Account-based marketing (ABM) is a marketing strategy in which web advertising is targeted at specific employees of specific companies. For instance, consider a retail software company that is seeking to expand into the European market. They may create an ad campaign that specifically targets fashion brands who have a presence at Paris Fashion Week, particularly those employees who will be attending the event.  

While such targeting may seem to be the polar opposite of privacy, the Privacy Sandbox’s Protected Audience API (PAAPI) can actually target specific companies (i.e. accounts) while still providing meaningful privacy protections. To understand the value of interest groups, let’s first examine how this type of targeting works in the current state of advertising and then how it will change for the better using the Privacy Sandbox’s proposed technology. 

Understanding ABM Today

To start with an example, take Jane, a fictional CEO of Big Company Inc., who would be a prime target for this type of advertising campaign. Let’s say Jane logs in to a web browser using her work email and visits a website that has a DSP’s tracking pixel on it. At this point, the pixel running on that site records several details about Jane, such as her Big Company Inc. web browser account, her CEO title, and any other publicly available information. 

Next, Jane visits the cooking column of a publisher’s website who wishes to sell an ad. In the current world, that publisher then provides that DSP who just recorded her information with a cookie, where it learns that Jane is also currently reading a recipe for a strawberry smoothie. After she’s visited a few more publishers, a marketer could start to build a relatively comprehensive picture of Jane’s life, even doing so unintentionally using a machine learning model trained with user browsing data as input. This picture could then be used to further target advertising toward her, such as showing maternity clothes and baby cribs under the assumption, based on her purchasing history, that she’s pregnant.

This type of hyper-targeted, assumption-based advertising is what the Privacy Sandbox’s technology is working to change, particularly with the advent of “interest groups.” In the Privacy Sandbox, an interest group is a piece of data stored within the browser that can be used to target ads. While it’s natural to assume members of an interest group all share similar interests, these groupings more indicate our interest in showing ads to this set of users. Interest groups persist in the browser for 30 days and may be refreshed once per day.  Although ad techs can store private data inside of interest groups, there is no mechanism to retrieve that information later.  

Interest Groups at Work

In the previous scenario, rather than using cookies, Jane’s details from her first interaction would be added to an interest group that stores that information. The ad bidding process would then use this interest group within the browser to target the ad. The only record received would be that someone visited the strawberry smoothie recipe page, rather than all of Jane’s personal information. The bidding function may make an extra-high bid because Jane is a CEO, but the ad shown to her must be identical to the same ad that is shown to many other browsers (this is referred to as k-anonymity).  If an ad is shown, the advertiser will be able to see the bid price and the details of the page on which it is shown. It is possible that the bid price will indicate the ad was shown to a person of high relevance, but the bid price itself is rounded such that no detailed information can be obtained about the person or the browser. If enough ads are shown to other Big Company Inc. employees (k-anonymity again), then an advertiser will be able to identify and report on which ads were shown to the Big Company Inc. account.

The critical value of interest groups is enabling advertisers to use private information during the ad bidding process. However, while this personal user information can be stored by the browser to help inform ad bidding decisions, the Privacy Sandbox developed interest groups to ensure this data cannot then leave the browser to be used elsewhere.

Why It Matters

Privacy Sandbox provides a mechanism for account-based marketing advertisement that will provide meaningful privacy protections while still enabling marketers to selectively target specific accounts.  While other approaches to ABM without third party cookies exist, most notably browser fingerprinting, those approaches do not provide privacy protections and we can anticipate that over time they will be blocked by improving privacy protections.  Future-proofing our business requires that we embrace privacy protections, and the Privacy Sandbox gives us what we need.

David Roundy is a Senior Software Engineer at NextRoll.

Cristian Rojas is an Engineer at NextRoll.