Lucile Dattin, digital director, Zenith
Advertisers and media agencies have been relying on third-party signals ever since the expansion of programmatic buying – initially presented as the best way to target the right audience at the right time.
But has third-party data ever been that reliable?
Using my own data as a guinea pig, I decided to delve a little deeper, accessing the ‘interest categories’ in my Facebook and Google accounts.
Cookie-based data represents the largest chunk of third-party data collection, using information on browsing habits to target users with personalised ads. There are millions of data signals gathered everyday to build those audience pools, but cookie-based audiences have their limits, especially when it comes to accuracy.
While the walled gardens of Facebook and Google do not only rely on third-party cookies to build their audience profiles, it provides a good indication of how our browsing habits are used to build those pools, as third-party data aggregators have a similar methodology.
First stop, Facebook. According to the platform, Facebook had identified no less than 176 Interests category for my profile including “Horseback Riding” (I don’t have any interest for this), “Sparkling wine” (I hate sparkling wine) and “Friday (1995 film)” (I have never heard of this movie). A lot of seemingly irrelevant categories are also part of this list, such as “Image”, “Tuesday”, or “Australians”.
On the other hand, while Google was spot on with my age category and gender, when it came to interests, this is where I was presented with a useful example of the platform’s data gathering limitations. Interestingly, Google associated my browsing habits to an interest in blues, country, and classical music (when my go-to are actually techno and rock), “Parenting” but also “Not Parent”, “Golf equipment” and “Volvo”. I’m still trying to connect the dots to understand how I fall into this eclectic mix of interests. Close enough Google.
As marketers, we’ve all bought some type of third-party data, including cookie-based audiences to target interests or in-market segments. Those two GAFA (Google Apple Facebook Amazon) examples of data collection to build interest categories are good demonstrations of what we’ve bought so far – for some categories, they were close enough to our interests, but for most they were completely irrelevant.
To move away from cookies, Google is currently building its new Privacy Sandbox proposal, Topic API. This solution will still be signal-based targeting, but suggest a more sophisticated approach than its initial FLoC (Federated Learning of Cohorts) solution, to overcome concerns related to personal tracking and anonymity.
The disappearance of third-party cookies will hopefully help shift advertisers and marketers minds to focus on valuable data sources. The usage of first-party data will be key to building profiles and similar audiences from real customers, as opposed to relying on probabilistic data.
In order to get the most out of first-party data, some vendors have developed ID-based solutions, to translate clients’ first-party data into anonymised people-based IDs, and reach them across the open web.
This is what Epsilon does. Acquired by Publicis Groupe in 2019, Forrester highlighted Epsilon as one of the strongest identity solutions in its research into identity solution providers.
Others include The Trade Desk and Yahoo, which have developed their own ID solution with a similar methodology. Epsilon currently has the highest reach in Australia, while The Trade Desk and Yahoo aim to grow rapidly and reach most of the population by the end of 2022.
Digital spend trends still suggest that Google and Facebook will keep the highest share of investment this year, however identity solution providers present a good alternative to the walled gardens.
In deciding which provider will be the best fit for your business, every client needs to identify, test, and learn what works for them. The best partner will be the one that delivers the highest match rate and reach to your first-party segment.
While approximately 80% of advertisers remain dependent on third-party cookies, it’s important to remember that the quality and reliability of this data could be questionable.
Focusing on people-based solutions will be key to limiting your brand’s reliance on third-party data, ensure that we are continuing to better address audiences in the ever-changing and dynamic digital ecosystem.