Contact us

Google’s User Account-Based Targeting On YouTube: 5 Key Implications9 min read

9 min read

On Friday Google announced plans for a significant new targeting technology development. “Over the next year” Google will develop changes that allow advertisers to reach audiences on YouTube that are defined by data in users’ Google accounts, and not by the data records held against cookies. This new data includes the fine detail of users’ Google Search and Maps history.

Google’s Customer Match (first launched in October 2015) is connected with the announcement. The Customer Match capability allows advertisers to match the email addresses they already know with the email addresses that Google stores against its users’ accounts.

With Customer Match, advertisers have always been able to target ads to their existing customers – across multiple devices – as well as to Google users who appear to be “similar” to their customers. Later this year, Google’s Diya Jolly writes, it will become possible to match against Google users in “new ways.” The industry will likely expect those new ways to include other common CRM data fields, like phone number and street address; and for Google’s Similar Audiences feature to offer bigger audience sizes.

There’s another important element in Friday’s announcement: that the operation of cookies and pixels on the YouTube website will be limited. The “legacy technologies of the desktop web” are to be abandoned. It seems highly likely that the limitation here will be on technology that’s not Google’s. Cookies that aren’t Google ones will neither be set on nor read from YouTube users’ browsers, but we can expect still to be able to reach Google audiences defined by the Google pixels that advertisers host on their own websites.

Google are quite explicit about wanting to see the end of “cookies and pixels,” but this ambition isn’t simply about wanting to evolve beyond the limitations of legacy technology. Google is promoting secure authentication – logging in – as an alternative to the cookie in advertising, and this has major implications for the industry. The authentication platform becomes a new – smaller – space for competition, and a new level of rivalry between Facebook and Google looks set to begin.

Friday’s announcement may have left anyone who followed the launch of Audience Centre 360 pondering that product’s positioning. We’re also likely to see a development in the online privacy debate as advertising technology and free online services are set to become inseparable like never before.

Here’s a bit more on those five key implications.

1. It’s the beginning of the end for cookies in advertising

The cookie, conceived as a technical feature of websites, has long been complained about as the way we identify people who might like to see an advert. Soon after most people had a smartphone, the cookie’s shortcomings in advertising began to look intolerable. They are as bad at crossing devices as they always were at crossing browsers (and at not being rejected by Safari). And compared with mobile apps’ robust device IDs, cookies look positively crumbly.

We identify ourselves online all the time: whenever we login to a site or app (when we “authenticate” with it). It’s a very logical idea that user authentication should be a better way to identify customers than trying quietly to attach a cookie to them.

Google’s move can be read as a bold statement of intent.

  1. We don’t think cookies are the most valuable way to define audiences.
  2. We think there are better experiences for users – and there is better business for Google – in evolving the way we define and sell audiences.
  3. We want to invest in this strategy by ceasing support (making life very difficult) for other ad tech businesses that prop up the old way.

Some of us might feel a surge of cynicism in considering part of Friday’s announcements. We’ll have to wait to see the precise detail of Google’s implementation as it goes live later in the year, but there is a certain element of the terrible old “legacy technologies of the desktop web” that we expect will remain strangely preserved. You should keep your DCM Floodlight and other Google “cookies and pixels” ad tech on your website, because you’ll still be able to use it to buy re-targeted advertising (from Google!) against the Google users who have previously visited your website.

And to be clear about “the end for cookies,” the course the ad industry steers doesn’t necessarily relate to the existence of the cookie on the web generally. Cookies remain useful in their original role: to help websites function. And this may sound a bit tricky, but while website authentication is now looking set to become central to advertising targeting, it seems likely to continue to be made possible on the web by the cookies that show you’re logged in once you have authenticated.

