How to handle media attribution in a cookieless world
By Media team. Top image: Open AI
The hot button issue right now in campaign tracking and media attribution is that support for third-party cookies is crumbling. How can you prepare for the cookieless future, and what are the alternatives?
While Google has pushed back its third-party cookie phase-out until early 2025, most of the marketing industry believe the end is nigh for the current user-tracking kingpin.
Third-party cookies are already blocked by default on Safari and Firefox, and it seems probable that Google will follow suit once it has answered the concerns of the UK's Competition and Markets Authority (CMA).
But while we know that the cookie switch-off means the end of the gold standard for measuring campaign performance, there is no obvious successor waiting in the wings.
Instead, there’s a Wild West of prospective alternatives...
Cookieless media attribution options
Cookieless attribution techniques use privacy-compliant methods to anonymously collect consumer data, and enable campaign impact measurement, while weaning the industry off its third-party cookie dependence.
The main contenders are:
Alternative IDs
Alternative IDs (also known as Universal IDs) are being developed specifically for the post-cookie era, generally by adtech providers in partnership with publisher consortiums.
There are many Alternative ID solutions in the market or under development. They typically devise a user ID by using cryptographic algorithms to convert personally identifiable information (PII) - such as email addresses or phone numbers - into a unique string of characters. This process is known as hashing.
These “hashed values” preserve uniqueness but conceal the original data thus protecting privacy.
The hashed data can then track opted-in users across participating websites and platforms, providing an alternative to the walled gardens of Google and Facebook.
Alternative IDs usually use deterministic matching to associate individuals with their different devices. This technique relies on gathering user profile data such as email addresses, phone numbers, date of birth and so on. Email addresses are most commonly utilized - as provided by publishers.
While deterministic matching benefits from accuracy, its weakness is scale. That’s because it depends on users logging in across multiple websites to track their campaign interactions.
The solution is to layer in probabilistic matching.
Probabilistic matching uses machine learning to link an individual’s activity across different devices, platforms, or sessions using anonymised data rather than personal log-in information.
The algorithms analyze patterns from multiple datasets to assess the probability that the signals point towards the same user. Like device fingerprinting, probabilistic matching correlates variables such as:
Device characteristics (e.g. operating system, screen resolution, browser type).
Network information (e.g. IP address, time zone).
Behavioral data (e.g. browsing patterns, click behavior, page visit sequences).
The upshot is you can generate a user profile without harvesting the personal data required by deterministic matching. In this case, the advantage is scale and the trade-off is lower accuracy.
But if an alternative ID solution combines deterministic and probabilistic matching, then the two techniques together can compensate for their individual weaknesses.
Good-to-know: Currently there isn’t a go-to alternative ID provider. It’s a fast-developing space that’s fragmented between many competitors and nobody has established a convincing lead or industry standard yet. The right solution depends on your market and objectives, hence we keep our eye on all the frontrunners.
Retail media networks
Like mammals surviving the meteorite that wiped out the dinosaurs, first-party cookies are unaffected by the imminent cookie apocalypse. So anyone sitting on a large trove of first-party data potentially has a valuable asset to sell. Especially as cross-site tracking becomes more difficult in a cookieless world.
Enter giant retailers such as Walmart, Best Buy, and the big kahuna, Amazon.
Amazon’s network includes owned entertainment properties like IMDb and Twitch, while BestBuy have announced a data-sharing partnership with CNET.
Shopify is also emerging as a strong competitor in this market as it can track user journeys across its multi-website ecosystem.
Good-to-know: Naturally, the relevance of each retailer’s campaign-tracking, media attribution product will vary by brand, but scaling your access to first-party data is a smart play in preparation for the cookieless future.
Zero-party data strategies
Zero-party data is personal information directly and willingly shared with a brand by consumers. In this instance, a customer is intentionally divulging something about themselves, usually as part of a value exchange.
Classic zero-party examples are competitions, quizzes, community engagement activities, UGC, reward programs, early access lists, co-creation initiatives, preference data, feedback forms, and promotions.
The clear advantage of zero-party data in a cookieless world is that it’s provided with explicit consent and it’s highly accurate.
First-party data differs because it’s implicitly collected through consumer interactions with a brand’s properties.
