“The New Walled Gardens Aren’t Just Bigger — They’re Blacker Boxes Than Ever”

For years, the term “walled garden” conjured images of Google, Meta, and Apple — massive platforms that controlled their own data and kept advertisers dependent on proprietary tools. But the new generation of walled gardens makes those legacy ecosystems look almost transparent by comparison. What’s emerging now across streaming, retail media, and connected TV isn’t just another set of platforms selling ad inventory. These are vertically integrated intelligence systems — companies that own the content, the ad server, the measurement layer, the optimization engine, and the data clean room, all under one roof, all invisible to outsiders.

Netflix is the most striking example. The company didn’t simply bolt an ad tier onto its existing product and hand the keys to a third-party ad server. It built the entire stack from scratch. The Netflix Ads Suite now includes an Audience Insights API that gives advertisers a curated window into member characteristics and viewing behaviors, a Reach Curve API for forecasting campaign delivery, and data clean room integrations with Snowflake and Amazon Web Services — with InfoSum to follow by end of 2026. On top of that infrastructure, Netflix is testing AI agents that autonomously manage, optimize, and purchase ads on the platform, while simultaneously running experiments with personalized ad loads and frequency caps that dynamically adjust based on individual viewing behavior. The company is even using AI to adapt advertiser creative assets for different placements — reformatting spots into vertical video or pause ads, and blending ad creative directly with Netflix content in tests with brands like DoorDash and Target. As Netflix VP of Advertising Nicolle Pangis put it, the suite exists because “Netflix is the only place that can leverage the best tech with the best shows and movies in the world.”

Think about what that means for competitive intelligence. When a rival brand runs a campaign on Netflix, you won’t know what creative they used, how Netflix’s AI modified it across placements, what audience segments they targeted, how frequency was capped, or what the AI did to adapt delivery in real time. That information lives inside Netflix’s clean room, processed by Netflix’s algorithms, measured by Netflix’s APIs. It never touches the open web where traditional spy tools can observe it.

Netflix isn’t alone. The entire premium video ecosystem is racing to build similar fortifications. Nine national TV publishers — including NBCUniversal, Paramount, and Warner Bros. Discovery — recently joined an initiative through OpenAP to create standardized ways to connect first-party outcome data to campaign exposure data, complete with unified conversion APIs and workflows that will eventually support agentic AI interactions. OpenAP CEO David Levy framed the effort explicitly as a response to walled garden dominance, arguing that “premium video needs its own intelligent operating layer to better compete with walled gardens.” In other words, even the publishers trying to offer an alternative to closed ecosystems are building their own proprietary intelligence layers in the process.

Meanwhile, Amazon continues to deepen its integration of retail data, streaming inventory through Prime Video, and its own DSP — creating yet another black box where purchase behavior, viewing data, and ad delivery fuse into a single system no competitor can peer into.

This is a fundamentally different competitive environment than anything performance marketers have navigated before. The intelligence asymmetry isn’t a bug in these systems — it’s the product. And it demands an entirely new approach to understanding what your competitors are doing.

“Why Transparency Is Dying Where the Money Is Flowing”

The structural incentive is elegantly simple and ruthlessly effective: walled gardens profit from opacity. When a platform controls the inventory, the audience data, the ad serving logic, and the measurement framework, it eliminates the advertiser’s ability to comparison-shop in any meaningful way. The result isn’t just a restricted marketplace — it’s a restricted epistemology. Advertisers don’t just lose access to where their ads run; they lose access to understanding why their ads perform the way they do.

This is by design, not by accident. Every walled garden has a financial interest in keeping advertisers locked into platform-specific optimization loops. If a brand running campaigns across Netflix, Amazon, and Disney+ could easily compare apples-to-apples performance data across those three environments, the inevitable result would be ruthless budget reallocation toward whichever platform delivers more efficiently. The platforms know this. So instead, each one offers its own proprietary metrics, its own attribution model, its own definition of success — ensuring that the advertiser’s “insights” are always denominated in a currency that only spends in one store.

