2020 Vision: The challenges facing Universal IDs
The winds of change are whistling with increasing ferocity through the adtech landscape.
With Google announcing its intention to stop supporting third-party cookies by 2022, and Apple’s Safari and Mozilla’s Firefox browsers already blocking third-party cookies by default, the death knell for third-party IDs has been well and truly sounded.
Universal IDs - data portability mechanisms which allow companies to track customers across the web without using third-party cookies - are being widely touted as an alternative and privacy-compliant way to get the identity information companies need to have the best hope of serving up the right product at the right time to the right audience.
However, despite these being heralded as a universal fix, that’s not quite the full story as they come with significant issues of their own, specifically around data and consent portability.
Tell me more about Universal IDs
Universal IDs can be probabilistic or deterministic.
Probabilistic IDs are probably the most straightforward, and therefore technically the easiest to scale, because they don’t require any additional inputs such as a significant volume of users being logged in, for example. Instead, they combine some first-party data with additional user behaviour clues such as the browser used, as well as the screen resolution and download speeds on the accessing device, to stitch together a user profile. This can then, ideally, identify that same non-logged-in user across different devices and browsers.
However, in the era of consent and consumer opt-out that swept in with GDPR, this way of working is problematic. Profiles created in this way collect information without users fully realising, meaning they can’t easily opt-out because it isn’t part of a formal consent process. This has led to this type of identity gathering being called out by the big browser companies as less than desirable from a privacy perspective .
Determining the way forward?
Deterministically-based Universal IDs, then, would seem to be a better solution here, as they rely on some form of constant identifier – such as a login, or a phone number associated with an account. However, they also run up against a couple of difficult-to-overcome challenges.
The first is scale – they can only work with known, as opposed to unknown audiences, so they need access to a critical mass of consent from any given user base to be effective.
The second is what happens to the data gathered, which requires a third party to make sense of it and build understanding of user behaviour across the whole company environment. That means the knowledge acquired, and the commercial power contained within the identity graph builds up, and inevitably ends up being centralised, held in a silo by the identity vendor. This again doesn’t result in much control for the user.
There is also a scale issue across different data sets and publishers. Do companies just match with one and accept the limitations of the identity graph that can be created, or match with multiple sets with the associated additional SaaS license cost that entails?
Deterministic Universal IDs can, at least, be decentralised, with identity data held and stitched together locally, while still allowing a common understanding of identity across multiple data sets. This is the way which Google, ironically given how much data it holds, believes offers the best way forward.
This approach, which means gathered knowledge is not centralised, addresses privacy concerns, because no data is handed over. Additionally, it also means being less confined, in theory, by the limitations of someone else’s identity graph.
Better together but apart
But there is a better way still - establishing identity using a federated approach. One that enables the ability to find out as much as possible about your users, and gain insights from their intersections with other data sets, while still respecting privacy. The one which we, and increasingly, Google, believe is the only logical way forward.
If you think of the two biggest data behemoths, Facebook and Google, they are both effectively single closed-shop environments. Massive, admittedly, but closed all the same. Amazon, too, is its own giant walled garden.
The conundrum the adtech industry has long been trying to solve is how all of the benefits and none of the downsides of being inside these walls can be re-created elsewhere.
The decentralised approach of InfoSum’s Unified Data Platform offers the answer - effectively mimicking BigTech’s unparallelled data matching, with none of the privacy risks and loss of control of sharing data with a third party.
How? Uploading data sets – which are never moved, shared or centralised and remain under the sole control of the data owner – to individual bunkers allows for a statistical representation of the data in each to be connected with other data sources and analysed to match existing identifiers.
No data movement and no sharing removes the need for trust while simultaneously unlocking more information than ever before.
The future of data? It’s decentralised and private. And it already exists.