close
6 Pages
Reading Time Estimate
31 min
What’s Happening
The advertising industry is taking a sizable hit during the pandemic, with US ad spend projected to be down 13% this year. At the same time, it is facing an enormous privacy-driven shift away from 3rd-party data, which has gained momentum over the past year.
The New York Times recently revealed in May that it is developing its own ad-targeting platform for advertisers, using 1st-party data directly collected from NYTimes readers rather than 3rd-party data from other sources. Leveraging the scale of its 6M paid users and even more registered users, NYTimes will initially offer advertisers access to 45 proprietary audience segments starting in Jul 2020 – slicing by age, income, occupation, interests, and other demographic factors. It plans to phase out the use of 3rd-party data for targeting altogether by 2021.
The move comes on the heels of similar efforts from other large publishers and retail brands. The journey to a post-cookie worldis reshaping advertising and will impact companies from automakers to consumer packaged goods. Those who can capitalize on distinctive 1st-party data assets are now repositioning themselves to do so.
Key drivers of the shift away from 3rd-party data
The shift away from 3rd-party data is being driven by 1) privacy regulation, and 2) sweeping changes by the major web browsers to phase out 3rd-party cookies (see our Dec 30 2019 brief on the California privacy law and our Jan 20 2020 brief on browser privacy). Both of these are in response to mounting consumer awareness and backlash to privacy breaches and abuse of data.
Privacy regulation
Since the EU’s General Data Protection Regulation (GDPR) took effect in 2018, other countries (e.g. Brazil, India, Canada, Australia) have followed in developing new national privacy bills or laws. These typically restrict how companies can collect, manage and use personal data, affording consumers rights such as the ability to delete their personal data upon request.
The California Consumer Privacy Act (CCPA) in early 2020 was among the most notable in its impact on the most populous state in the US, resulting in some companies treating it as national law. The CCPA takes aim at 3rd-party data, requiring companies to notify consumers if their data is being sold and allow them to opt out.
A new Congressional proposal targeting Section 230 (see our May 29 2020 edition on Trump’s Section 230 executive order) could make it harder for tech firms to use even their 1st-party data for behavioral ad targeting.
Privacy updates by major browsers
With privacy now a prominent front in the browser wars, most major browsers – including Google Chrome, Mozilla Firefox, Apple Safari, Microsoft Edge, and others – have added features to either block tracking by default or make it easier for users to block tracking. These new restrictions have focused on 3rd-party cookies, which are commonly used to follow users across the web and gather browsing data for, among other purposes, retargeting of consumers with ads on other sites.
Google’s Jan 2020 announcement that it planned to end Chrome support for 3rd-party cookies by the end of 2022, in particular, shook up the ad industry given its leadership in the global browser market with a 64% share. It has already started moving in this direction – the latest Chrome 83 lets users block 3rd-party cookies in Incognito mode as well as manage/delete cookies on a per-site basis. (Google is currently facing a $5B lawsuit for tracking users in Incognito mode.)
Apple Safari has also placed significant emphasis on privacy over the past few years, resulting in a 30-60% decline in ad pricing on Safari. Most recently, Safari updated its Intelligent Tracking Prevention to block a broader array of trackers including the widely used Google Analytics.
The shift away from 3rd-party data to what?
The natural next question becomes: If companies can’t rely on 3rd-party data, what options are left for targeting consumers with relevant ads and personalizing their experience? Models are emerging (or re-emerging) across industries for consumer targeting and personalization that may not rely on 3rd-party cookies, or any cookies, or any personal data at all. The approaches described below are not necessarily mutually exclusive, and in some cases, are complementary.
Use of 1st-party data and other variants
These alternative models will frequently use 1st-party data and its variants, instead of aggregated and hard-to-trace data from 3rd-party brokers:
1st-party data is collected directly from users, typically through identifiable interactions (e.g. logins, settings) or 1st-party cookies set and read by the website the user is on (e.g. to help stay logged in). 1st-party data can be gathered through any channel – such as cloud services, website, app, call center, email, text/chat messaging, social media, mailers, purchases, point-of-sale interactions, in-store beacons, and loyalty programs. While 1st-party data can be limited in scope and not every organization has access to 1st-party data (or is capturing it effectively), there is growing momentum behind it for its accuracy and reliability.
2nd-party data is essentially 1st-party data sold by its owner (perhaps a publisher or retailer) to another party to use as its own (often for advertising), typically through an alliance or private marketplace. 2nd-party data, which comes from a vetted source and is often more reliable than 3rd-party data, has become increasingly attractive as a way to enrich companies’ own data and expand their customer reach. It has emerged as another revenue stream for media brands, despite concerns about control over their data and how to manage user consent.
