Personal Data - Stores, Lockers, Vaults

Session Topic: Personal Data Ecosystem (TH4D)

Convener: Kaliya Hamlin

Notes-taker(s): Wendell Baker

Tags for the session - technology discussed/ideas considered:

Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:

The identity ecosystem from the perspective of an entertainment provider consists of an audience side and an advertising side. The audience side of the house creates experiences which are fun to play and cause the users to return to the site. The advertising side is often a key aspect of the monetization strategy for the site. The current structure of the industry maintains both of these systems in parallel though that is changing. This diagram is part of that story.

Audience sites typically manage user identities using a "screen name" or "email provider" approach. This works well to identify the person sitting behind the program or device that is contacting the service. This works well when "the browser" is the only interface to the site. However, with the rise of the "app" economy (installed-on-client programs that are not obviously a JavaScript-enabled HTML browser) there is interest in identifying both the application and the user behind the application. Audience sites are increasingly finding ways to chain identities together using open protocols such as OpenID or by other means. Additionally the audience sites use information about user behavior on other sites to help personalize the experience. This can be done with offline data feeds that are provided to the audience site to help customize the experience.

The advertising side typically manages user identities in a "force-placed" approach where the browser or applications are assigned a unique identifier. This is done to manage the user experience along such dimensions such as frequency capping, recency or intensity and to associate personalization with the advertising experience. In these systems, the advertising identifier stamp is typically tied one-to-one to the device or program. However, a user may have multiple computers or browsers; also a computer or browser may be shared among multiple users. Both of these effects confound intensity and personalization systems which are actually directed towards individuals, not devices. Both online and offline sources of data are used to enhance the ad selection process. For a given opportunity, ads are often selected using an "exchange" which clears a trade much like an open-cry stock or commodities market. In order to increase liquidity in these marketplaces, systems of real-time bidding have arisen to allow the opportunity to be traded across multiple marketplaces. In order to ensure that the original levels of privacy and anonymity are preserved across the marketplaces, the user identity stamp is typically transformed (cryptographically hashed) when the opportunity is offered on a different marketplace. This preserves the privacy of the user and ensures that the publisher's data rights are preserved.

The dashed arrow in the middle is a new development in the entertainment industry. In that model the audience side identity, which nominates a user, is chained to the advertising side identity system. This allows the advertising system to personalize and to limit the advertising towards an person rather than having to guess (or not) about that based on device or browser use.