fraud because of Apple’s decision to remove IDFA. In fact, removing IDFA further removes a critical layer of transparency for advertisers and could potentially lead to the increase in application fraud—specifically attribution fraud.”
Fortunately, Ahlberg also said the absence of the IDFA won’t make it harder for White Ops to detect fraud, either.
A NEW (OLD) PARADIGM: LEAVING USER-LEVEL TRACKING BEHIND Not all identity solutions rely on consent. In most US jurisdictions today, consumers aren’t required to consent to the use of third-party tracking cookies, for example. Many of the various workarounds and alternatives that have been built over the years, such as fingerprinting, are conducted without getting opt-in consent or providing a meaningful opt-out ability. But marketers know it’s only a matter of time before these workarounds become obsolete, whether that’s through regulatory means or because of moves by the likes of Apple and Mozilla.
Many in the industry are positioning this challenge as an opportunity to (finally) educate consumers about the importance of targeted advertising to their internet experiences, genuinely convince them to consent to sharing their information for that purpose, and enjoy a result where willing consumers volunteer accurate data in order to receive useful and relevant marketing messages.
“First, we are optimistic that when the value exchange of the internet is transparent and consistently communicated that we will not see notably higher opt-out rates,” said Dave Pickles, co-founder and chief technology officer at The Trade Desk. “It’s also important to remember that full ID coverage will not be necessary to fuel successful attribution and targeting. Advertisers will need only enough scale to create statistical significance for models fueled by deterministic data.”
But others caution that large shares of consumers will not consent to such data sharing. Several experts warned that it’s time to move away from the idea of user-level tracking and toward sampling-based measurement approaches, like panels.
“We’re increasingly in a world where user-level attribution can’t be relied upon to give one comprehensive read on performance,” said John Goulding, head of strategy for the US at programmatic partner MiQ. “Instead we’ll have a patchwork of attribution sources. On the one hand, there are solutions like SKAdNetwork and any APIs made available in the Chrome Privacy Sandbox that will provide scaled attribution, but with major limitations and obfuscation methods. On the other hand are deterministic IDs and any opted-in IDFAs that will provide rich data for optimization and the ability to develop custom attribution within data clean rooms, but at incomplete scale. The trick is going to be how to reconcile these different data panels to model accurate performance across 100% of activity whilst retaining the ability to optimize and customize measurement toward the specific goals of a given brand.”
There’s no question that this will be difficult, but most experts interviewed for this report agreed that the opportunity is there for genuinely better attribution than the user-level tracking in place today.
“In the future, and as IDs become scarcer, we can expect new attribution models to emerge where algorithms will allow extrapolation of accurate measurements on a portion of the audience to the entire ad reach,” said Romain Job, chief strategy officer at Smart AdServer. “You can think of the panels already in use in other traditional media, but with a larger scale, more accurate measurement, non-ID-related data points, and a strong layer of data science to cook all of this together.”
Anne Hunter, head of product marketing at consumer insights platform DISQO, noted that many of those trying to simply replace the cookie are “ignoring consumers wishes not to be tracked by companies they have never heard of. When building a single-source panel, you must start with the consumer in mind and get their permission to provide insights. With that baseline of transparency and trust, the problems of traditional measurement go away. You can understand consumer behaviors and marketing impact at a granular level—with high accuracy—because you can see the entire digital footprint of an audience member, confidentially.”
CONTEXTUAL TARGETING Panels may help with measurement after the fact, but advertisers will still look to target their messages as they’re being delivered. Contextual targeting—common in the days before programmatic display advertising scaled a decade ago—has gotten renewed attention in the wake of the GDPR. Publishers and other advertising partners are continuing to work on these offerings, which have gotten more sophisticated thanks to advancements in computing.