Zero-party data– a newer term of art coined by Forrester around 2018 – draws the distinction between the often-passive collection of 1st-party data (which includes observed and inferred data) and the collection of data actively and intentionally provided by a consumer for a given purpose. Zero-party data might be collected through a product or subscription on-boarding process, a loyalty program, a social story, poll, questionnaire, pulse survey or contest. While zero-party data is not new (and arguably a subset of 1st-party data), it’s been getting more attention lately in conjunction with the growing emphasis on consent and transparency.
Contextual targeting
One of the more established alternatives to 3rd-party data, contextual targeting relies on the context of the interaction rather than the identity of the user. For instance, an ad may be served up based on where a user came from and the content of the webpage the user is browsing – from the image or video in the user’s view to the tone and meaning of the text being read.
In its low-tech form, contextual targeting emerged in the early days of the internet and was used by Google AdSense before behavioral targeting gained traction. It fell out of favor for a while after advertisers saw the high ROI from behavioral targeting and retargeting. Lately, with the rise of GDPR and other privacy regulation, contextual targeting has re-captured the attention of advertisers as an approach circumventing the need for personal or identifying information.
Contextual targeting has advanced to incorporate and analyze granular context, with more adtech vendors like Adlede, Essence, GumGum and Admantx working on these modern solutions. AI-powered text analysis and semantic algorithms can gauge tone and nuance, as well as incorporate metadata, video transcripts, comments, and related keywords. With a better understanding of the content, contextual targeting can also lower the risk that a brand’s ads are served next to undesirable content.
There is some nascent evidence that, for publishers, behavioral targeting with personal data might be only a little better than contextual targeting – one study last year found that use of cookies resulted in just a 4% gain in revenue for publishers. There have been examples of publishers such as the New York Times cutting off behavioral targeting only to see their ad revenue grow. A Digiday poll found that only one-third of respondents had seen greater ad revenue from behavioral targeting. (There is significant variance depending on the publisher – Google’s own study, in contrast, found that less relevant advertising without cookies resulted in 52% less revenue for publishers.)
Publishers are joining forces to make it easier for ad buyers to target contextual ads. For instance, UK publishers representing 45M+ users – including The Guardian, Telegraph, and News UK – formed the Ozone Project to establish a unified taxonomy for tagging and defining “context.”
Advertising “walled gardens”
The term “walled gardens” is usually used to refer to the digital-advertising behemoths – Google, Facebook/Instagram, Amazon, and sometimes Apple. These big tech firms will continue to have access to the massive streams of 1st-party data that flow through their respective products and platforms, though how they use it may be constrained by privacy rules. This access to 1st-party data means their advertising business models can continue to thrive despite the decline of 3rd-party cookies (which Google and Apple, as major browser vendors, are playing an active role in facilitating).
Large publishers with reach and scale are also now positioned to build their own advertising walled gardens, using 1st-party audience data gathered through 1st-party cookies and consent mechanisms such as registrations and logins. Some have called publishers’ user-identification data the equivalent of a “publisher-side cookie.” Data from large publishers – which can be authenticated, high-quality, and enriched with user interactions across a large body of content – can be very attractive to advertisers if there’s an overlap in addressable audience. Some publishers – such as the NYTimes, Washington Post and DPG Media – are already working on data and ad-targeting platforms (see below).
Data alliances & “clean rooms”
Publishers are exploring ways to collaborate and share data sets that pool their 1st-party data – kind of an “ecosystem of walled gardens.” For publishers both large and small, data-sharing can help increase the leverage and value of their data, as an alternative to Google or Facebook’s platforms. 2nd-party data deals, while not new, have not gotten to scale because of the challenges associated with protecting both user privacy and proprietary data, Mechanisms are now arising to enable pooling in an anonymized way, matching users without identifying them.
Data clean rooms (such as Google’s Ads Data Hub and Amazon Marketing Cloud) are an approach to pooling data from multiple entities in a controlled space without direct access by any entity. Clean rooms are growing in demand as a way to combine aggregated data from the big walled gardens and customer data from advertisers (such as Unilever), while maintaining privacy.
There’s an emerging class of brands, publishers and data vendors that are pooling data to train “propensity modelsusing machine learning that can predict audience behaviors. Prosper, for instance, can combine 1st-party anonymous consumer data from its large-scale surveys with clients’ 1st-party data in a clean room to generate customized models.
Identity systems
To maintain compliance with privacy regulations, marketers will largely have to transition from anonymous cookies to fully-consented identities. This approach depends on the effectiveness of anonymized people-based identifiers in matching users to activities across websites, sessions and devices, along with good consent mechanisms. Players are considering how to position themselves in an ill-defined future environment with a 1st-party “universal identity infrastructure.”
