We’ve covered publisher first-party data, advertiser first-party data and second-party data, but there is still yet another key type of data – third-party data.
What is third-party data?
Third-party data is any information collected by an entity that does not have a direct relationship with the user the data is being collected on. Often times, third-party data is collected from a variety of websites and platforms and is then aggregated together by a third-party data provider such as a DMP.
By aggregating data from a variety of disparate properties, DMPs are able to create comprehensive audience profiles. These profiles contain information on users’ web interactions and behaviors, which are then used to categorize users into particular segments, such as auto intenders or sports fans.
For example
A user might visit a sports website and select their favorite team. They might then navigate to an auto enthusiast blog, and finally fill out a credit card application on a banking site. The sports site doesn’t necessarily know that the user loves luxury cars or has a high income. Unlike the sports site, the data provider does not have a direct relationship with the user, however since the provider collects data from the user as they travel the web, not just on an individual property, they can form a more complete view of that user.
What are the benefits?
Third-party data provides a breadth of information that can’t be matched by an individual entity. While first-party data is usually considered the most valuable since it is free and developed deterministically, it simply can’t match the breadth and scale of third-party data that is modeled probabilistically. While an individual publisher might have highly specialized knowledge about a user from interactions on their site, since most sites have specific content or products that attract users, it is likely they won’t know much about the rest of the user profile.
Additionally, these third-party data providers distill the data down to targetable audience segments, removing analysis that the advertiser may have needed to perform otherwise.
Third-party data and audience data in general helps shift away from using content as a proxy, allowing advertisers to buy audience segments based on key characteristics.
There are consumer benefits to audience targeting as well. Numerous studies have shown that consumers often prefer personalized advertising, as long as it doesn’t overstep and invade their privacy. A well-known study, run by Adlucent, found that 71% of consumers prefer ads tailored to their shopping habits. It is intuitive that consumers engage more with ads that make sense to them. In fact, a Nielsen study found that 78% of respondents said they had discovered a new product from a TV or internet ad, compared to just 56% that found out from friends or family.
How is the data sold and activated?
Data providers sell this aggregated, anonymized data to advertisers to facilitate targeted ad buys, allowing advertisers to target and tailor ads to effectively engage those particular audiences. These bundles are typically sold on a CPM basis with the hopes that a more targeted, engaged audience is worth the additional cost since the ad might otherwise be exposed to audiences not interested in buying.
For example
A luxury car brand might wish to target in-market car buyers — individuals with expiring leases or old cars — who make more than $80,000 per year. Although the CPM will rise to accommodate the cost of this targeting, and the overall scale will decrease, the advertiser is much more likely to reach real prospective customers.
Historically, this data was targetable at the demand-side platform (DSP) level. DSPs would integrate with DMPs to make these targeting options available to the traders on their platforms and to their advertising clients. However, with the increase in private marketplace transactions to greater than 50% of all programmatic spend in 2020, there is increasing pressure to move data targeting to the supply-side platform (SSP) side of the transaction since SSPs have greater visibility into the inventory of their publisher clients. This is particularly useful to target premium inventory from publishers who are only comfortable making their inventory available programmatically in private marketplaces when it is protected by services such as Audience Lock.
Moreover, publishers are starting to utilize third-party data to bolster their own first-party data with complementary audience information or simply to add value to their inventory without first-party data.
For example
A publisher who runs an auto enthusiasts site might have specialized car interest information from their users’ logins and profiles. This publisher could enhance the scale it offers advertisers by combining their first-party user information — built from users that selected luxury cars in their “favorites” profile or from clicking on a certain number of articles — with the high-income, in-market car third-party data from the example above. The publisher could then activate this data for advertisers across their entire footprint of sites, including on non-auto related content.
At SpotX, our most successful publishers are often receiving double the CPM for data-layered campaigns than for those without. The advertisers are happy because they are reaching the audiences they want to target in premium, brand-safe environments and our publishers are happy because their CPMs are so high.
What are the risks?
As we shared in our what are cookies… blog post, desktop and mobile web third-party data is currently collected primarily using cookies. You may have read that all major browsers plan to discontinue third-party cookies. While this paradigm shift won’t harm the audience-building process of CTV or in-app environments, which primarily use persistent device identifiers to build audiences, it will likely impact desktop and mobile web. Without cookies, how will the various interconnected parties in the ad tech supply chain sync audiences? There are a number of potential solutions in the preliminary design stage to help solve this industry-wide conundrum. However, none appear to be the silver bullet and none are remotely ready for primetime. Check out our blog post on the demise of third-party cookies to learn more.
Learn more about audience data
Now that we’ve covered the different parties involved in the data collection process and different types of data, we’re ready to get more granular. If you found this topic interesting I would recommend you read this post to learn about the probabilistic and deterministic methods of audience identification and the benefits of each.
This article 
was refreshed by Eric Shiffman, director of product marketing at SpotX and was originally written by Lexie Pike, product marketing manager.