We’ve all been there: You look up a product out of sheer curiosity with no intent to actually spend money on it, and suddenly you’re inundated with ads for similar products.
The challenge for advertisers is distinguishing the passive browsers from the people with genuine purchase intent.
In a recent development, zillow knows that challenge well. Its rental site often serves as a place for people to fantasize about their future home, even if they’re not actually looking to move.
In a recent development, “We are known for dreaming,” said Ravi Kandikonda, SVP of marketing for Zillow Group. But that’s not necessarily a bad thing. Welcoming the dreamers is a great way for Zillow to drive awareness of its brand and get potential future renters into its platform, Kandikonda added. From there, it’s a matter of striking a balance between building awareness and driving conversions.
That’s where Zillow turns to its media agency partner, Canvas Worldwide.
Sources indicate that in September, Canvas announced a newly introduced division called ART Dept., which provides a collection of AI solutions to help brands and agencies optimize their campaign budgeting and pacing. Zillow is using ART Dept.’s AI agents to better separate the renters from the dreamers and tailor its marketing accordingly.
According to recent reports, the “ART” in ART Dept. stands for “AI Applied Research and Technology.” (Though, if we’re being pedantic, wouldn’t that be AART – or AIART?).
Quibbles about acronyms aside, Canvas’ latest division offers a mix of AI optimization features that are becoming commonplace for ad agencies and marketing platforms. Though, like other agentic platforms, Canvas states its differentiation comes from how it applies those optimizations to its agentic layer.
For example, offering a pacing dashboard is “table stakes” by now, rather than a distinguishing feature for an agency, said Sally Lee, SVP of programmatic and ART at Canvas.
But Canvas wanted to make its pacing dashboard “multidimensional” by connecting it with a budget source of truth and building an agentic layer on top of it, stated Lee.
In a recent development, the agentic layer, called Talos, determines whether a campaign is underpacing or overpacing and optimizes for better performance. Before pushing the newly optimized campaign into the platform, a member of the Canvas team reviews and approves (or denies) it.
Canvas can “pick and choose from different search and behavioral categories” to optimize against, noted Lee. In Zillow’s case, that might include whether a user is researching renter’s insurance or looking into moving truck rentals.
Canvas gets those behavioral signals through its data partnerships, including its MCP connection to predictive audience targeting organization Dstillery.
Access to behavioral data has helped Zillow reach more specific audiences and be smarter about exactly where and when to target them, said Kandikonda.
For instance, targeting multifamily renters in Seattle using a generic, off-the-shelf “multifamily apartment buildings” audience only goes so far, he noted. “I need to know when they are on a particular platform” to bid for that audience at the right moment with the right content, he added.
Dstillery’s audience modeling predicts which websites the target audience will visit and when based on real behavioral data. And Canvas’ AI can help surface audience and behavioral overlaps that might have previously gone overlooked by its team.
Canvas’ reliance on data-backed reasoning was a major draw for Zillow when the two first started working together in 2023, Kandikonda said, and it continues to be a major focus of ART Dept.
The goal behind the newly introduced product suite is to save media teams time figuring out what media is performing and how much to spend on different audience segments.
According to recent reports, increasing ROI is crucial for any brand. But it’s particularly important within the real estate market, Kandikonda stated, as marketing dollars are constrained and the entire industry “is bouncing around the box.”.
So, while Zillow appreciates its dreamer audience, it’s feeling the same pressure most brands are to focus on lower-funnel conversions, such as “actual movers,” Kandikonda said.
That’s part of the reason off-the-shelf renter audiences weren’t cutting it for Zillow’s ads business, he added.
According to recent reports, an off-the-shelf audience may indicate people who are interested in renting more broadly, he noted, but ART was able to select for people who were actively in the process of planning a move. And more precise targeting means better lower-funnel ad performance.
The destination, not the journey.
From Zillow’s perspective, the biggest change from using Canvas’ ART Dept. is the results, rather than the process. After all, the AI tools are used by the Canvas team, rather than by Zillow.
Zillow just tells Canvas what outcome it wants to achieve. Canvas then translates the desired outcome into a scorecard that weights different goals and metrics within the interoperable tech infrastructure that powers the ART Dept., which is called +Suite.
Industry observers note that from there, Canvas’ Audience+ product (which is similar to a data management platform) connects to data partners and the Talos AI agent to build custom audiences. Finally, Talos activates and optimizes the campaign.
The system creates an ongoing feedback loop, noted Canvas’ Lee: “Build an audience, activate it, see how it’s performing against what that client cares about and identify where to make changes to drive a better outcome.”.