Login alliances – which let users use a single login credential across different media brands – are still early stages but growing in adoption, especially in Europe. In Germany, an alliance of publishers and ISPs launched a unified login ID program under the nonprofit European NetID Foundation, with now 95+ companies involved. In addition to the 38M+ German accounts that are “NetID ready” so far, there are also plans to open up the program across Europe. Login unions have also formed in France, Switzerland, Finland and Portugal. Asia is more nascent in identity-focused alliances but Alibaba-owned South China Morning Post, which is working on its own persistent ID, has lately been advocating for a regional unified digital ID initiative for Asia Pacific.
IAB Tech Lab, the nonprofit technical-standards consortium affiliated with the Interactive Advertising Bureau, is orchestrating an industry-wide collaboration called Project Rearc. Project Rearc aims to establish standards for a “consumer-provided, consented identifier tied to privacy preferences” that would enable 3rd-party vendors to participate in the ecosystem. It specifically will focus on standards rather than an identifier product or service.
Existing unified ID consortia based on 3rd-party cookies (originally formed as an alternative to Facebook and Google) will need to evolve. For instance, the Advertising ID Consortium founded by LiveRamp (formerly Acxiom) largely relied on cookie-based identity-matching initiatives – such as the consortium’s own Open Ad-ID, IAB Tech Lab’s DigiTrust ID, Trade Desk’s Unified ID, and AppNexus ID (now called Xandr). Now DigiTrust plans to deprecate its shared ID once 3rd-party cookies go away (browsers are beginning to block its ID from being set as 1st-party cookie). AppNexus/Xandr pulled out after its acquisition by AT&T, in favor of an advertiser-friendly “community garden” based on AT&T’s 1st-party identity system with data on 170M subscribers. Industry watchers say the Advertising ID Consortium could evolve into a shared ID based on anonymized profiles, leveraging LiveRamp’s IdentityLink non-cookie, people-based identifiers.
Vendors are also working on approaches to shared ID and identity resolution that don’t involve 3rd-party cookies. Solutions from “independent ID” players (e.g. LiveIntent, BritePool, Net ID) that use hashed/encrypted versions of email addresses as identifiers are becoming more common. Others are using phone numbers or other personally identifiable information (PII), or combining with device-based IDs (see below), to identify users across sites. Some vendors like Merkle and ID5 are taking the approach of storing their shared ID or tag on a publisher’s 1st-party domain/cookie – though it’s not clear whether the browsers will view this as a workaround and crack down on this.
Google’s Privacy Sandbox proposals
Google’s Privacy Sandbox, launched in Aug 2019, is looking to bring a technical solution to life over the next two years that can balance user privacy while enabling personalization and an advertising-based business model. Current proposals include:
  • Changes to the real-time bidding (RTB) interface that allow publishers to opt for stronger privacy protections;
  • Federated Learning of Cohorts (FLoCs), which uses browser-based machine learning to group people (based on behaviors and characteristics) into audience segments that can be targeted by advertisers;
  • TURTLEDOVE (Two Uncorrelated Requests, Then Locally-Executed Decision On Victory), which can enable remarketing by having the browser send two separate requests with the contextual webpage and the advertiser-identified interest, respectively, followed by on-device ad selection;
  • A Privacy Budget that constrains how much user data is revealed to make it harder to identify a specific user; and
  • A cryptographic Trust Token API that would allow sites with user interactions that indicate real users to share that knowledge with ad buyers, to detect non-human traffic and ad fraud.
Other approaches & avenues
3rd-party cookies disguised as 1st-party cookies: This approach moves 3rd-party cookies to a publisher URL, effectively turning them into 1st-party cookies. While these would work only on the publisher’s web domains, they could still relay data to 3rd-party aggregators (e.g. website, username, IP address, browser information) – allowing them capture perhaps 80% of the value compared to a normal 3rd-party cookie. Other dubious related methods include the use of link shorteners, link decoration, and user redirection to tack on information to URLs, allow a third party to set a 1st-party cookie, or otherwise track users.
Device-based advertising IDs and their alternatives: Mobile advertising is heavily reliant on device-based, user-resettable advertising IDs, rather than cookies, people-based identifiers, or permanent device IDs (which are not supposed to be used for advertising purposes). Advertising IDs may not last in their current form, however – they are legally considered personal data in California and Europe, and Google recently faced scrutiny under GDPR in Europe for its Android Advertising ID tracking. Apple also recently announced it would prompt iOS device users with the option to allow or reject tracking by individual apps. With the two leading mobile operating systems also the leading browser players driving the move towards privacy, some industry watchers expect advertising IDs to gradually go away in favor of native attribution APIs (that cut out middlemen) or app-based identifiers that bar cross-app correlation.