The measurement crisis in connected TV and streaming makes this dynamic especially acute. As OpenAP CEO David Levy put it, “Premium video needs its own intelligent operating layer to better compete with walled gardens,” a frank acknowledgment that even major TV publishers recognize they’re losing the structural battle. The initiative he’s leading — bringing together nine national publishers including Paramount, NBCUniversal, and Warner Bros. Discovery to standardize how outcome data connects to campaign exposure — exists precisely because the current landscape makes cross-publisher measurement nearly impossible. Without that standardization, every advertiser is trapped evaluating each platform on the platform’s own terms.

Now layer AI on top of this asymmetry, and the gap becomes compounding. Netflix is actively testing AI agents to manage, optimize, and purchase ads on its platform, alongside personalized ad loads and frequency caps that dynamically adjust based on individual viewing behavior. On the creative side, the company is using AI to adapt advertiser assets across different placements — vertical video, pause ads, content integrations — and combining brand creative with Netflix programming in ways tested with partners like DoorDash and Target. The advertiser gets a performance report at the end. Netflix gets a continuously learning system that understands, at a granular level, which creative variations work for which audience segments in which content contexts at which moments. That intelligence doesn’t leave the garden. It accumulates inside it.

This is the asymmetry that should keep independent marketers up at night. The platform’s AI gets smarter with every campaign it runs across every advertiser. Your campaign is one data point in their model; their model is invisible to you. You receive outputs — impressions, completions, estimated conversions — but the underlying logic, the cross-advertiser patterns, the behavioral signals that actually drive those outcomes, all of that stays behind the wall. Measurement inside these gardens isn’t designed to illuminate strategy. It’s designed to justify spend. The report tells you what happened. It never tells you what your competitors did differently, what creative approaches are gaining traction in your category, or how the platform’s own algorithmic preferences shaped your delivery.

For marketers who built their competitive edge on reading the landscape — studying competitor creative, reverse-engineering what’s working, iterating faster than the next brand — this information asymmetry isn’t just an inconvenience. It’s existential. The walled garden accumulates compounding intelligence. The advertiser gets a receipt.

“The Multi-Channel Illusion: You’re Diversifying Your Spend, Not Your Intelligence”

Every growth marketing playbook published in the last two years repeats the same mantra: diversify your channels. And the advice is correct — concentrating your entire user acquisition budget inside a single platform is a well-documented path to diminishing returns, rising CPAs, and existential vulnerability to algorithm changes. But there’s a critical distinction that almost nobody makes: spreading your spend across five platforms is not the same as gaining competitive visibility across five platforms. In practice, most advertisers who “go multi-channel” are simply replicating the same blindness in more places simultaneously.

The standard framework for thinking about this — and it’s a useful one — divides the landscape into three tiers. First, the walled gardens like Meta, Google, and Apple Search Ads, where massive reach comes paired with proprietary measurement and opaque auction mechanics. Second, ad networks like Unity, AppLovin, and Mintegral, which maintain direct SDK integrations with millions of apps and offer more stable CPIs through fewer intermediaries. Third, programmatic DSPs, positioned as the “global brain” that finds high-value users across the open web through real-time bidding engines powered by machine learning. This taxonomy is genuinely helpful for budget allocation. It tells you where to spend. What it doesn’t address — what almost no multi-channel framework addresses — is what you can actually see once you’re inside each tier.

Consider what happens when you activate across all three. Inside Meta, you get your own campaign performance data — your CTRs, your ROAS, your frequency metrics — but zero visibility into what your competitors are running, how long their campaigns have been live, or which creative angles they’re testing. Inside Unity or AppLovin, the situation is arguably worse: you’re buying placements across a fragmented app ecosystem where even your own data attribution is imperfect, let alone any line of sight into rival strategies. And inside a DSP, for all its algorithmic sophistication in analyzing millions of variables to decide if a specific user is worth the bid, you still can’t pull up a competitor’s landing page, dissect their ad copy, or track their campaign cadence.