2. We can expect more aggression between authentication providers

We all hate passwords. We want online experiences to be easy, fast and secure. Wouldn’t it be nice and lovely if there was just one company that helped me log into everything?
Yes. Especially for that one company. Pretty much anybody can peg you with a third party cookie. Your identity by cookies can therefore be present on lots of different platforms at once. There’s lots of competition for the sale of your views, clicks and conversions.
But in a world where authentication is the basis for your identity as a prospective customer in the advertising marketplace, that marketplace will have far fewer audience vendors offering your identity for sale than it does today, and your identity will become more expensive.

With Friday’s announcement, we’re seeing Google give cookies a big shove towards obsolescence. We can expect continued efforts from Google, too, towards becoming pretty much the only platform that lots of people ever use to log into online services. It may not be long before, for lots of prospective customers, the only place that they can be bought online is from Google.

3. Google is going to be good for things where until now we needed Facebook

Facebook is great in pitches, because it’s so often a unique solution for reaching people with weird interests, or people in the throes of a “life event.” On the other hand, Google, today, can’t really sell those niche interests and fleeting circumstances. Whether or not you think Google is a moral organisation, it is not normally good business to sell a product that doesn’t perform as promised. With cookie-based audiences, Google can’t be sure enough whether a person really does collect Panini stickers.

The shortcomings of cookie-based identification mean that audience vendors don’t know with so much certainty whether the behaviour and interests implied by your cookie do actually represent your real behaviour and interests. A cookie-based audience, as a result, has more error in its definition than one based on definitively authenticated user accounts.

This means a cookie-based audience has poorer definition. It’s fuzzier, and unless an audience vendor is chancing it, the detail of cookie-based audiences must be more limited than login-based ones.

It’s a key reason why Facebook is able to sell 19-year-old ferret owners who have just started a new job, while Google currently cannot. It seems likely that already, today, Google can see plenty of users who might be in that audience.

When today we ask for a cookie-based audience similar to a list of people (i.e. when using Customer Match) the available audiences are often smaller than if we buy a login-based similar audience (from Facebook). From later this year, if audiences are matched by Google from user account to similar user account, there should be a lot more matches available with high confidence, yielding much bigger audiences.

4. Audience Centre 360 might take some time to get going

The DMP-like component of Google’s relatively recently launched GA 360 Suite of premium analytics products was offered as having third party integration at its core. We would be able to transfer insight into audience buys from across the ad tech ecosystem, and we’d be able to validate and optimise those audience buys in a single place. It’s not clear right now how that proposition would work when Google is planning to become a barrier in some significant contexts to other vendors’ ad tech. We might not be surprised by a slight pivot on, or a slow burn with, the ongoing tech development of Audience Centre.

5. This may dampen the online privacy debate

Authentication seems consensual in a way that cookies aren’t. Logging in is part of an online service’s user experience. But, on the other hand, so much of what we do today can’t happen unless we authenticate. With cookies, we’ve always had ways to choose whether or not to allow them: private browsing, deletion, rejection settings, ad blocking. With authentication, I can’t use Facebook unless I log in; and if I can’t login to Google, I might as well just go for a walk, and look at the sky for a bit.

One way to define privacy is as the capacity to choose to be unidentifiable. Fairly soon, individuals will not be able to prevent their identification for use in ad tech, because that identification will be an inseparable part of the essential digital services that we have to use pretty much every day.

Perhaps the role of government legislation in online privacy will get clearer. Free digital services, monetised with advertising, have always argued that the removal of privacy is part of the deal, i.e. “that’s why it’s free.” But advertising cookies are technically separable from the free service, which seems to have made things complicated. If access to free services and the core component of ad tech are one and the same… that may prove simpler in law.

Exactly where we would express choice about advertising is likely to continue shifting, and – ironically – end up firmly behind the ad tech login itself. We will have to login with a service provider/advertising business, and express preferences in our user accounts to say that, although we have linked our identities with lots of other information, the advertising company does not have our permission to act with that linked knowledge. Not, surely, coincidentally, Google launched this very service (quite quietly) into all Google accounts, in June 2016: My Activity.