Device/browser fingerprinting: Considered somewhat unsavory for its invasive approach to privacy, device fingerprinting pairs probabilistic methods with attributes like operating system, browser information, language settings, screen size, fonts, installed plugins, and especially IP address to “fingerprint” the unique device. Fingerprinting may be used by attribution providers and other adtech firms, as well as consumer-facing websites (e.g. banks) with a greater need to validate logins. Google, Apple and Mozilla, however, have all been taking steps to make it harder to fingerprint devices in their browsers.
Geo-targeting: An off-shoot of device-based tracking, geo-targeting uses location indicators (e.g. real-time GPS, IP address) to drive contextual ad targeting. At the zipcode level, location can be a proxy for demographics (e.g. income) and business information (e.g. SIC code). Location data, however, faces privacy hurdles depending on how it’s used, since it can be very hard to anonymize granular location data (see our Apr 15 2020 brief on geolocation tracking & surveillance). Apple has been active in releasing privacy features to limit geo-targeting, such as prompts asking users whether they want to continue to allow background location tracking for an app and allowing users to share only approximate (rather than precise) location.
Direct compensation to consumers: Some players are experimenting with paying users to contribute data, consent to its use, or watch ads. Facebook launched the Viewpoints market research app that pays people for taking surveys. The Financial Times asks non-registered users to take a survey and “pays” them with access to an article. Privacy-focused browser Brave and startup AdWallet are paying users to watch ads. Startups such as Killi, Wibson and UBDI are positioning themselves as consumer-focused brokers of personal data. Direct compensation, however, is rife with potential landmines – firms have drawn criticism for targeting teenagers (Facebook), adding affiliate links to URLs (Brave), operating a pyramid scheme (Zynn), and exploiting low-income communities (personal-data brokers).
Alternative media channels: Marketers are rethinking how they are deploying advertising dollars to be more strategic. There has been renewed interest in targeted advertising on media channels that don’t rely on cookies, such as connected TV (CTV), podcasts, and retail media.
  • Connected TV is expected to reach $9B in ad spend this year, driven by an increase in TV-watching during shelter-at-home. For audience targeting, OTT (over-the-top) players such as Roku, Hulu, AT&T/Xandr can tap 1st-party data based on logins, device IDs and content. Last year, LiveRamp extended its IdentityLink ID to connected TV, and it is now being used by adtech firms to identify a household during an ad buy. Traditional pay TV firms Comcast and Charter and cable network conglomerate ViacomCBS are also stepping it up with a recently announced joint venture in Comcast spinoff Blockgraph, which matches data sets from advertisers and ad sellers while limiting the personal data shared. Hulu is working on more actionable ads with better attribution (called GatewayGo ads), in addition to integrating with Nielsen Media Impact to measure reach. Disney Hulu XP, which will let advertisers buy video campaigns across Disney properties, is expected to launch in Oct 2020.
  • Retail advertising or “retail media” is the category carved out by Amazon, and more recently Walmart, Target, Kroger and other retail brands (see below). It has the advantage of tapping into massive troves of 1st-party data that goes well beyond demographics to granular behavioral data such as purchases, browsing and search history. Depending on the retailer, some of that data may be available for both online and in-store transactions. Retail advertising has the advantage of being close to the point of conversion – advertisers can get a “view into the shopping basket” and achieve high-quality attribution using the retailer’s transaction data.
The response from ecosystem players: Publishers
Publishers are investing in 1st-party data, as well as new ways to monetize it. Registration walls for non-subscription readers are growing in popularity among publishers such as the New York Times, Hearst Newspapers, GateHouse Media, and Tribune Publishing. Publishers are also experimenting with different types of value exchange – from offering an “ad-light” experience for a reader’s email address (Forbes), to unlocking content or a contest entry (USA Today) in return for a registration. A handful are building proprietary advertising platforms for their advertisers and to license to other publishers.
  • In Mar 2020, The Washington Post revealed it was tripling investment in its Zeus Technology Suite launched last year. The suite is comprised of Zeus Insights, an ad-targeting platform pairing 1st-party data and audience segments with AI-powered contextual targeting; Zeus Prime, a real-time ad-buying platform allowing advertisers to buy ads across a network with other publishers and directly from specific publishers (which could see $10 per thousand impressions vs. $2 CPM through middlemen today); and Zeus Performance, an adtech stack focused on page load times and ad viewability that signed 50 publishers in the 6 months since launch. The Post is licensing Zeus to other publishers around the world, integrating it with its Arc publishing/subscription platform that has been in-market since 2016. Revenue targets for Zeus Prime alone run into the “8 figures.”