The fragmentation problem runs even deeper in premium video and connected TV. As OpenAP CEO David Levy put it when announcing a new cross-publisher measurement initiative, “Premium video needs its own intelligent operating layer to better compete with walled gardens.” When nine national TV publishers — including Paramount, NBCUniversal, and Warner Bros. Discovery — need to band together just to create a standardized way to connect outcome data to campaign exposure data, you begin to see how thoroughly balkanized the competitive intelligence landscape has become. If publishers themselves struggle to offer unified measurement across their own properties, the idea that advertisers can maintain competitive awareness across those same properties is aspirational at best.

This is the multi-channel illusion in its purest form. You’ve correctly identified that concentration risk is dangerous. You’ve spread your budget across social, ad networks, CTV, and programmatic. You’ve built dashboards tracking your own performance in each silo. But you’ve solved a portfolio problem while leaving an intelligence problem completely untouched. In every single one of those channels, you remain unable to answer the most fundamental competitive question: what is everyone else doing?

There is, however, one layer of the digital advertising ecosystem where competitive activity remains observable — where creatives, landing pages, and campaign duration are not locked behind proprietary walls. And it’s precisely in that layer where spy tools shift from being a nice-to-have research supplement to becoming the primary mechanism for competitive awareness.

“Native, Push, and Pop: The Last Transparent Battlefield”

While the biggest platforms retreat further behind proprietary walls and the TV industry scrambles to build standardized cross-publisher measurement frameworks just to approximate the visibility that walled gardens withhold, an entire category of digital advertising operates on a fundamentally different architecture — one where transparency isn’t a concession or a reform initiative, but a structural feature baked into how the technology works.

Native ads, push notification ads, and pop/popunder traffic all share a common trait that separates them from every walled garden channel: they serve ads through public-facing placements on the open web. When a native ad appears on a publisher’s site, it renders as a visible element on a crawlable page. When a push notification fires, the creative and the destination URL are discrete, observable components. When a pop or popunder opens a new browser window, the landing page it loads is a standard URL — not an in-app experience locked inside a proprietary SDK. Every link in the chain, from the creative asset to the landing page to the offer itself, exists in indexable, inspectable space.

This matters enormously for competitive intelligence. Because these ads live on the open web, they can be systematically discovered, catalogued, and analyzed by spy tools in ways that are architecturally impossible inside Meta’s feed, Netflix’s ad suite, or Amazon’s sponsored placements. A tool like Anstrex can show you the exact creative a competitor is running on a specific native network, how long it’s been active, which geographic regions it targets, and what landing page it drives traffic to. You can see whether a campaign has been running for three days or three months — a proxy for profitability that no walled garden will ever voluntarily surface. You can deconstruct the entire funnel, from headline to pre-lander to offer page, and reverse-engineer what’s working before you spend a single dollar testing it yourself.

Contrast that with the reality inside walled gardens, where as VideoWeek has documented, even TV companies are struggling to move measurement beyond a “box-ticking exercise” because opaque systems make it impossible to build useful data assets for optimizing future performance. If major broadcasters can’t get transparent measurement out of these ecosystems, individual performance advertisers certainly aren’t getting competitive intelligence out of them.

The strategic implication is counterintuitive but important: in a world where the largest advertising channels are becoming progressively darker — more proprietary, more algorithmically mediated, more resistant to outside observation — the channels that remain structurally transparent don’t just retain their value. They become disproportionately valuable. Native, push, and pop traffic aren’t legacy formats clinging to relevance. They are the last arenas where an advertiser’s skill at reading the competitive landscape translates directly into lower costs and faster iteration. The ability to see what’s running, analyze why it’s working, and adapt before your competitors do is a genuine edge — and it’s an edge that only exists where the underlying infrastructure permits observation.

The open web isn’t a limitation. It’s the last battlefield where intelligence still compounds.