"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Nico Neumann, assistant professor and fellow, Centre for Business Analytics at Melbourne Business School.
Last week, Google announced that it would phase out third-party cookies and even provide a solution to stop fingerprinting over the next two years. The project has been termed Privacy Sandbox and puts another nail into the third-party data coffin, in line with similar browser changes from Mozilla and Apple Safari.
Many companies and experts now fear what the proposed tracking limitations mean for the wider ad tech ecosystem, particularly for third-party targeting, attribution and ad tech downstream businesses.
However, if we take a sober look at the upcoming changes, we can conclude this is a necessary system cleanup that will not only benefit consumers through increased privacy protection but also advertisers.
Attribution never really worked
First, predictive modeling rarely provides causal insights – such as whether an ad caused sales uplifts or if there was just a correlation (more ads served to people who convert) created by some targeting algorithm to make the ads look as if they worked wonders. The latter is more often true than the former.
Second, attribution models nearly always analyze digital data only, leaving out the complete picture of what is happening. For example, most brands will see a spike in their website and search traffic after a TV spot aired. However, if TV is not considered in the attribution analysis, search will get the conversion credit even though it was just a means to an end for the impact of the TV ad, which does not get a chance in a digital model.
So, forget attribution. Marketers are better off relying on experimentation (within platforms or geographically) and survey-based methods, such as brand trackers. Marketing-mix models will likely see a return too. This type of predictive model, leveraging aggregate data, still suffers from the first issue (questions about true causal inferences), but at least it attempts to take into account the entire marketing mix and typically a long-term view (branding).
Third-party data is often not worth the money
Many companies rely on third-party data targeting, driven by the idea that targeting would reduce wasted ad spend and thus media costs. This is not always the case in practice because ad tech has amplified many costs, while the advantages of using black-box data and complex tech are unclear. When we investigated the accuracy of more than 90 popular third-party audiences in a recent research study, the quality of the data segments was often poor and not worth the additional costs for using targeting.
Returning to first- and second-party data sources as well as contextual targeting is more likely to improve ad targeting accuracy and campaign outcomes.
Shifting power back to large publishers and retailers
The demand for first-party data clearly plays into the hands of large publishers and retailers who have enough reach to build their own logins and pools for targeting inventory, including Facebook, Google, Amazon, Disney, Walmart, AT&T, Comcast and News Corp.
Many other ad tech suppliers, whose business is built on piggybacking their tracking pixels to other audience owners, will struggle (see Criteo, whose share price declined strongly following Google’s announcement). A power shift to larger players is inevitable in the long run and similar to the retail trends in the 20th century, when big store chains replaced many independent, small shops. Economies of scale apply to data businesses too, probably even more.
Having consumer data in the hands of several players only should also facilitate handling consent frameworks and decrease the risks of data leaks.
Media buying should become easier, reducing agency relevance
Some media buyers from agencies believe they will profit from the upcoming industry changes. This may be true in the short term as agencies can quickly react to market disruptions. Agencies often create a new business unit for any new ad platform and can offer expertise because they more quickly accumulate knowledge by serving ads on new platforms for multiple clients. (This happened to search, programmatic and Amazon.)
However, a concentration of market players makes it easier for clients to have direct relationships with important ad sellers, reducing the need for agencies, which benefit from highly fragmented markets. Anyone with a credit card can buy ads from Google and Facebook – the beauty of self-serve platforms and automation. The same will likely happen with many other platforms of large suppliers.
And this is a desirable trend to reduce the costs of media buying. Future automation should improve operational efficiency and make media buying easier, cheaper and more accessible, not more complex and more expensive as Programmatic 1.0 did.
Back to the roots: advertising as a creative business
The current third-party ad tech system unfortunately has many flaws (fraud, hidden fees, long supply chains) and probably was a costly undertaking to create (the illusion of?) cheap audience targeting everywhere.
But many marketers seem to have forgotten that audience targeting is only a small part of advertising among many other exciting avenues to help companies market their products effectively. Moving away from unregulated and messy third-party data has many advantages and could help the industry prosper by refocusing our efforts away from “who to serve ads to” and toward “what ads to serve.”
Let’s not forget: Advertising, at its best, can be seen as art. In the past, it often has felt as if we spend 70% of budgets deciding who is allowed to enter the art museum instead of spending the money on the artists and beautiful exhibition rooms. The end of third-party tracking could provide an opportunity to change this.