  • In Dec 2019, Vox Media (which owns Recode, The Verge, SB Nation and New York Magazine, among other properties) launched Forte, a contextual ad-targeting platform that uses only 1st-party data collected across its 13 media brands from 125M+ readers. Forte can leverage a reader’s viewing habits on one Vox property to serve ads on another. Advertisers also get the opportunity to communicate with consumers directly. Vox’s portfolio of software solutions for publishers also includes programmatic ad marketplace Concert, publishing platform Chorus (which competes directly with the Washington Post’s Arc), and open-source comment/community management suite Coral (acquired from Mozilla in early 2019).
  • News Corp has been building an identity graph using 1st-party data on nearly 100M unique user profiles in the US and 590M+ anonymized user IDs globally, tracked across its media properties. Its News IQ global programmatic ad platform, launched in Dec 2017, uses the 1st-party data for contextual targeting – leveraging machine learning to gather context (e.g. category data, sentiment) and enable advertisers to target audiences based on behaviors and preferences. News Corp has engaged in regional deals to enrich its data, such as its recent Apr 2020 partnership in Australia with consumer loyalty platform Flybuys. At the same time, it has been offloading less profitable adtech units such as Unruly.
  • Condé Nast recently launched a 1st-party data product for advertisers in May 2020 called “Now|New|Next.” An evolution of its Spire audience-targeting platform launched in 2015, it offers advertisers a dashboard with insights into “how consumers are spending now, how those behaviors might change and who will be the next customer.” Condé Nast has been moving away from open auction towards guaranteed deals with partners – the source of more than half its programmatic revenue today, a significant shift from two years ago. It has invested in initiatives to demonstrate more value to advertisers – such as a deal with measurement firm Nielsen for in-store data to support attribution, guarantees of business results for high-dollar advertisers, and a Prime suite of high-end ad products including video.
The response from ecosystem players: Retail & consumer brands
Like publishers, retail and consumer brands have access to swaths of 1st-party data – from user logins, transactions and online searches. Some of the larger retailers have followed Amazon in developing their own advertising platforms to capitalize on their data:
  • Walmart over the past couple years has brought its digital ad business in-house under Walmart Media Group (see our Mar 26 2020 brief on the renewal of Walmart). It has been working to build a larger advertising business competing with Amazon for supplier marketing dollars, with more integrated campaigns in stores and on digital properties. Walmart launched a self-service ad platform in Jan 2020 where advertisers and adtech firms can buy Sponsored Product (search results) ads. While its efforts are still nascent, the promise of Walmart’s trove of data (275M+ customers visit its stores and ecommerce sites each week) unlocked for targeting and attribution has advertisers excited. Retail and audience data are already available for display ads (though not for search ads, which are typically keyword, SKU or contextual).
  • In 2019, Target rebranded its ad network Roundel and began pitching its 1st-party data assets to large advertisers in an effort to grow its retail media business. Target has built 147M+ customer profiles using its 74M customer visits/transactions every week across offline and online. While its audience is smaller than other retailers, Target is taking the angle that the demographics are attractive – “outdoor leisure aficionados” who are 33X more likely to be college graduates. It has a partnership with Disney (which has a 70% overlap in customers) for attribution, connecting Disney TV campaigns with Target sales data.
  • eBay, which began bringing its ad business in-house in 2017, has taken the approach of catering to large complementary advertisers, such as financial services, auto brands, and telecom. It has targeted Amazon’s weaknesses, pitching itself as being easier to work with, sharing data, and offering a coordinated experience for marketers. It was reportedly on track to generate $900M from its advertising business in 2019.
  • In May 2020, adtech platform Criteo launched a self-service portal enabling marketers to buy personalized placements across a network of ecommerce sites without 3rd-party cookies. Its partners at launch include Target, Best Buy, CVS Pharmacy, and Macy's.
What It Means
Depending on your role in the ecosystem, this might be an exciting or alarming time. To understand who’s on either side of that line, let’s rewind back to the genesis of the cookie. When cookies were first invented in 1994, these small text files were intended to be set and read by websites for the convenience of their users – to help stay logged in, support online shopping carts, and remember language preferences – not as a general tracking mechanism.
In the past, consumer data typically came with implied consent – through choice-based interactions with ecommerce retailers, banks, airlines, hotels and other sites. While data aggregators did exist, the ways in which personal data was collected, shared and sold were fairly limited. The technology to gather granular data about an individual’s behaviors across platforms, analyze it with machine learning, and apply it in real-time automated auctions did not yet exist.
Over time, consumer data gradually became decoupled from the consent of the human beings that the data describes. It became bought and sold purely as an information asset in liquid advertising markets, not unlike how financial derivatives can be decoupled from an interest in their underlying assets. An entire ecosystem arose to capture, aggregate and sell data – from data aggregators who existed solely to transact in personal data to hordes of app developers who sold user data to wireless carriers like Verizon and AT&T selling customers’ location data.
