Yes this is my view of how the ecosystem could evolve and how a potential solution might work.
Why am I writing about this?
Google is attempting to restrict ad-blocking and tracking-blocking capabilities, which I find problematic from a user experience and user-control perspective. Firefox has responded by deciding to enable third-party cookie blocking by default for new users of Firefox.
I am not particularly comfortable with this battle for users and their privacy. Both approaches reduce the user’s ability to make an informed choice, while content publishers become collateral damage. One company is effectively trying to increase publishers’ advertising revenue, while the other is introducing measures that may reduce it.
Google’s approach is also aligned with maximising its own revenue. Advertising accounts for approximately 84% of Alphabet’s revenue and an even larger proportion of Google’s business. Given Chrome’s global browser market share, including browsers based on the Chromium engine, Google operates from a position that is close to a monopoly in several parts of the digital ecosystem.
The financial upside of restricting ad blockers could therefore be substantial. Users may be exposed to more advertising and more measurement technologies. From Google’s perspective, the ideal scenario would naturally include limiting competitors’ access to data as well.
This can happen through a series of incremental changes: transferring Facebook-related data into Google DoubleClick, accessing data associated with Twitter, using Gmail data to identify users’ purchase histories, and connecting behaviour through Android, Chrome and authenticated Google accounts.
In summary, Google probably holds one of the world’s largest collections of user-level behavioural data—and Firefox does not want to give it unrestricted access to even more.
What could a future without third-party cookies look like?
A website visitor would still download a standard JavaScript file provided by an advertising or measurement platform. The file might be larger and more sophisticated than today’s tracking scripts. However, it would be served from the publisher’s own domain, even though its functionality and configuration would originate from the advertising platform.
The advertising platform’s measurement JavaScript would store a first-party cookie and potentially additional information in local storage. It would then begin collecting behavioural data without requiring the user to download or accept a third-party cookie owned by the advertising platform.
The JavaScript could calculate statistics such as page views, time on page, scroll depth, ecommerce interactions and traffic acquisition sources. These calculations and behavioural signals could be processed and stored locally in the user’s browser.
The advertising platform would not necessarily receive the user’s raw behavioural data. Instead, the browser could maintain a local model describing the user’s interests and behavioural patterns.
This could be implemented as a relatively simple scoring model or as a more advanced machine-learning solution—for example, a pre-trained neural network running locally through TensorFlow.js. Advertising targeting would then be based on the output of that behavioural model within the context of a specific website.
The browser or website would transmit the model’s output, such as an interest category or propensity score, rather than the underlying user-level event data.
The more content a website provides, the more behavioural signals it can collect and the more accurate its targeting model can become. Publishers would also be motivated to encourage users to create accounts and sign in. Authentication would provide a longer measurement window, improve audience recognition and enable more accurate targeting.
The same infrastructure could support both advertising and standard website personalisation.
Advertising platforms would become much more deeply integrated into publishers’ websites. We would probably also see more questionnaires asking users what they like and which topics interest them. In return, users might receive access to premium content, gamification features or small loyalty rewards.
Connecting a social profile could become particularly valuable because it would provide publishers with additional declared and inferred data. User-uploaded photographs could also become valuable signals, as image-analysis technologies may identify interests, preferences and other attributes relevant to personalisation.
Of the possible futures for digital advertising, this is probably the approach I would prefer. It could enable relevant advertising and personalisation while keeping a greater proportion of behavioural data on the user’s device.
I use smart devices and online services extensively, including many Google products. However, Google and Facebook already control an exceptionally large amount of data about individuals, and both companies have repeatedly demonstrated that they are willing to use this data in the name of delivering a more convenient user experience.
What do you think the future of advertising targeting will look like? Will third-party cookies disappear?
P.S. Here is Lewis Hilsenteger’s perspective on ad blocking from the point of view of a content creator: Unbox Therapy.