Building Identity in a Cookieless World
Q&A with Adi Raz, Head of Product at Clinch

September 2, 2020 | Katie Arena

The cookie started crumbling years ago. This is nothing new or unexpected. Just as you wouldn’t build a house on a cracked foundation, Clinch did not build our identity solution to be reliant on third-party cookies. For one thing, third-party cookies alone do not provide an adequate understanding of your audience. Now more than ever, advertisers should consider multiple user signals to build effective, hyper-personalized creative experiences.  

As Google’s deadline to eliminate all third-party cookies from its browser grows closer, Adi Raz, Head of Product at Clinch discusses how Clinch’s ID solutions will deliver ad personalization in the absence of third-party cookies.

How will Google’s decision to eliminate third-party cookies impact the structure of the adtech ecosystem?

Anytime the #1 web browser makes a significant business decision, you can expect there to be a shakeup across the adtech industry. But at Clinch, we view this as an evolutionary process that will allow the best technology providers to outshine those who can’t deliver on their promises. As the industry experiences another accelerated round of consolidation, if you’re not delivering a truly omnichannel approach, you will become irrelevant very quickly; on top of that, if you’re not in a position to add value to a brand’s first-party data either to enrich it or to utilize it in a way that benefits the brand, you will have a hard time once this becomes a fixed item on a brand’s tech partner checklist. 

Google plans to phase out third-party cookies entirely by 2022.
How does Clinch deliver personalized ad experiences without the use of third-party cookies?

Clinch collects both neutral (IP, browser info, etc) and behavioral signals from users on the advertiser’s website and during ad campaigns, in an anonymous fashion. This allows us to create unique user profiles across various devices, IPs, and browsers. These anonymous user profiles are applied to a machine learning algorithm that builds a knowledge graph of behaviors, vetting them to potentially unique users. The output is a probabilistic prediction for each user that allows us to serve personalized ads with high degree of certainty. 

How does Dynamic Creative Optimization (DCO) work with first-party cookies/data?

Clinch employs multiple methods of working with a brand’s first-party data to personalize messages to a user. One way is to add a Clinch pixel onto the advertiser’s website, which allows us to gather anonymized user information directly from the advertiser. Another way is to integrate with a client’s CDP or CRM, which allows us to digest consent-based user information from the advertiser and map it to our knowledge graph. This enables us to fine tune a message to a higher degree, based on the consumer. We are also able to take the information we gather on users from ad campaigns and feed it back to an advertisers CDP, which enriches each customer record, and allows for better insights for the client to understand their users, and build custom audience segments based on behavioral data. 

Safari/Firefox have already phased out third-party cookies.
How does Clinch execute DCO across these browsers?

We use a Clinch ID to segment users according to their behaviors, to provide the best message for each segment. This also allows us to serve a user a story-based campaign, whether it’s during the same session, or across multiple sessions. By integrating our Clinch ID with the client’s CDP, we can still identify user’s specific interests, and serve them personalized creative.

It’s important to note here the value of true omnichannel advertising. Your user will never be confined to a single browser type or channel, so by implementing a personalized omnichannel campaign strategy, you will do a better job of overcoming certain roadblocks, to more effectively reach your audience. 

How will we work with The Trade Desk’s Unified ID? And others “Unified ID solutions” that come up?

The Clinch ID solution was created to easily unify user and campaign data across platforms, channels and devices to help you better understand your customer and deliver more relevant, personalized advertising experiences. We understand that one single user ID solution will never be adopted by everyone, so we built our tech stack to allow for synergy between the ID sources we are integrated with and baked into our platform, as well as any ID solution that clients choose to implement.

How does Clinch obtain user signals?

Clinch collects various signals including IP, browser headers, user agents, carrier info, as well as user behavioral information such as websites visited, time of day/day of week behaviors, ad engagement, etc. These signals are pulled into our knowledge graph to build a potentially unique vector for each user. When Clinch encounters a new ad request we map the request to the best vectors from our graph and use it to personalize the creative.

What degree of certainty can you achieve with Clinch ID Solutions? 

We aim for a 90% certainty level when identifying a specific user in an anonymous fashion. There is an ongoing process to constantly test and improve this level of certainly on currently identifiable, yet anonymous audiences by comparing match rates via user cookie match, versus a Clinch ID match. 

Define “Deterministic” and “Probabilistic” in the context of the Clinch ID Graph Builder.

A Deterministic ID refers to scenarios in which we are 100% sure the uniqueness of the user is true. For example, when a user enters their email address into a form on an advertiser’s website, that email address is considered to be a deterministic signal. There is no further assumption to be made. The email entered is “true”.

A Probabilistic ID refers to when we conjecture based on multiple signals. Each of those signals may not necessarily be unique on their own, but when grouped together, we are able to establish a pattern that is unique. For example, IP and user agent are not unique signals by themselves, but when incorporated with behavioral signals such as a user visiting an advertisers’ specific product page, they provide enough of a pattern to uniquely identify the user. 

How does Clinch execute personalization in a mobile environment?

In a mobile web environment where we have third or first-party cookies,  personalization is achieved in virtually the same manner as display. In a situation where we lack cookies, we refer back to the Clinch ID pattern recognition. 

iOS 14 will be released September 2020, virtually killing IDFA (user opt-in rates est. 10-20%).
How will this impact our ability to operate across in-app environments?

While it’s undeniable that there will be an impact on the effectiveness of in-app mobile campaigns, the Clinch ID knowledge graph can still identify most users that are unique or belong to a specific audience segment.

What are they key data sources that go into a Clinch ID?

We leverage anonymized user signals from a variety of proprietary sources and data integrations:

  • Clinch Creative Elements
    • Placement IDs, campaigns shown, DSP signals, products clicked, social activities, and ad engagements 
  • On-site Tags/Pixels
    • Browsing behavior, device ID, anonymized IP, time of day, day of week, ad ID (iOS, Android) 
    • Contextual signals, domains visited, webpages surfed, consumer interests
  • Integrations with Third-party Identity Providers
    • LiveRamp user identity graph
    • Tapad cross device IDs
  • Clinch ID API
    • Seamlessly integrate your own user databases and first-party data for the most effective match rates and personalization

True omnichannel execution requires connecting all data sources and signals to match and anonymously identify the right user to deliver the most relevant experience. As it is obvious we are heading into some kind of a cookieless world, it’s important to be independent through multiple data sources and integrations, and remain committed to providing value to your audience throughout the entire campaign lifecycle.