Today, over 90% of websites have 3rd-party cookies, which are often set to be longer-lived than just that browser session or that day. A news site like The Washington Post might have 40 cookies. Cookie-based real-time bidding frequently results in a slower experience for users (which 5G won’t do much to help with) – an ironic outcome for an invention originally intended to make the user experience better.
However, the norms around consumer personal data are changing rapidly. We should view the decline of 3rd-party cookies as part of the current “deployment” stage of this technological cycle, in which technologies penetrate the traditional core of the economy, spurring society and public policy to catch up. The momentum behind privacy regulation is no accident – the pendulum has swung in the other direction and we are now well past the Wild West of the digital age. Google’s announcement in Jan 2020 that it planned to phase out 3rd-party cookies in two years just confirmed the death knell for an artifact that has played a central role in the diminishment of privacy in everyday life.
All paths going forward will lead to and through consent. If the original sin was the decoupling of personal data from consent, then the solution lies in bringing them back together. Consent is becoming the standard for privacy regulation – and moving beyond just checking the box. GDPR, for instance, requires that consent be “freely given, specific, informed and unambiguous.”
Consent is the reason why 1st-party data is garnering so much attention, and why certain large publishers and retail brands are gaining more power and leverage. Under emerging privacy laws, consent is required even in companies’ use of their own 1st-party personal data, and it can be easier to gain and manage consent through direct interactions with consumers.
For publishers, size makes a big difference
Publishers have had mixed reactions to the looming death of the 3rd-party cookie depending on their size. One survey found that 46% expected to see revenue loss. However, smaller publishers are expected to be disproportionately impacted by not being able to readily access data and technology from outside sources. Publishers in Germany, for instance, saw a 45% decline in revenue when Firefox blocked 3rd-party cookies in 2019. With less data, smaller publishers are also less able to offer users a personalized experience.
On the other hand, large publishers like the NYTimes have advantages that stem from their scale – notably the ability to operate their own walled garden. NYTimes’ recent investment in a 1st-party data ad-targeting platform was only viable because of its scale and trusted brand. Advertisers are seeking closer relationships with publishers who have large pools of 1st-party data and engaged audiences. These strategic relationships allow them, among other affordances, to share data and apply AI to a larger and richer data set. As a result, large publishers will see more guaranteed deals, higher CPM, and a gradual shift away from auctions and middlemen. Their large pools of data also mean they are more attractive partners for data alliances or 2nd-party data arrangements with other publishers and ecosystem players.
The scale advantages for larger publishers in a cookieless world also extends to the size of their war chests and ability to invest in technical capabilities. The infrastructure required to collect and manage 1st-party data and generate insights can be a costly undertaking – one reason why the likes of NYTimes, Washington Post, News Corp, and Vox are among the few attempting it at scale. It’s also the reason why these publishers are licensing their technology – to share the costs and gain a recurring revenue stream. In a survey, large publishers were more likely to emphasize non-advertising revenue streams such as subscriptions and technology platforms.
Smaller publishers and local news outlets – already being squeezed and seeing closures and layoffs during the pandemic – have fewer options. The recent wave of AI-powered news aggregators (see our Nov 9 2019 brief) has the potential to either inflict damage by making content harder to discover or elevate local news among the audience that cares about it. Some smaller publishers may find a “way in” through alliances that pool 1st-party data, enable a unified identity (e.g. login alliances), or make it easier for advertisers to spend ad dollars with local news outlets. Others may move into a specialized or premium niche and survive through subscriptions, or convert to nonprofit journalism and survive through charitable contributions. Those staying independent will likely use the big ad platforms, plugging directly into the identity graphs and demand from those platforms. We can expect to see continued industry consolidation, especially amongst the smaller high-quality players.
It’s still early days in this wave of publisher innovation
We’re seeing significant experimentation in consent mechanisms such as registration walls for non-subscription readers and progressive profiling during on-boarding of new subscription customers. These help publishers collect more 1st-party data, in addition to supporting the core business (e.g. subscription conversions, metered paywalls, better user experience). Registration walls can be risky, however – only a fraction of non-subscribers (perhaps less than 30%) will log in and some publishers may see a bounce rate of 90%. Progressive profiling can also result in a negative experience for new customers, who might be impatient to get to the content.
Publishers are experimenting with incentives for readers to give their information – such as unlocking content, an “ad-light” experience, or contest promotions. They are working to drive more interactions across their properties including mobile apps, which have notifications that can draw readers back into the publisher’s walled garden. We’ll see more engagement mechanisms as publishers seek to gather more “zero-party data” – such as polls, short surveys, chats, and social media-like features. Community, loyalty and VIP programs can also help collect more 1st-party data in addition to reducing churn.
