In January 2017, AdExchanger ran a headline that could have been written any year since: Facebook measurement, the walled garden is cracking a window.
That’s the whole walled garden debate in nine words. A giant platform with logged-in users and closed-loop data lets advertisers see a little more, the industry calls it progress, and the wall stays up.
We went back through the past decade of the trade press in our archive to see how that argument has evolved. The short answer is that it hasn’t. What changed is who’s behind the walls.
2017: “Can anyone break the duopoly?”
In 2017, “walled garden” meant two companies. The trade press called them the duopoly, and the open question was who might challenge it.
The candidates came and went. One agency executive argued that Snapchat could break the Google-Facebook duopoly. AdExchanger wondered whether Verizon’s Oath might break it. Another contributor argued that as the duopoly fragments, agencies win.
Even then, the challengers were careful about the label. AOL said it didn’t want to be a walled garden, and Time Inc.’s Viant argued that “people-based” doesn’t have to mean “walled garden.” Publishers, meanwhile, were honest about the appeal. Digiday asked them what they admired about the duopoly, and the answer was mostly logged-in scale and data.
None of the challengers broke it. Snap became a solid niche platform. Oath was folded and sold. The duopoly didn’t fragment. It got company.
2018: The duopoly becomes a triopoly
The new member was Amazon. Digiday had already called it a true frenemy for retailers in 2017, and by December, AdExchanger noted that Kroger wanted to be a walled garden too. By September 2018 the word was “triopoly,” and Philips said “there’s no need for middlemen” because it now worked directly with all three.
The walls showed up in television at the same time. In December 2018, Digiday reported that ad buyers were frustrated by the limitations of OTT walled gardens. The same year, AdExchanger collected predictions from TV ad chiefs that included “Netflix will have ads,” and when Netflix tested ads that August, people freaked out. It took a few more years, but the prediction came true, and Netflix built its own garden when it did.
2019: The gardens win the privacy fight
Then came the part of the story that should have ended the debate. Privacy regulation and browser changes were supposed to hurt data-hungry platforms. They hurt everyone else more.
In January 2019, AdExchanger’s roundup headline was that walled gardens weathered the privacy storm. By May it was explaining how the walled gardens were eating open programmatic, and Digiday asked why advertisers find it so hard to quit them.
The logic is simple once you see it. A platform with logged-in users doesn’t need third-party cookies to recognize people, and it doesn’t need to share data with anyone to measure what happened. Every rule that made cross-site tracking harder made the gardens relatively more valuable. Then Apple, in 2021, did the same thing on mobile with App Tracking Transparency, which Digiday called a power play of throttled competition and concentrated leverage. That made Apple one more garden, and a very large one.
2023–2024: Everyone builds a garden
By the early 2020s, the question was no longer how to break the walls. It was how to build one.
Every company with logged-in customers and transaction data reached the same conclusion. Retailers built retail media networks. Streaming services built ad tiers. Commerce and delivery apps built ad businesses. AdExchanger asked whether embracing open programmatic would be retail media’s savior or its downfall, which is really the walled garden question asked from the inside.
The data layer followed the same path. In January 2023, AdExchanger asked whether the walled gardens would turn clean room technology into yet another platform plaything. Independent clean room companies like Samooha were working their way through the walled gardens, and Digiday asked whether an agency’s AI partnership with a cloud provider was another loosened brick in the walls. Notice the verbs. Cracking a window in 2017. Loosening a brick in 2023. Six years of small openings, and the wall still standing.
The independents found their pitch in the contrast. In August 2023, AdExchanger noted that The Trade Desk was powering on with its “we’re not a walled garden” pitch, even through an ad slump.
By February 2024, the gardens had started running into each other. When Walmart agreed to buy Vizio, AdExchanger’s headline was simply when walled gardens collide.
What actually changed
Read all of this in one sitting and the argument sounds stuck. In a sense it is. But three things did change, and they matter more than the headlines suggest.
The walls got shorter where it counts. In September 2023, AdExchanger explained why Facebook, Google and Amazon were embracing media mix modeling. That’s a significant concession. Media mix modeling works on aggregated data across every channel, which means the platforms are accepting being measured alongside everyone else rather than only by their own numbers.
The gardens are more numerous, so none of them is the whole market. A marketer in 2017 who wanted to reach people at scale had two real choices. A marketer today can buy logged-in audiences from search, social, several retailers, streaming services and commerce apps. Each one is closed, but there are enough of them that the buyer has leverage again, because none of them can be the only answer.
The real wall is consent and identity, not the company. The thing that makes a garden a garden is a direct, logged-in relationship with the customer. That’s no longer something only platforms have. Any brand or publisher that earns that relationship has the same raw material, in smaller quantity.
What we’d tell a marketer now
After a decade of this argument, we think “open versus walled” is the wrong question. The useful questions are narrower and more practical.
Can you verify it? Every garden should be willing to be measured by something other than itself, through media mix modeling, incrementality tests or a clean room you have real access to. If it won’t, discount its numbers accordingly.
Can you take your learning with you? The risk of a garden isn’t that it’s closed. It’s that everything you learn about your customers stays inside it. Insist on getting audience and conversion insights back in a form you can use elsewhere.
Are you building your own? The brands and publishers in the strongest position today are the ones that took the gardens’ lesson and applied it to themselves: a direct, consented relationship with their customers, and the data infrastructure to use it.
The walls aren’t coming down. That debate is over, and the gardens won it. The better news is that the thing that made them powerful turned out not to be exclusive. It was a relationship with the customer, and you can build one of those too.
Every headline in this piece links to the original story. Browse the archive’s coverage of walled gardens, the duopoly and media mix modeling.
Kevin Reynolds is the founder of Advertising Perspectives, MadTech Consulting and data4orce.ai. He is a marketing technology architect and C-level advisor, with a focus on Salesforce and data-driven advertising.