Identity resolution is the lifeblood of retail media

Retailers who can connect an individual or household identity to a purchase are positioned to fill the gap left by third-party cookies. The ability to understand and connect with the customer at a foundational level is the very lifeblood that will sustain and power effective retail media strategies.

Not all cookies are bad

Web sites have used small files containing discreet data strings since 1994. This was an existing computer science pattern even before Lou Montulli employed them at Netscape and applied for the first “cookie” patent.

Cookies can be used for lots of “good” things, like remembering the user is logged into the site or communicate back to the server how often a user visits.

“Third party” cookies are used to track behavior across multiple sites over time.

Imagine you are browsing an article on The New York Times Web site. At the top of the page is an advertisement. That ad gets served by a third party. There’s an exchange of data between the browser, the NYTimes.com server and the server hosting the ad.

Then you stop reading The Times and go to Yahoo. The same cookie can be employed by the ad tech provider to track you had just visited The Times, along with a host of other information.

This is a simplified version of how ad technology can track your behavior throughout the Web[1].

“Give me three data points and I’ll find you.” – Anonymous Engineer

It’s not necessary to have personally identifiable information in the cookie to track a user. “Fingerprinting” is the practice of collecting a confluence of variables, such as IP address, browser version, screen size and operating system, to create an independent identifier.

Ad networks can also embed their own identifier in a cookie to help figure out who you are over time.

These various methods of collecting digital identity signals are the fuel for the engines of digital advertising, powering complex marketplaces where ads are bought and sold in real-time.

How do the ads do this?

To understand how these identity signals are significant, we also need to understand the world of ad networks.

Digital advertising is bought and sold on exchanges that more closely resemble stock market day trading than the days of Mad Men.

Going back to The New York Times example, the ad that appears could be the result of multiple bids on your data. The times is a publisher, essentially a Web site (or App) with advertising inventory to be sold.

A real-time ad exchange is a marketplace for publishers and advertisers to buy and sell ad inventory. Advertisers can bid on certain audiences, targeting things like demographics or behavior, and the publishers can accept those bids to fulfill the ads.

Just like day-trading, the parties involved need to have a tight strategy going into the exchange. It moves too quickly for a human to make decisions on the details. Those details get set ahead of time and executed during the exchange process.

I was amazed to discover bids might go through dozens of platforms in milliseconds before the ad is served.

Now imagine the confluence of ad networks logging various data points over time to your “fingerprint.” Smart networks record competitor IDs alongside their own and tease out overlaps. 

With billions of ads shown to billions of users, the networks develop a robust collection of data, forming probabilistic graph of user identities. The network may not literally know your name, rank and serial number, but they have a pretty good idea of where your computer is located and what you look at and purchase on it.  

Loyalty data is better than cookie data

Retail media can cut through the complexity and unreliability of cookies. If a user is transacting online directly, it’s straight forward to connect an ad view with an online cart purchase.

But what makes retail media exciting is its loyalty data. A retail loyalty program is an exchange of data with a user in return for incentives, like discounts. When a customer connects their loyalty account at the point of purchase, the retailer can match the transaction to an identity.

Unlike the probabilistic nature of cookie-based tracking or fingerprinting, which infers identity and behavior based on digital signals, loyalty data provides a deterministic link between a customer's identity and their actual purchase history. This verified connection allows for much more accurate targeting, personalized messaging, and, critically, precise closed-loop measurement of ad effectiveness directly tied to transactions.

While cookies offered a temporary pulse for digital tracking, the verified identity provided by loyalty programs is the true, sustainable lifeblood that fuels precise targeting and measurement in retail media.

Combining Ad Tech and Loyalty Data for stronger identity resolution

Retail media networks can leverage both the ad tech tracking data and the customer’s loyalty data to create a much stronger marketing profile.

Here’s how this works. Imagine you go to your local Gertie’s Grocers and swipe your Gertie’s Guests loyalty card before buying a can of house brand soup[2].

Gertie’s now knows that you, with Gertie’s Guests loyalty number GG123, are a likely soup buyer.

The corporate behemoth, Warhol Soups, is creating a targeted campaign to target its competition, Good Soups. They have paid Gertie’s Media Network to target likely soup buyers on a series of digital ads.

Gertie’s RMN whips up a targeted audience and an ad campaign. A user sees the ad and the Ad Network assigns that user ID AN456. When the user clicks to Gertie’s site or makes a purchase in the store, Gertie’s can correlate their loyalty ID with their ad network ID.  

Now Gertie’s RMN knows GG123 is also AN456.

Gertie’s gets a sale. Warhol Soups gets a sale. Gertie’s RMN demonstrates return on ad spend (ROAS) and gets more ad business. But crucially, Gertie’s RMN has collected more identity signals to better target users in the future.

Beyond the value of everyone making money, the ultimate motivating factor, Retail Media Networks can match the deterministic loyalty data, like name and physical address, with the ad network’s probabilistic data. This powerful combination, fueled by the constant flow of first-party identity as its lifeblood, creates a dynamic and highly effective marketing profile.

Conclusion

The era of relying solely on inferred data and fragmented tracking is giving way to a more precise, identity-driven approach in retail media. By centering around first-party data gleaned from loyalty programs, retailers gain a distinct competitive advantage and unlock significant new revenue streams. For advertisers, this means moving beyond guesswork to reach known customers with relevant messages and measure the direct impact on sales. As Retail Media Networks continue to mature, the strength of the underlying identity framework will be the essential engine driving profitability and proving incremental return in an increasingly complex digital world.

[1] I have been careful to craft a Web example, but tracking between mobile apps employs similar, albeit technically different tactics. Although Google has waffled about third-party cookie support in Chrome for several years, Apple is a privacy pioneer. iPhones have offered users an opt-out of cross-app tracking since 2021.

[2] One purchase doesn’t make a pattern or propensity, but we will keep it simple for the example.