The loss of third-party cookies is often seen as a negative for the advertising industry. However, according to The Trade Desk (TTD), this may not be the case.
“Ending cookie support will bring better data, better personalization,” said Ben Sylvan, vice president of data partnerships at TTD, during a session on retail data at the eCommFronts Night Market event last week in New York. “It will lead to a much cleaner and more sustainable internet.”
The session was moderated by Olivia Giaimo, group director of e-commerce media science at Horizon Media Night Market.
Sylvan said programmatic advertising is moving away from behavioral data, which tracks the online content people consume and interact with through anonymous identifiers such as cookies. Broad demographic targeting — “using proxies for what people will buy or do” — is also falling out of favor, he said.
A step into the void is retail data.
“When I get a loyalty card from my local grocer, I give them my email and phone number,” Sylvan said. He predicts that advertisers will use hashed email addresses and phone numbers, as well as new deterministic identifiers (eg TTD Unified ID 2.0) for CTV, audio and display purchases.
In the past, third-party providers “inferred a lot about people based on one visit to a website,” Sylvan said. For example, just because someone clicks on motorcycle content once, it doesn’t mean they’re buying new bikes.
Similarly, a diaper brand that historically bought ads on linear TV using basic demographic targets (say “female” or “25-35”) was “basically playing a puzzle game” about who had kids instead of making confident data-driven decisions about which consumers to target, Sylvan said. The brand wasted impressions and media spend on people who would never buy diapers.
On the other hand, retailers don’t have to infer what people want. They have years, even decades, of purchase data that can be used to create detailed customer segments. If a diaper brand wants to reach diaper buyers from a particular retailer, such as Albertsons, Kroger, Walgreens, Sam’s Club, or Instacart, it can target a large number of verified diaper buyers and eliminate people who are not interested in diapers. audience.
Get our editors’ roundup delivered to your inbox every weekday.
One of those verified buyers might be Sylvan, who was a ramen and mac and cheese sports journalist who never bought diapers… until he became a father of three who bought diapers for more than eight consecutive years. And retailers have his transactions to demonstrate exactly what lifetime value their ads can generate.
Even non-endemic advertisers – with products not sold in chain stores, such as insurance or auto – can use retail purchase data to target consumers. For example, a credit card company may target someone for an enhanced consumer credit offer if they have recently started buying premium or expensive food, as a signal of general consumer purchasing power.
Retail data is especially useful for financial institutions, Sylvan said, because fair lending laws mean demographic targeting is off limits to them. (You can’t target women as “women,” but you can target Walgreens customers who shop for makeup and women’s shampoo.)
Not so fast
While retail data can support more targeted campaigns, it has its drawbacks.
For example, targeting still needs to be broad enough and the audience large enough to be worth reaching out to. Because the data is based on actual purchases and only from customers with signed-in accounts or credit cards that share data with the retailer, the volume threshold can be difficult to reach, even for large retailers.
“A best practice when it comes to audiences is to think at the category level,” said Sylvan.
For example, Hershey’s could create its own segments targeting M&Ms buyers and Hershey bar buyers separately. But it’s not worth it for advertisers to target every little niche, even though in theory they can target “people who only buy blue M&Ms, who dyed their hair blue with hair dye they bought at the grocery store, and who also they buy a lot of diet coke,” Sylvan said.
TTD “separated retail media from inventory,” Sylvan said. This means advertisers have more options for retailers to work with.
DSP even has some retailers offering purchase data for free – as long as advertisers use their DSP to serve ads.
But other retailers are successfully linking data to inventory and managing where ads appear, Sylvan said. He mentioned Roundel, Target’s retail media network, as a PMP solution that TTD advertisers liked because of the “ease (with) inventory selection.
Whatever advertising platform model retailers choose for their media business, TTD’s foundation is the ability to connect media spend with actual sales. That’s why, according to Sylvan, “retail audiences are the future of data when it comes to programmatic.”