Consent strategies are being updated as well with consent management platforms, which can track consent at the user level and share with advertising partners, and authenticated consent solutions that sync user preferences across platforms and devices.
Non-cookie identity mechanisms are another major front for innovation. Anonymized identity lies at the heart of the challenge presented by the loss of 3rd-party cookies – how do you target and personalize at scale without maintaining privacy? We’ll see further traction in technology-led solutions such as the proposals in Google’s Privacy Sandbox and ecosystem-based approaches such as login alliances. Also on the rise are “group identity” schemes such as differential privacy mechanisms and probabilistic ID graphs that rely on group patterns.
Not all the innovation around identity and consent will be for the good. Some players will try to circumvent browser constraints and pursue less privacy-centric and more dubious approaches – such as device fingerprinting and disguised 3rd-party cookies. Internal audits for GDPR are turning up dodgy practices like consent-string fraud – which Condé Nast reportedly ran afoul of in Dec 2019. Much of this is happening behind the scenes – publishers don’t have a lot of incentive to share how they get around a constraint, since tech firms can and will plug any holes in their systems that arise.
Publishers are also working on enhancing their value propositions for advertisers – from partnerships to enrich data and support attribution (e.g. Condé Nast and Nielsen) to continued build-out of the proprietary advertising platforms described above. Those with 1st-party ad-targeting platforms will seek to differentiate based on the quality and breadth of their audience segments and the ways they’re able to enrich that data with specific insights about their users (e.g. preferences, emotions, likelihood to purchase). Better attribution will be key, given that these ad-targeting platforms are competing with retail media with a view into actual transactions.
With the rise of contextual targeting, publishers like the Washington Post are exploring ways to track content consumption journeys and build “lookalike” models. For instance, for a given anonymous user, publishers can use the referring site and a single click on a piece of content to help advertisers place an ad with a 90% likelihood of engagement. We may also see more creative partnerships between publishers and advertisers afforded by closer relationships (e.g. co-branded content, sponsorships).
Large retailers and consumer brands become advertising companies
The extraordinary success of Amazon’s advertising business – which generated $14B in revenue in 2019 (up more than 4x since 2016) – spurred other large retailers and some consumer-facing platforms to follow in its footsteps. While few companies can replicate the full range of Amazon’s assets and capabilities, the larger retailers can capitalize on the most critical leverage points – the full-cycle view of the customer journey and the direct tie between ad spend and customer purchase.
For advertisers with products being sold through a retailer or platform, we can expect the direct-attribution value proposition to be very attractive. Certainly, search-based advertising has proven to be an effective model – it is now the largest segment in global digital advertising, accounting for about $140B of the $365B in digital ad spend worldwide. It’s not yet clear how attractive advertisers with only complementary products and services (not sold directly) will find retail media channels, though eBay appears to be finding success with this approach.
Like publishers (see above), retailers and consumer brands are actively working to drive more “zero-party data” interactions across their properties. For instance, mobile apps and membership programs can generate fully-consented 1st-party data, in addition to making the user experience better and deepening the customer relationship. A pipeline of high-quality zero-party data has the potential to enact a lucrative strategic flywheel, as the big tech firms have demonstrated.
Expect to see more consumer platforms follow the path of Uber Eats into advertising. The advertising opportunity – which has proven to be highly lucrative for the big tech firms – is alluring for fast-growing companies that generate swaths of 1st-party consumer data through their core day-to-day business. This is especially true for the likes of Uber and Instacart, which are seeking avenues for profitability. Consumer apps already embed consent in their login process and typically have mobile analytics built in without having to contend with browser strictures. Here, privacy-related constraints in one industry are creating opportunities in another.
Advertisers will take a short-term hit but eventually adjust and benefit
Advertisers face an adjustment period in which they may see lower ROI on their ad spend – stemming from overreliance on the Big 3 digital walled gardens (which together represent nearly 70% of all digital ad spend today), as well as contextual targeting.
There's no silver bullet approach on the horizon – the more likely scenario is that different identity solutions will emerge to address each part of the consumer experience, which will have to be pieced together for targeting. Advertisers will need to build new adtech partnerships and competencies. Some – including smaller advertisers – will revert to the known walled gardens of Google and Facebook to ease the adjustment, at least in the near term.
Advertisers will face greater limitations in cross-site tracking – as one publisher executive put it, “[E]verything is becoming more site-centric and publisher-focused.” Omnichannel marketing will be much harder for a while.
