In plain English: How does cookieless targeting work?
By Rachel Foskett, Senior product marketing manager
A blog series explaining some of the concepts, processes and technologies we need to do our jobs – in plain English.
Cookies are on their way out. (Take a read of our previous post to find out why).
In the next couple of years, the morsels of data used by ad tech companies to track users around the internet for targeting and measuring ad success will soon be blocked by Google Chrome (the largest web browser), making them essentially defunct as a tool for online targeting.
But once cookies are gone, brands will still need to build audiences. They will still need to tailor advertising to different groups. And they will still need to measure the results of their campaigns (ideally, more successfully).
So, what does the cookieless future look like?
Beyond cookies – the two different targeting methods
Right now, the cookieless future on the open web (ie the internet outside the closed platforms of Google, Facebook etc) can be roughly divided into two groups – authenticated targeting and anonymous targeting. Let’s take a look at both.
Authenticated solutions work by getting explicit consent from an internet user to use their data. Usually, this takes the form of a login screen on websites before you can read or view the content. So, as well as agreeing to the site itself recording data about your visit (first-party), you’re agreeing that other companies (or third-parties) who’ve placed pixels on the site can collect data too.
Essentially, this isn’t that dissimilar to the old third-party cookies. But the big difference is the addition of explicit consent from the user.
The approach faces a big challenge, namely that if a user logs onto an individual website, that data can’t then be ported around the web to track them – and most sites simply don’t get enough traffic on their own to make this a worthwhile exercise.
Ad tech companies are currently working to solve this problem. The ‘Unified ID’ from The Trade Desk is one example that aims to provide technology that allows for sharing the same ID across a group of websites, so if a user logs into one website it matches up with logins from other opted-in websites. A user can then be tracked across all of the websites within that ecosystem.
A different approach is called ‘ID resolution’. In this method, rather than providing a user with a single ID that can be used everywhere, they focus on matching all the different IDs a user has within their technology. Using this data, they can create what’s called an identity graph – a database of all the different data points about a single user that are tied together.
LiveRamp’s IdentityLink is a front-runner in this area, with over 200m user IDs available to be connected to other datasets. With ID resolution partners, you can combine the identity graph data with your own first-party data, essentially offering the best of both the first-party and third-party worlds.
Anonymous targeting takes the opposite approach to authenticated methods. Rather than getting explicit consent to identify and track a user, it does not identify them at all. Targeting can be anonymous in one of two ways – either by using contextual targeting, or aggregated targeting.
Contextual targeting relies on non-personal data based on where a user is viewing content rather than who they are. For example, if a user is reading an article about yoga, you can take a guess that they’ll be receptive to ads for yoga gear or fitness classes on the same page.
If a user then engages with such an ad, you can use technology to scan the page for other contextual elements and find pages that have similar features, for example healthy living content or exercise gear content, and target people there too.
This method relies on a lot of data and pretty sophisticated machine learning to process it all – but done right, it can be just as effective as audience modeling.
There are also a lot of companies working on proposals for aggregated solutions – which groups users together in a cohort for segmentation or analysis so that any one individual user is never identifiable or trackable. Google Chrome themselves have built what they call their ‘Privacy Sandbox’ where developers are putting forward various proposals for this kind of solution.
There are also solutions appearing that focus on aggregated targeting via what is called ‘edge computing’. This focuses on bringing data science to the data on-device or on-browser – “on the edge” – as opposed to removing the data from the site, to a set of servers and then analyzing and segmenting the data there for targeting, which makes it easier for that data to be exploited. Add edge computing techniques to federated learning – a machine learning technique that trains an algorithm across multiple decentralized edge devices – and you can get similar real-time segmentation and targeting as we do now.
And while authenticated and anonymous targeting are the two main routes for replacing third party cookies, it’s also important to remember that plenty of channels and datasets, especially newer ones have never used third-party cookies and will be relatively unaffected: Connected TV, programmatic audio, and mobile in-app targeting are all examples of media channels that rely on non-cookie based anonymous identifiers for targeting and measurement. And real-world data such as weather data, location data, temporal data (about the time or day) or macro trends around current affairs, health or social events can all be used for more personalized targeting without using personal data or third party cookies.
So, both authenticated and anonymous targeting will allow advertisers the ability to identify and target their audience in the world beyond cookies. But the level of measurement that’s possible with each is not the same. (In fact, the future of measurement is a whole topic in itself, so we’ll save that for another blog post.)
Evolution is good
The story of digital advertising has always been one of change – and the end of third-party cookies is simply a part of that. Advertisers should not be waiting for a ‘silver bullet’ to cookieless targeting – as always, it will be a process of trying things out and iterating, and likely taking a final approach that combines more than one of the methods we’ve outlined today. Now is the time to start testing new options to shift away from cookies, because in the end, that is the only way to really prove the value of any of these solutions. Undoubtedly, it will mean that we all have to do new things and think differently about how campaigns are run. But we think building towards a world where people don’t have to be suspicious of advertisers using their personal data while at the same time still funding a free internet is a good thing.
If you would like to learn more about how to prepare for a cookieless world, book an MiQ Unlocked session or get in touch.