In the flurry of new initiatives, while everything feels very new and progressive, with a great deal of activity and innovation underway, there is every possibility the outcome will inspire a sense of déja vu in anyone who’s been in the business a decade or more. Because in order to achieve the same aims, some older techniques will once more become de rigueur.
Before jumping into these new-old techniques, let’s have a quick recap of what we generally use online identity for:
- Targeting: to show ads to the people we think will be interested in and respond to them
- Frequency capping: to avoid showing too many of the same ad to people
- Attribution: to ascertain if an ad (and whose ad was it) led to a purchase
The humble third-party cookie is a wonderful tool for all these uses, and actually very privacy-oriented, as it doesn’t reveal identifiers like name, email address, etc., and is easy for users to delete. The most significant benefit of the third-party cookie is its universality. It works everywhere, with any vendor, any browser, any country. It itself is simple, even if we crowbarred it into providing complex solutions.
A future without the humble third-party cookie will be characterized by complexity, since every identity solution will only cover a part of the whole.
With this in mind, what are the runners[1] and riders for this future world?
- Project Rearc: led by IAB Tech Lab, so an AdTech-focused solution. This approach is trying to leverage the experience from the broadly successful Transparency & Consent Framework and the German netID initiative to provide some kind of user-volunteered non-cookie pseudonymous identifier.
- Privacy Sandbox: kicked off by Google, now led by W3C, so a browser-based solution.
- Private solutions, such as ID5, LiveRamp IdentityLink and TheTradeDesk Unified ID.
The key thing about all of these solutions is that they will be partial solutions – either linked to a specific set of publishers or websites, or to a certain set of use cases.
What marketers therefore need to start thinking about is what drives value for their brand (and ultimately, their customers), rather than “what can I track and measure right now?”
Realistically, marketers will likely be looking at a plan with the following components:
- A campaign running in one of the walled gardens – Amazon, Facebook, Google – to give broad coverage, with good targeting and reporting down to the user level thanks to their logged-in users.
- Campaigns offered through open environments – these will offer a variety of solutions where innovation can take place, but will also require a patchwork of targeting and attribution methods. These may involve one of the new mechanisms described above, or equally one of the tools that’s been gathering dust in the workshop for years, such as:
- Contextual targeting: There’s some discussion about what can be achieved via next-gen or advanced contextual targeting, but ultimately, it’s about the page, not the user – this approach can be useful for targeting, but won’t help with attribution.
- IP address & device-based household-level ID: This tactic effectively targets family units rather than individuals. There’s no reason why this can’t work well, but it offers less precision than cookies, which might be fine for targeting, but less suitable for attribution.
- Measuring success via panels: This is widely used now of course, but will be more significant post-cookie. Evaluating brand lift or other web activity from a panel then extrapolating to the whole campaign will become an essential tool.
- Measuring plain old sales, or footfall: This is powerful today, but in the future will be less attributable, i.e. as a marketer you’ll get an overall number, but difficult to apportion across different vendors.
- Offer codes: As old as the hills. In the ads, you specify a particular code which the user then enters at the point of purchase. Still used today on podcasts, but could come back in a big way on visual digital channels too.
Let’s also remember that while it’s possible that some of the innovation occurring right now may yet yield an effective user-level identifier, it still likely won’t be as comprehensive as the system we have right now (and for the next 18 months).
My advice to marketers and advertisers is to think about how you’re providing value to the brand, how that value can be measured, and which partners can provide what you need. Not everything is cookie-based – for example, reach, views, engagements – all of these in-unit metrics can be reported on by vendors with control of the creative executions. Flexible vendors can work with different panels to tailor the measurement to the campaign. Innovative vendors will be able to leverage whatever the best-in-class open identity solution(s) that exist post 2021. Thus, there’s no reason to panic, but every reason to get informed and start thinking this through, as those 18 months will fly by.
[1] We’ve already had a few fallers with The 5th Cookie and DigiTrust now out of the running. These were both primarily cookie rationalization solutions, and still relied on a third-party cookie.