The forward-looking large advertisers recognize that opting out of publisher and retail audiences would limit their opportunities, and are exploring direct strategic relationships with large publishers and retail brands. Advertisers with large pools of their own data are better equipped to work with large publishers and engage in statistical approaches like probabilistic ID graphs that can ensure a specific audience without matching specific identities.
Large consumer-facing advertisers such as banks, credit cards, and insurers – some of which already have teams of data scientists applying AI and predictive analytics in-house (e.g. for cross-selling) – are also working to leverage their rich 1st-party data sets for paid advertising. This involves facilitating real-time interaction between proprietary data from their owned channels and the major digital ad platforms (e.g. Google), with the data staying resident in their own architectures.
In the longer run, advertisers will adapt to a world without 3rd-party cookies and be the beneficiaries of more advanced tools from publishers and retail brands, better attribution, and less fraud. With contextual targeting, they also benefit from improved brand safety and lower likelihood of ads being placed near misinformation or inappropriate content.
Adtech as partners rather than intermediaries
The end of 3rd-party cookies and shortening of the value chain between brands and publishers will be hugely disruptive to the advertising ecosystem – resulting in an “adtech winter” and massive discontinuity. There are a number of adtech players with business models that are completely predicated on 3rd-party cookies, such as data management platforms and attribution vendors. Many of these firms will go out of business. We can also expect to see a shake-out and wave of consolidation among supply-side platforms (SSPs) serving publishers and demand-side platforms (DSPs) serving advertisers.
There will be an opportunity for adtech players to help companies enrich their 1st-party data with more data and insights (e.g. through machine learning or propensity models). 3rd-party data isn’t going away altogether, though it is becoming harder to use. The players raising recent rounds are offering non-cookie solutions – such as GumGum, which applies computer vision to webpages to enhance contextual targeting, and Bluecore,which offers AI-powered non-cookie personalization in email marketing.
Adtech players can also help publishers and retailers improve how they track, assemble, and monetize their 1st-party data. There is still much to be figured out in how to link anonymized user IDs across website, properties, and companies without the use of cookies.
Not everything is changing – big tech will still be among the winners
One major theme in all of this is that being big offers an outsized advantage – for publishers, advertisers and big tech firms. Google, Facebook and Amazon – together 70% of all US digital advertising spend in 2019 – will assuredly emerge among the winners here. If the one clear effect of the end of 3rd-party cookies is that the owners of large pools of 1st-party data will gain leverage and power, then the big tech firms with their massive stores of data will come out on top.
Google, for instance, can track users of its own products as 1st-party data. It has access to an enormous amount of 1st-party data from its Search engine and popular consumer applications (e.g. Gmail, Maps, YouTube), which it uses to dominate search advertising. Google can also use the 1st-party data collected from its code that is run on the many websites and apps using Google’s analytics, adtech, and other applications and plugins. Facebook has a nearly endless inventory of 1st-party data from user-generated content on Facebook and Instagram. Amazon has a view into ecommerce and transaction data at scale that positions it strongly to capture the attribution opportunity.
As one knowledgeable industry watcher put it, “View-through attribution and omnichannel measurement will not go away, but they will only come from walled garden-specific first-party data integration and clean rooms such as Google’s Ads Data Hub and the Amazon Marketing Cloud.” Real-time bidding also isn’t going away, though new approaches will be needed for targeting that don’t involve 3rd-party cookies. Google and Facebook are among the best-positioned with the technical capability, scale and toolsets to shape what comes next.
The biggest risks to these big tech firms are consumer trust and antitrust. With all three of them facing antitrust scrutiny right now, it’s a delicate time to be enacting changes that disrupt industries and disproportionately impact smaller players while leaving the big tech on top.
Consumers come out on top
The real winner here is the consumer. 3rd-party cookies have allowed, to some degree, marketers to become less connected to potential and existing customers. If 3rd-party cookies go away, it means that advertising will need to be more relevant to real consumer wants and needs, and brands will have to be more dedicated and diligent about building direct relationships with their consumer audience. Consumers will have more control over their data and greater assurance their activity isn’t being tracked or shared with the wrong actors. At the end of the day, that’s good news for us all.
Our thanks to Nitin Gupta (Director of Product Management, Advertising Technology, Pegasystems), Mike Banuelos (Vice President of Strategy, Hawke Media) and others for their comments on this brief.
Disclosure: Amazon and Google are vendors of 6Pages.
Have a comment about this brief or a topic you'd like to see us cover? Send us a note at tips@6pages.com.
All Briefs
See more briefs

Get unlimited access to all our briefs.
Make better and faster decisions with context on far-reaching shifts.
Become a Member
Become a Member
Get unlimited access to all our briefs.
Make better and faster decisions with context on what’s changing now.
Become a Member
Become a Member