“Contextual targeting can be as effective as user-level targeting if deployed intelligently, i.e., finding the contexts that are a good proxy for not only the desired users, but also for the right state of mind for the desired users,” said Kasper Skou, CEO and co-founder of semantic custom audience firm Semasio. “This is, of course, not an easy task and quite dependent on the individual campaign goal.”
“All these privacy protection measures are threatening the targeting and optimization strategies adopted by marketers and platforms these past 10 years,” Job said. “User IDs fuel most programmatic advertising spend today and as they disappear, algorithms will need to rely on other data points to drive and optimize spend. Context has proven to be a relevant data point to drive advertising performance in the past, and we expect it to become more and more strategic to display advertising spend decisions, especially because contextual solutions in 2020 and beyond will benefit from robust technological innovation like AI and machine learning.”
Job reported that Smart AdServer measured clickthrough performance three times higher than average for automotive ads delivered on pages with matching semantic context. The company found similar results in all languages, and similar correlations for other types of ads and contexts, including travel. He did note that they were not able to analyze post-click performance as part of this research.
And it’s important to note that contextual targeting isn’t a panacea for privacy problems. “The page is a proxy for the user, and as such, it needs to be used responsibly and compliantly just like user-level targeting,” Skou said. “For example, if the topic of a page was deemed sensitive by some future industry self-regulatory frameworks or legislation, then we must take that into account before serving an ad.”
THE PROBLEM OF CONSENT REMAINS Many of the options discussed in the previous sections still rely on getting consumer consent for marketers—and their partners—to use their data for targeting, measurement, and attribution. The ad tech ecosystem sees consumer education as the key to getting that consent, especially education about the idea that targeted advertising is what pays for internet content, and that without consumers agreeing to exchange their data for it, the open web won’t be viable and only the walled gardens will be left.
This is not a new point of view.
In June 2008, former eMarketer principal analyst David Hallerman published a report on behavioral targeting that said, “As people make the internet an increasing part of their daily lives, privacy issues will become more of a battleground.” But he believed the problems could be resolved through “a clear process of informed consent,” pointing out that marketers had already learned to do this with email.
But a little over a year later, we noted research from the Annenberg School for Communication, University of California Berkeley School of Law, and the Annenberg Public Policy Center that found two-thirds of US internet users did not want to see ads tailored to their interests—and that their opposition increased when they learned how such ads were delivered. Almost two-thirds (63%) thought advertisers should be legally required to delete information about consumer internet activity immediately.
In 2010, covering other consumer research expressing discomfort with ad targeting practices, we wrote, “Education without effective empowerment with regard to their own data may not be enough for consumers to get comfortable with targeting.” The same year, we warned, “If consumers are not given effective controls over how and what is tracked about them, they may believe their only resort is government intervention.”
There has not been a clear process of informed consent, or radical transparency. There hasn’t been a capability for consumers to manage their preferences conveniently. There hasn’t been even the smallest effort to explain to the public how RTB works and what it means about who processes their data. But there has been a massive increase in spending on targeted digital advertising. And there has been regulation.
Research conducted in the UK in February 2019 by the ICO, Ofcom, and Harris Interactive found that consumer support for viewing digital advertising in exchange for free content dropped from around two-thirds (63%) to around one-third (36%) when survey participants were given some basic information about how personal data is used in RTB.
See our August 2019 report, “Consumer Attitudes on Marketing 2019: Privacy Concerns Mount, and Ad Blocking Isn’t Going Away” for more on the ICO/Ofcom/Harris Interactive survey.
In 2010, we cautioned: “The fact that over time, as behavioral targeting has featured prominently in the news, Americans continue to regard it as invasive, creepy, or otherwise undesirable further suggests that educational efforts may not be enough.”
Still convinced education is the answer? The following are highlights of recent research on consumer attitudes toward data-sharing and targeted advertising:
■ Almost half (43%) of US adults surveyed in February by Econsultancy and Cheetah Digital said they were limiting the brands they gave data to. Another 41% said they deleted cookies regularly, and 30% reported blocking ads.
