Walled Gardens Represent A Measurement Problem, Not A Marketing Problem


The digital advertising industry continues to struggle with the phenomenon of “Walled Gardens,” in part because walled gardens are mistakenly viewed as a problem waiting to be solved. But in reality, walled gardens represent an additional measurement constraint, not an insurmountable marketing obstacle.

Google, Facebook, Amazon and other walled gardens have earned industry hegemon status by attracting large, loyal, logged-in audiences of consumers. These audiences power high-performance ad networks across numerous devices and formats. And rather than avoiding walled gardens, marketers should continue to include these ad networks on digital media plans for as long as a brand’s target audience continues to spend time engaged within their respective walls.

Companies like Amazon and Facebook have increasingly begun to offer robust solutions for consumer journey path analysis and ROI measurement—in part because they are in a very strong position to offer these solutions due to the vast audiences who log in to their networks. Walled gardens’ access to persistent consumer identification enables them to manage control groups and run sound A/B tests when it comes to advertising campaigns. These tests can be used to measure the ROI of media investments made within an individual walled garden.

Therein lies the true problem with walled gardens. It’s not a marketing problem—it’s a measurement problem. Marketers have more high-performance, cross-device, multi-format ad networks in front of them offering up more ROI and lift studies than ever before. Robust targeting and optimization is available within each walled garden included in the media plan. But impressive ability to measure themselves aside, herein lies their fatal flaw: walled gardens have no real way to measure each other. Without unified measurement across walled gardens and non-walled media investments alike, marketers are left without a way to compare and measure performance and ROI across walled gardens—and more importantly, across their entire media plan.

The fact that walled gardens constrain holistic media measurement should not be a surprise. Walled gardens’ job is not to offer plan-wide optimization; rather, their business is to sell as much of their own media—and that business is booming. Walled gardens rightly view user data as their “golden goose,” and protect it accordingly.

Marketers’ inability to measure their full media plan inclusive of walled gardens brings this article right back to where it began: with a critique of the industry’s misplaced efforts to “solve the problem” of walled gardens. Despite their walled nature, holistic plan-wide media ROI measurement is still possible. However, doing so through the industry’s favorite methodology—i.e. bottom-up, impression-level attribution and user matching—is off the table. Facebook is never going to be beaten at its own user-level identification and measurement game. So marketers have to change how they score the game.

Marketers can and should use a measurement methodology that does not require the type of data that walled gardens withhold. One option is to implement a series of randomized controlled experiments designed to measure media plan ROI without the use of user-level data. Another option is experimental design—an approach automated through software and directly integrated into a media buying platform such as a demand side platform (DSP). Experimental design remains the best method to date of showing the causal link between media investments and business outcomes such as actual sales or brick-and-mortar foot traffic. The experimental design approach effectively transforms the entire media plan into a controlled experiment that has no need for user-level data—and in doing so, ensures that measurement works seamlessly across the full plan, including walled gardens.

Walled gardens constrain marketers’ ability to measure media ROI, but only so far as those marketers insist on using a methodology that walled gardens are strongly incentivized and able to obstruct. Marketers who insist on taking an impression-level approach to holistic media plan ROI measurement will not succeed. Marketers open to trying a new measurement methodology however, might just find a way around those pesky walls. The Greeks found their way through Trojan walls without tumbling them to the ground. Marketers can do the same with today’s walled gardens through experimental design and plan-wide measurement.

Our solution is encoded in a product called Market Pulse. You can hear one of our customers describe it in this case study.