Video Advertising 101: Cookie Targeting

Across the advertising technology industry, the term ‘cookie’ continues to be used in increasingly inaccurate and confusing ways. Some vendors contrast ‘cookie-based’ with ‘people-based’ approaches to targeting, which is misleading to some extent because both technologies only work in combination.
Simply put, a cookie is the basic technical method a web browser uses to identify users to a web server between page requests. On the web, every page request (when a user clicks a link) is independent from the previous one, so without cookies, no web server would be able to identify a user or react in a meaningful way to a user’s recent behavior on a site. Fundamentally, a cookie is just a code number that the web browser sends to the web server with every request so that the server knows it’s still the same user.
In that sense, every website is “cookie-based.” When a user logs into Facebook, it sets a cookie in their web browser to remember that it’s the same user. The same goes for YouTube, Google and any other website that with a log in.
For example, when a user logs in to YouTube, the server sends this:

This login cookie expires two years from now, and during that time the YouTube server can identify a user based on this code. In the background, Google & YouTube can match this cookie against Gmail, a mobile phone ID, cookies on other web browsers, Apple TV or Roku IDs, etc. Once a user log into any Google service on a web browser or phone, Google can track their behavior across their various sites and apps.
Other sites perform similarly, so in a sense, all sites that also have also a mobile app (e.g., Uber, Snapchat, LinkedIn, Twitter, the New York Times) are “people based.” Social video platforms have a wealth of data available to them, as users spend a great deal of time on these platforms. Other platforms know less about a user’s behavior, as users spend much less of their online time on these domains.
The real trick from a marketing perspective is in matching a cookie to a particular user identity that contains as much useful information as possible for targeting purposes. Facebook and Google are the leaders because they have the richest user profiles by far.
In contrast, most open web ad tech uses anonymous cookies, so they can only match user specific behavior on certain sites to a given web browser. They have no idea who users really are. This leads to sometimes extremely inaccurate information.
For instance, here’s what BlueKai (Oracle) and its data partners think they know about me:
Apparently I travel through time frequently, because I’m at the same time 30-34 and 50-54 years old (both incorrect):


I’m both young & hip (very true!) and an elderly parent (not quite that elderly yet, hopefully):


And apparently I’m very busy, working in four separate industries at the same time:

Compare that to the Google’s greater degree of precision on user interests. Not perfect, but orders of magnitude better because they can track users much more consistently.
Here is what Google thinks my top interests are (most of it quite accurate… minus Country Music and Fast Food, of course):

Facebook’s profile is even better, but not by much. (Apparently I’m interested in garbage bags and V8 engines, neither of which I count amongst my hobbies.)

Not only can Facebook and Google track user behavior on their own sites, they also track much of a user’s web surfing behavior. Across the web, those ubiquitous Facebook share buttons phone home every time (using the login cookie), informing Facebook of which page a user is visiting. Currently, Facebook only targets ads against this information on an opt-in basis, but that might change in the future.
Same with Google: almost every web page in the known universe uses Google Analytics or any of Google’s other free services (all of which match against a user’s login cookie too), which enables Google to build an almost perfect profile of a user’s web behavior. They’re not necessarily using all this information, but they could.
Both Facebook and Google have strong profiles of user interests and behaviors, most of which they already use for ad targeting. These platforms use cookies for web behavior tracking, but the real gold is in building backend user profiles.
Social video platforms are also continuing to develop other advanced targeting solutions beyond cookies, such as targeting against a user’s Google search history on YouTube. As social video platforms continue to make strides in tracking user behavior, it’s clear that these platforms will continue to offer a tremendous opportunity for advertisers looking to reach and engage their core audience.