PubMatic’s CRO Emma Newman explores scalable audience addressability solutions
Scalable audience addressability is one of the cornerstones of success in digital advertising. Now is the time for publishers and advertisers to knuckle down and define their strategy and roadmap for identifying, targeting, and engaging audiences in a compliant, brand-safe way that does not compromise user experience.
There are three main options available to do this: matched first-party data, universal identifiers, and browser-based audiences. Brands and publishers need to understand and weigh up the benefits and drawbacks of each solution on a channel-by-channel basis while keeping in mind the need for omnichannel measurement and optimisation.
Matched first-party data
When a publisher and advertiser are able to unequivocally identify and match an individual consumer using robust data - such as email address or mobile ID - addressability is an ideal option. For example, if someone uses the same email address to log into a retailer’s website and a premium publisher’s website, that email address provides verification that it’s the same person on both sites. This enables the retailer to retarget that consumer with personalised creative and messaging. Furthermore, matched first-party data creates a feedback loop between both parties that enables publishers to measure and optimise performance using the same data to increase yield and return-on-investment (ROI).
Matched first-party data is also highly-scalable through cross-publisher collaborations such as NetID in Germany and Project Ozone in the UK. These initiatives enable publishers to collectively compete with the walled gardens by identifying synergies between their audiences and combining their inventory in a single platform.
It is also possible to use matched first-party data to create cross-channel addressable audiences, e.g. when the same username or email address is used to log into an app on a smartphone and then on a tablet. From a CTV perspective, the vast majority of broadcaster VOD platforms, content apps, streaming devices, and other video services require individuals to sign-up with an email address to access their content. The increasing popularity of these services means CTV providers are rapidly building large-scale addressable audiences. Additionally, other identifiers such as location can be leveraged across multiple devices to create another layer of addressability.
The drawbacks to the matched first-party data approach are that it’s relatively complex to set up, potentially excluding many smaller publishers and advertisers, and there are challenges in creating multi-market solutions that can ingest and understand multiple languages and currencies and adhere to local-market regulations. However, the addressability technology space is evolving at a rapid pace with new, innovative solutions coming to market that could well solve these challenges.
Universal IDs
With the demise of third-party cookies, Universal IDs have been touted as an alternative to tracking and targeting consumers across the web. As these IDs are delivered in a first-party environment, many have considered them a more future-proof way of collecting identity information. To achieve data portability (and thus audience addressability), first-party IDs and associated data are shared with third-party Universal ID providers. These solutions use this data to build an identity graph which is made available to media buyers and brands for targeting.
There are essentially two styles of Universal IDs, each with their own benefits and challenges. Probabilistic Universal IDs (commonly referred to as device fingerprinting) combine some first-party data with, user behaviour clues, traffic-level attributes (e.g. browser and frequently visited websites), screen resolution, and download speeds on the accessing device. The combined data is used to create individual user profiles that have a high likelihood to match with other non-logged-in user profiles across different devices and browsers.
While relatively simple in design and set up, the longevity of probabilistic IDs is uncertain, as the three leading browsers have committed to blocking fingerprinting in the future. Rather than seeing this as a challenge, as an industry we should view this as an opportunity for further innovation.
The second style of Universal IDs is deterministic, or de-anonymised. These work in a similar way to third-party cookies, essentially tracking on-site user behaviour. The difference is that deterministic IDs use first-party cookies thus the data collected can legitimately be used for advertising purposes. The drawback is that first-party cookies do not provide cross-site or cross-channel insights because they are tied to the domain of the website owner.
Today, a big opportunity for publishers and advertisers looking at the Universal ID approach is to consider a blended approach. PubMatic combines various probabilistic and deterministic ID solutions which are made available to buyers in a single platform. This enables us to identify the best mix for individual advertisers.
Browser audience segments
The final option, which is still being defined, uses browser-based audience segments. Initial analysis shows that this approach is highly scalable but lacks transparency and visibility. Essentially, using browser audience segments results in an environment that is not dissimilar to the old network model - blind buys on unsold inventory. In order for this approach to be successful, there needs to be further innovation that significantly reduces the threat of brand-safety issues and enables publishers and brands to measure and optimise performance.
Audience addressability & a cookieless future
Over the next 12 months there is likely to be further innovations in the audience addressability technology space - especially in programmatic. It is important to harness the current opportunities and those that present themselves in the future in order to continue to provide publishers to compete with the walled gardens and give media buyers the data they need to plan and activate campaigns using independent addressable audiences.
Publishers and advertisers should focus on exploration and experimentation with multiple solutions in order to understand the efficacy of each solution against business objectives. Working with solutions like PubMatic’s Identity Hub gives publishers the opportunity to A/B test different solutions by activating, managing and curating ID solutions in one platform creating scalable, addressable audiences to advertisers to target.
For advertisers, creating private marketplace deals with PubMatic’s unique ID targeting creates an opportunity to improve personalisation, boost media efficiency, and campaign ROI across omnichannel formats: mobile, in-app, CTV, and web display.
The final juncture is layering additional audience data in a way that gives data owners control enabling buyers and sellers to transact in a mutually beneficial way. Within PubMatic’s Audience Encore buyers can create fully-custom data segments in private marketplace deals and target audiences across a multitude of first-party audience data readily available through our premium and exclusive partnerships, including: H-Code, Captify, Audigent, Lotame, and more.
Download PubMatic’s report ‘The Present and the Future of Audience Addressability’ here.
Posted on: Friday 9 April 2021