■ About four in 10 US and UK adults surveyed in February 2020 by Adobe said cookies, along with pop-up ads and alerts, were the most frustrating part of engaging with a brand’s content.
■ Almost two-thirds (63%) of US adults surveyed in June 2019 by the Pew Research Center correctly understood that cookies allow websites to track user visits and site activity. That made cookies among the best-understood digital topics Pew was researching, second only to phishing scams.
■ And Performics and Northwestern University’s Intent Lab found in March 2020 that more than seven in 10 US adults wanted to see regulatory reforms of how companies treat their data.
% of respondents
Attitudes Toward Government Control of SelectCompany Practices According to US Adults, March 2020
Would like to see the government reform how companies treatmy data
71% 16% 13%
Would like to see the government reform monopolistic practicesof companies
68% 20% 12%
Would like to see the government increase regulation ofcompanies online
61% 21% 18%
Agree Neutral Disagree
Note: n=1,500 ages 18-70Source: Performics and Northwestern University's Intent Lab, "DigitalSatisfaction Index: Brand Trust Spotlight," Aug 20, 2020258130 www.eMarketer.com
And to the extent that users are willing to share their data, it’s with the expectation of meaningful choice and control—which is often still not present. About three-quarters of US adults surveyed in November 2019 by geofencing firm Herow said they would be more likely to share their location data if they were given a clear and easy way to control if and how it was used. Seven in 10 said they would be more likely to share if apps explained clearly why they were asking for location and what the value for the user would be.
% of respondents
App Attributes that Make US Adults More Inclined toShare Location Data, Nov 2019
Provides a clear and easy way to control if and how user data isused
74%
Explains clearly why it is asking for location and what is thevalue for the user
70%
Provides a method to select when and for how long the locationis accessible by the app
65%
Has an established brand59%
Is a brand that users are attached to59%
Complies with a major privacy regulation such as GDPR in Europeor CCPA in the US
53%
Note: read as 59% of respondents would be more inclined to share theirlocation if the app has an established brandSource: Herow, "Mobile Location in the Age of Privacy," Jan 28, 2020252996 www.eMarketer.com
“Marketers, agencies, and media companies are going to have to make a choice, and it’s a simple one,” Hunter of DISQO said. “They can either use various technical workarounds to replicate the cookie, or they can do what consumers actually want: get permission before collecting data.”
KEY TAKEAWAYS ■ Commonly used identifiers are going away, and marketers must respond both tactically and strategically. Advertisers and their partners must figure out how they are relying on current infrastructure and prepare to patch what they’re doing as necessary. But workarounds won’t be a long-term solution.
■ Regulation is here, and it’s going to get only more stringent. The CCPA already requires marketers to respect global opt-out mechanisms including browser-based “do not track.” The CPRA, which seems likely to pass, will go further and isn’t likely to be the last US data privacy and protection law to get on the books.
■ Privacy-forward platforms are working to make user-level tracking and attribution impossible. Platforms like Apple’s iOS have the power to truly require permission for marketers to use data—and most experts don’t think users are particularly likely to give it. That means marketers will have to go beyond user-level tracking to understand the effectiveness of their digital ads.
■ Now is the time to prepare. No solution is perfect yet, but all will take time to develop, and the clock is ticking.
Overcoming Identity Headwinds in a Cookieless World
This article was contributed and sponsored by Verizon Media.
It’s no secret that consumers are concerned about their data privacy. Today, there’s lack of transparency around the data exchange that fuels the digital economy. Consumers have been left in the dark, sounding the alarm for change, and forcing governments and industries to take action.
Desire for personalization outweighs data concerns
The new consumer perspective around data usage must be balanced against the benefits of personalized brand experiences. Even as major players in government and tech move to curtail tracking and expand users’ control, consumer data tells a different and somewhat more complicated story.
