Commentary

The Match Game: Cookie Syncing Is So 1995

Cookie syncing and audience matching is a process by which disparate pools of cookies are mapped to each other between two platforms such as a DSP and a SSP. Cookies, by nature, are domain-specific so the purpose of cookie syncing is to help platforms communicate with one another through a common language. Today, cookie syncing is the primary way that publishers and advertising technologies transact; where the cookie sync tables act as the translation layer between the partners. 

A cookie sync, as a stand alone solution for finding consumers, is problematic. For starters, cookies can fail for technical reasons such as expiring on a site. Within a cookie pool, there can also be a high proportion of users on iOS which blocks third-party cookies. Users can also abandon a page before a pixel request successfully fires. 

While there have been attempts by the industry to create a universal cookie pool, for any common identity framework the identities would be constrained to activities within a single browser. This does not solve for intra-device connectivity on mobile smartphones and tablets. Nor does it begin to attempt to connect a single person across multiple devices — not with deterministic data nor probabilistic methodologies.

A simple and possibly more effective strategy would be to complement cookie syncing with server-to-server de-duplication between disparate cookie-pools. By combining cookie syncing and probabilistic association, match rates between two different cookie pools would increase as opposed to a model which relies solely on the successful firing of two pixels simultaneously. 

Even more so, matching through a third-party who offers server-to-server capabilities will prevent possible data leakage as there is no need to piggyback a third-party pixel to initiate a sync. Because device IDs are less fractional and persist for much longer than browser cookies, the match rates and re-identification capabilities will also be significantly higher and will result in persistent identification of an individual across all of her browsing behavior, traversing both individual browsers as well as multiple devices.

Partners today are ecstatic at a 70% match rate post-sync. Add in a third partner and the acceptable, aggregate match rate drops to 49%. I personally do not accept this as a sound solution given that we have relied on this technology for 22 years after originally deploying the functionality for a totally separate use-case. It is time that inter-platform communication saw some innovation.

 
Next story loading loading..