Google Topics API Is a Step in the Right Direction for Privacy, But Still Lacks Granularity | Industry Insights | All MKC Content | ANA

Google Topics API Is a Step in the Right Direction for Privacy, But Still Lacks Granularity


As 2022 continues to fly by, third-party cookies rapidly approach their demise. There are many unanswered questions about the logistics of Google Topics API, but admittedly, Topics is a privacy-safe step forward. Agencies that drag their feet to find a new solution are only hindering their brand clients' success.

The deprecation of cookies cannot be postponed further, and advertisers must begin (if they haven't already) to test out privacy-safe alternatives. The relationship between a brand and consumers is one that cannot be taken for granted. It's a difficult pill to swallow, but there isn't going to be a single, easy alternative for third-party cookies.

How We Ended Up Here

Google's previous alternative targeting mechanism for third-party cookies, Federated Learning of Cohorts (FLoC), which identifies a user based on their "cohort" or people who have similar interests as the user, met some pretty harsh criticism from major players within the advertising industry. The biggest criticism was that FLoC used fingerprinting, which allowed third-party players to harvest data about users' devices to help compile profiles on users enabling them to be tracked online.

Following the missteps of FLoC, Google released Topics API as their latest alternative to the third-party cookie. Topics was created to address the initial privacy concerns of FLoC. Google stated that Topics will provide human-understandable transparency, reducing the risk of fingerprinting and avoiding sensitive categories.

Topics API should help reduce the chance that individuals could be identified based on their unique topics of interest. Using the feedback received from FLoC, Google proposed Topics as a new form of interest-based advertising in the Privacy Sandbox to be inherently more transparent and provide a more meaningful experience for users.

Topics API's original design included a limited taxonomy of 350 topics. Google eventually plans to expand their taxonomy to include hundreds to thousands of topics, attempting to exclude sensitive topics. As users browse different sites, topics will be inferred by the browser. The same user may have distinct topics assigned to them across various sites, mitigating the risk of revealing users' identities. When users browse websites on Chrome, they will be able to clearly understand the topics they are associated with.

Google's set of proposed topics will be public and manually curated to avoid following the IAB's footsteps of introducing sensitive topics to the set. The 1,500-row 2020 IAB taxonomy is one static Excel sheet of categories. Categories are mostly broad topics, but some are just specific enough to have been flagged by privacy advocates. When we look back years from now, many consumers will not be comfortable being categorized by some of the more sensitive categories in the IAB taxonomy.

The ultimate goal of Topics is for the taxonomy to be sourced by an external third-party with incorporated feedback and input from across the ad industry to build an even better category set than the IAB's.

This is a step in the right direction, because for too long the industry has been overly reliant on a governing body who can only agree to an updated taxonomy once every few years. But there remains the issue of waiting for a continuously changing committee to manually review, which will paralyze the opportunity for true understanding of content across the bidstream.

Overall, Topics is a great step forward, but it is not a silver bullet. Google even admitted "the mapping of sites to topics will not always be accurate," claiming that the training data is imperfect since it is created by humans and the resulting classifier will be as well. The uncertainty over its accuracy poses the question on the granularity of Topics, especially considering the relatively small number of initial Topics proposed.

Google also admitted the intention is for the Topics labeling to be "good enough to provide value to publishers and advertisers, with iterative improvements over time."

With a finite number of broad strokes topics with admitted imperfections, advertisers cannot rely on Topics to provide them the precise targeting and granular audiences that they've become accustomed to.

What Alternatives Advertisers Can Implement Now

Advertisers should start testing alternatives now so they aren't scrambling in the cookies' final hours to find a solution. Start A/B testing different solutions and ensure you are diversifying your cookieless advertising strategies.

Increasing transparency and leveraging zero-party data are great starting points. Personalize advertising experiences for consumers and provide some sort of value exchange. Doing so not only allows advertisers to get desired data needed, but it also builds consumer trust with your brand by personalization and remaining fully transparent with the data you're getting from consumers.

Explore the power of advanced contextual advertising strategies that are more granular than Topics. Contextual advertising is based on users' interests that are derived from websites they visit and has become more advanced in recent years. Not only does contextual provide consumers with a more personalized ad experience, but the ads are able to reach relevant targeted audiences on a more precise and granular level.

Advanced contextual targeting is a helpful solution for advertisers and brands to remain both privacy-safe and brand-safe. Advertisers can weave contextual targeting strategies into their strategies now to start tracking the results of these relevant brand experiences in the right moment.

The path forward to a cookieless future will look different for all advertisers, but one thing that will look the same for all — a focus on fostering trustworthy relationships with consumers, while ultimately respecting their privacy. Finding alternatives for cookies may be daunting, but the industry has the opportunity to create a new, privacy-first digital landscape for advertising in the future.

The views and opinions expressed are solely those of the contributor and do not necessarily reflect the official position of the ANA or imply endorsement from the ANA.

Doug Stevenson is the CEO and co-founder of Vibrant Media, which is powered by its targeting platform Quintesse. Doug and his co-founder launched Vibrant Media in 2000 as a pioneer in contextual ad tech. He was named as a winner of the 'Entrepreneur of the Year' from Ernst and Young LLP in the New York Metro area in this role.