Amidst a greater consumer awareness of ad-tech's use of personal information, major players in the online advertising ecosystem have been adopting increasingly privacy-centric solutions. But what exactly does online advertising look like in a world where persistent, cross-publisher identifiers, like third-party cookies or IDFAs, are disappearing?
In this blog feature we discuss with Michael Vinson, SVP, Analytics Strategy, Comscore’s vision for a privacy-focused future without the ad-tech artifacts that have been used, or abused as some say, in online advertising for decades.
Q: How are stakeholders in the digital ecosystem and online advertising being impacted by privacy changes?
Consumers spend a large fraction of their time and attention online and have developed a greater awareness of the collection and use of their personal information than ever before. Both the buy-side and the sell-side are being challenged to continually improve engagement and ROI metrics, while also respecting consumers' privacy and maintaining their trust.
There is no one-size-fits-all measurement solution that will work for all stakeholders. Brands and agencies on the buy-side may or may not have direct relationships with their customers; either way, they often want to provide customer-specific experiences and messaging to improve the relationships they do have, without using “creepy” knowledge of their customers. Measurement solutions that allow the brands to find their targets online while maintaining trust and respect for privacy will be an ongoing challenge. Comscore sees future measurement as a complex interactive ecosystem among the advertisers that fund the open web, the publishers that provide the content and context for the ads and sometimes have knowledge of their users that can be useful, and ultimately the web users whose legitimate interest in privacy must be respected. By working with all the players, Comscore can connect the dots to provide value to the brands, serve the needs of the publishers, and ultimately be the trusted currency used to measure value and identify opportunities across the web.
On the sell-side, publishers and platforms can sometimes better understand their logged-in users based on their hashed login identifiers. When available and permissioned for use, first-party identifiers will be a significant part of future measurement. When they are not available, we will continue to use our extensive opt-in panels to measure reach, demographic propensities, and other metrics that our clients want. Comscore, as a trusted third-party measurement provider, is committed to putting consumer trust as a front-and-center principle, and generally to support the move toward privacy-forward, transparent measurement on the open web.
To facilitate the use of first-party and other identifiers along with non-deterministic strategies such as the Virtual-persons IDs (VIDs) of the World Federation of Advertisers (WFA) proposal, Comscore has adopted the theme of interoperability in its measurement methodologies: the need to develop a framework that is adaptable to every stakeholder and can work with different identity spaces. Comscore is committed to developing its entire measurement stack: panels, census network infrastructure, television return-path data (RPD), first- and third-party integrations, and more, in a privacy-focused and interoperable manner.
Q: How can a publisher or advertiser work with Comscore to develop privacy-focused frameworks for their business?
Comscore’s existing digital audience measurement solutions have a long history of adapting to changes in the industry, including the trend toward greater attention to privacy. As the browsers have changed how they handle the data exhaust created by web activity, Comscore has adjusted to keep pace. As regulations have changed in various jurisdictions, Comscore has been ready to comply. At the core of our measurements are our panels, which provide detailed information about online behavior in a transparent manner based on affirmative consent from our panelists. We use the panels to give us the human dimension to what are otherwise purely digital datasets.
For example, we recently launched our Predictive Audiences solution, which is the industry's first cookie-free audience targeting capability at scale. It is built on granular consumer behaviors and applied through privacy-friendly contextually-based ad placements.
Looking forward, the industry is already driving new privacy-focused frameworks. One such solution is the cross-media campaign measurement proposal of the World Federation of Advertisers (WFA). Comscore is working on this blueprint, put forward by Google and Facebook, alongside a number of other companies. It has a privacy-safe, probabilistic assignment of identities to traffic at its core. These VIDs (Virtual-people IDs) are designed to break the deterministic linkage between measurement identifiers and the actual people behind the traffic. The design is intended to produce aggregated results (reach and frequency, with demographics) that are consistent with what you would get if you used deterministic linkages. Essentially, a VID is assigned in a repeatably-random way at each publisher, and then combined within each demographic category to allow calculation of reach and related metrics. The VID is privacy safe in the sense that each VID is only attached to a single person within a publisher. Once combined with VIDs from other publishers, there is no longer an actual single person that the ID is derived from. Moreover, if run through the “sketch” methodology (a sketch is a probabilistic data structure that allows calculation of aggregate metrics only), there aren’t even single VIDs attached to each user, and no event-level data leaves the publisher. By assigning these identifiers or defining the sketches based on them, the methodology guarantees that the resulting metrics will be internally consistent and logically coherent. The design guarantees that no VID represents any one actual human being, but rather stands in as the behavior that could have happened, even if it didn’t. The training sets used to create the VIDs are the panels, which Comscore has used for decades.
Of course, not every publisher has the resources to execute on this fairly sophisticated methodology, so Comscore has also been working to develop solutions to make it easy for publishers to participate, as well as alternatives for publishers who do not want to work within the WFA framework. We are also prepared to work with the various unified IDs (UID) such as UID2.0, to the extent that they survive privacy scrutiny. If IDs exist and are properly permissioned, we intend to work with them in an interoperable manner, putting them into a common framework along with VIDs to give our clients the ability to use all of them. In addition, we are developing proprietary enhancements on top of the WFA methodology, to give our clients added value.
Q: With privacy regulations and consumer choices ever changing, how is Comscore supporting consumer privacy efforts as a measurement company?
At Comscore, privacy has always been a significant part of our focus. Though the last five years have shone a spotlight in terms of consumer awareness of online tracking, for over 20 years Comscore has continually innovated to provide privacy focused solutions across its measurement stack. We strive to be a trusted currency for our clients and partners to enable the planning, transacting, and evaluating of media across platforms.
In particular, we are aware of consumers’ fear of feeling exploited or manipulated using data collection methods. By building out a comprehensive privacy framework and basing our measurement methodologies on opt-in panelists, we hope to play a significant role in upholding transparency, integrity, and ethics in online measurement. We strive for continuous improvement in all aspects of panel management and are committed to transparency in data collection. For our advertisers, platforms, and other partners who work with us, this goal should be the same: to make users feel safe when engaging with properties across the web. With the demise of third-party cookies and other persistent identifiers, we now have an opportunity to change consumers’ perception of how they are treated online.
One of the great benefits of the digital age is the abundance of long-tail, niche websites. Are you interested in mid-century Flemish three-legged stools? There are probably a dozen websites out there for you. No matter how obscure or delightfully idiosyncratic your interests are, you can find information and community on the open web. And these small, independent sites can often stay in existence by using advertising to pay for the content. It’s a win-win for the publishers and the consumers. The open web needs independent, third-party measurement that is credible to all advertisers and able to show the value of all properties, large and small. The alternative is that these small, independent sites will end up with very low CPMs as remnant inventory sold by ad networks and other aggregators, and may struggle to survive. As measurement, especially cross-publisher impression and reach measurement with valid demographics, becomes more challenging in light of the legitimate privacy demands of consumers, there is a risk that some publishers will lose the ability to attract advertisers. Comscore’s approach is to continue to develop measurement methodologies that will both protect consumers’ right to privacy and enable their access to all content, whether mainstream or niche.