In a 2019 MAGNA IPG Media Lab and Verizon Media study, 88% of respondents said they find personalized ads useful, relevant, or interesting. Eighty-six percent said they see how the use of their data benefits their online shopping experience, stating that websites are more enjoyable and easier to navigate when they offer personalized experiences. A majority of consumers are willing to voluntarily offer some forms of personal information in exchange for tangible benefits such as discounts or recommendations.
Crucial to this equation? Consumer trust. Consumers won’t share data at the cost of personalization if they don’t trust the partner. Conversely, if they trust the partner, they’ll be open to sharing even if they can’t explicitly state what personalization feature they’re getting in return.
Where do we go from here?
The dissolution of third-party cookies and other tracking mechanisms is forcing organizations to build or find alternative identity solutions to connect with audiences. According to research from Winterberry Group and the Interactive Advertising Bureau, 60% of North American data professionals expect to increase their spending and emphasis on the use of first-party data, as well as ramp up efforts around third-party identity resolution solutions (38%) and contextual advertising (34%). Marketers must take an integrated approach to tackle a future without cookies—and
to keep the benefits of identity a reality without compromising consumer privacy. Verizon Media shares three ways to look at an end-to-end sustainable identity solution.
1. Next-gen audience solutions. Next-generation audience solutions will enable marketers to target based on interests, demographics, and contextual signals to deliver and maintain ad relevance—all without depending on any identity, third-party cookies, etc. Verizon Media’s aggregated measurement solutions leverage cohort-level data to access insights, evaluate impact, and drive performance.
2. Open partnerships. New solutions will require deep collaboration between industry partners to develop privacy-friendly solutions for adoption at-scale to ensure publishers are able to monetize, marketers are able to deliver relevant ad experiences, and consumers are able to exercise meaningful choice.
3. Unified ID. We don’t rely on third-party tracking mechanisms, cross-site tracking, fingerprinting, or other invasive and non-sustainable approaches. Instead, Verizon Media’s Unified ID is powered by Verizon Media’s proprietary identity graph. This solution empowers advertisers to draw on hundreds of millions of logged-in, opted-in users from across Verizon Media’s global portfolio of owned and operated media properties, as well as the industry’s only independent ad platform with a full-stack DSP and SSP to protect data integrity across demand and supply. Because our users are logged in, all signals aggregated by Verizon Media’s Unified ID are entirely consent-based, enabling marketers to create cross-screen campaigns that reach the right consumers at the right time and in the right context.
Giving consumers the personalized experiences they want while respecting their data privacy is a delicate dance. Our solution walks the line by offering our brand partners an opted-in audience to safely target, build, and optimize campaigns. In a shifting identity world, future-proof your strategies and reach your audience with transparency through Verizon Media’s identity solutions.
We believe connections and identity begin and end with the consumer. And we should know - as the 2nd largest digital media company, we engage 900+ million global users per month. We offer marketers identity at scale, built from 2 billion consent-based, cross screen data signals with direct consumer relationships at its core.
The cookieless world is coming. Stay truly connected and future-proof your business to power better consumer experiences.
Your account team is here to help:Email [email protected] to submit a request for research support, or contact [email protected] or 866-345-3864 to discuss any details related to your account.
To learn more about eMarketer advertising and sponsorship opportunities, contact [email protected].
The leading research firm for marketing in a digital world.
Coverage of a Digital WorldeMarketer data and insights address how consumers spend time and money, and what marketers are doing to reach them in today’s digital world. Get a deeper look at eMarketer coverage, including our reports, benchmarks and forecasts, and charts. Receive daily insights to your inbox. Subscribe to eMarketer’s complimentary newsletters.
Confidence in the NumbersOur unique approach of analyzing data from multiple research sources provides our customers with the most definitive answers available about the marketplace. Learn why.
Customer StoriesThe world’s top companies across every industry look to eMarketer first for information on digital marketing, media and commerce. Read more about how our clients use eMarketer to make smarter decisions.
