Build, Buy Or Die? The Existential Question Of Cross-Device Identity

Considering its prevalence in trade publications and conference tracks today, it’s incredible to think back and realize that the entire notion of “cross-device identity” entered our lexicon only a few years ago.

It’s a multi-device, mobile-first world that’s increasingly fragmented. The advertising and marketing industries are prepping for a cookie-less world.

With cross-device conversations in full swing, there is a marked shift in how companies are approaching identity solutions. It’s build, buy, or die for every player in the ecosystem.

Who’s building and how?

For certain companies, it makes sense to build their own cross-device solution from scratch. Such companies have their own data, either first-party deterministic data or from third-party partnerships, or have enough of a window into anonymous data to take a machine-learning focused probabilistic approach.

They probably also have a team of engineers who have already built a bidding engine, optimization techniques, or measurement platform, which they believe can pivot or expand into working on bridging identity.



Companies are choosing to build their own cross-device identity solution in order to maximize scale, or to absorb costs of licensing from a third-party. When identity extension for media execution occurs within one platform, there is generally less audience drop-off.

Where a third-party identity graph is leveraged through a separate DSP, for example, the identity data would then be merged into that separate data/bidder and face a drop-off in scale, since the inventory sources would be different.

When the data and bidder are part of one platform, the data is built on the exact inventory being used in the platform, and thus more efficient.

There are different ways companies are going about “building.” Some are building identity from the core – bringing on dozens of engineers, accessing large processing power and infrastructure in various regions, and most importantly, developing a sustainable and global data acquisition strategy.

Other companies are building more of a co-op by comingling data, requiring the buy-in of the largest players across every vertical.

Finally, we see a hybrid model where intellectual property and engineering resources are acqui-hired to build a first-party solution, while the enterprise also relies on third-party partnerships to achieve scale and coverage.

Who’s buying and why?

On the other end of the spectrum, a lot of companies are looking at what it takes to build a passable identity graph -- and quickly opting out. That’s why many choose to partner with a recognizable brand.

The thinking here is that if marketers partner with those known leaders, they’ll attract a larger share of marketing dollars. There are classic examples of DSPs trying to get into the creative business, either through building its own or partnering with creative shops.

There are a few different ways of “buying” identity technology.

Some go for the buffet model, wherein they partner with everyone and build a mega-graph. While great in theory, this can be an expensive exercise. Unless there’s a back-end engineering team that can synthesize this data, enterprises are likely signing up for more than they need, both from an infrastructure and cost standpoint.

Many companies are realizing that the cost in combining too many partners is high, and the investment is tough to justify. These players are consolidating their partnerships, and only focusing on one or a few partners.

Who’s Dying?

Given global device proliferation and the resulting consumer identity fragmentation, the companies who aren’t looking at identity as a necessity in the near future, will likely be left behind. It's not just a battle advertising companies face; it’s everyone in the greater martech space.

This isn't about jumping on a bandwagon; it’s about moving with the changing environment of devices and platforms.

We’re on the cusp of a shake-up, and the winners will be the ones that find their identity – whether from building or buying.


Next story loading loading..