Zero-party data can be integrated with first-party sources to create more accurate segments for campaign tracking and media attribution.
Good-to-know: Zero-party data is low volume and can require some creativity to generate once the obvious sources are tapped. Happily the Dialect Creative team are awesome at ideating zero-party moments that drive audience engagement.
Device fingerprinting
Also known as digital fingerprinting, this old-school technique creates user profiles by combining multiple data points such as:
Browser information
Screen resolution
IP address (which provides an approximate location)
Operating system
Hardware information (e.g. CPU, GPU, available memory)
Installed fonts and plugins
Language and time zone settings
Fingerprinting works by utilizing so many variables that each profile is effectively unique.
If a profile is exposed to an ad and then logs into a game for the first time then a purchase has almost certainly been made.
This circumvents the dearth of attribution data escaping the walled gardens of Steam, Microsoft, Sony, and Nintendo.
The same method also works for micro-transactions so it’s a smart way to measure gaming software campaigns.
Good-to-know: Inevitably fingerprinting raises privacy concerns, especially if your campaign is subject to stringent regulation such as CCPA or GDPR. Thus ensure your device fingerprinting platform gains explicit user consent and has robust data security protections in place.
UTMs
The wonderfully named Urchin Tracking Modules (UTMs for short) are text codes attached to URLs that track users via your analytics software.
UTMs can reveal:
Traffic source: e.g. search engine, social media platform, email newsletter, or website.
Traffic medium: e.g. organic search, CPC ads, social, email, or referral.
Campaign ID: identifying the specific campaign.
Campaign content: distinguishes between different versions of the same ad or link e.g. different button colors or call-to-action variations.
Campaign term: shows the keyword that triggered the ad in a paid search campaign.
UTMs are attached to campaign URLs. When a user clicks through from an external source via a UTM-tracked link, the UTM parameters are captured by analytics tools like Google Analytics.
From there, you can analyze the data collected by UTMs to understand which campaigns, channels, and content are driving traffic, conversions, and engagement.
Good-to-know: UTM’s reliance on URLs means they’re a post-click technique that are silent on the post-view leg of user journeys.
Google’s Privacy Sandbox
The 100-Ton elephant in the room is Google’s proposed solution to the cookieless future: the Privacy Sandbox.
Its purpose is to enable targeted advertising and campaign measurement to continue while designing in greater privacy for individuals. Once Sandbox rolls out, Google says it will finally allow third-party cookies to go the way of the dodo. (Android device IDs are set to be eliminated, too.)
In the new world, ad targeting on Chrome and Android will be governed by the Topics API.
The API will generate a weekly list of a user’s top five interests (or topics) derived from their browser activity. Publishers and adtech providers can access a version of that list to serve ads aligned to an individual's interests.
Meanwhile, the Attribution Reporting API will manage optimization and measurement of ad campaigns. It’s a post-click and post-view solution, while other Sandbox tools will provide a level of cross-site and cross-device visibility.
Good-to-know: Sandbox is still under development but is available for testing. The emphasis is on preserving underlying user privacy and the signs are that Google’s solution will not offer the granularity of cookies. Moreover, cookies are an open standard. Some have voiced concern that a Google-controlled framework will not support the competitive adtech environment we’ve become used to. That said, much remains to be seen.
The steps to take to preserve media attribution in a cookieless world
While Google could keep kicking the can down the road, most industry professionals believe that the cookieless world will arrive in 2025. Hence our “what next” recommendation is to:
Plan and prepare now. Many brands are actively doing so because there is no magic bullet solution to the loss of third-party cookies. The answer will lie in finding the right mix of alternatives for your brand.
Devise new strategies to expand the scope of quality zero-party and first-party data flowing into your business.
Create new partnerships with agencies and platforms that are proactively working to solve the challenges presented by the cookieless world.
Consider which cookie alternatives will allow your brand to maintain campaign performance and visibility - while staying on the right side of privacy regulation - in the short and medium term.
Test in-market solutions aligned with your brand’s priorities, audience, and tech stack.
We are actively devising best practice strategies to deal with cookie deprecation and the realities of the privacy-by-design ecosystem. Get in touch if you’d like to talk through your brand’s challenges.
If you would like to discover more about our integrated approach and work together on a project, get in touch.
