Netflix has historically dismissed the notion of an advertising-supported membership tier for its service. But last week, at Morgan Stanley’s 2022 Technology, Media & Telecom Conference, Netflix’s CFO, Spencer Neumann, indicated that the company may have softened its reticence related to advertising:
It’s not like we have religion against advertising, to be clear…But that’s not something that’s in our plans right now… We have a really nice scalable subscription model, and again, never say never, but it’s not in our plan.
My read from this comment is not that Netflix is earnestly considering offering an ad-supported tier (known as AVOD in the parlance of connected TV). But, as is argued in this piece by SparrowOne, any company that operates a scaled digital product must consider advertising as a revenue model, almost as a compulsion of fiduciary duty. So whether or not Netflix has sincerely evaluated an advertising business recently, they almost certainly have at some point.
I’d argue two things regarding any potential foray into advertising by Netflix. First, that Netflix already operates a scaled personalization engine that functions as an ad network. And second, that Netflix would be better served commercially by using its advertising technology to erect a Content Fortress than by selling its placements to advertisers.
Netflix already operates a scaled personalization engine
Netflix’s personalization engine dictates the content that is recommended to users across almost the entirety of the surface area of the product. For a comprehensive overview of how Netflix uses machine learning to personalize the Netflix product experience, see this one-hour talk from a Director of Machine Learning at the company.
The influence of Netflix’s machine learning infrastructure is ubiquitous throughout the user experience in the product. And the economic benefits of this to the company are substantial: in a 2015 paper, researchers at Netflix estimate that the “combined effect of personalization and recommendations save [Netflix] more than $1B per year.” That number would almost certainly be substantially higher if reported today: according to streaming analytics company Antenna, Netflix’s active monthly churn rate is not only the lowest of the premium streaming services, but it is astonishingly stable at slightly more than 2%.
Personalization is an effective weapon in combating churn, assuming a product features enough content to address a large number of distinct audiences (I speak to this notion in Mobile ad creative: how to produce and deploy advertising creative at scale). This idea is addressed in the video linked above: part of Netflix’s investment into content production is directed by its personalization apparatus, which can determine which audiences are underserved with content on the platform. According to one analyst, Netflix may spend $19BN on content production in 2022, up from $17BN in 2021.
The components of a scaled personalization engine are not substantially different from that of an advertising platform. With the exception of a bidding system and an interface for creating campaigns, Netflix’s personalization infrastructure contains everything needed for serving personalized ads, and it could be co-opted to do that. Netflix even operates a contextual bandits-based creative optimization mechanic that surfaces the most effective creative for a given title based on genre- and theme-based preferences.
I’d argue that Netflix possesses all (or most) of the machinery required to launch an advertising business. Netflix’s personalization engine is already generating tremendous value for the company and it could be cross-utilized for serving third-party advertising to the same effect. But I don’t think Netflix should use its personalization engine for serving third-party advertising (and, frankly, I don’t think it will). Rather, I believe that Netflix could create a powerful Content Fortress using this technology that would unlock far greater commercial value than an advertising business could.
Netflix as Content Fortress
In this piece, I describe a Content Fortress as:
Ultimately, a Content Fortress fits into the post-ATT reality by subsuming third-party content interactions into a first-party setting. Note that a Content Fortress is different from a Walled Garden: a Walled Garden engages users with (often user-generated) content and monetizes that engagement by allowing advertisers to purchase ad impressions that lead the user to an external destination. The most sophisticated of these platforms then monitor the conversions that take place subsequent to ad clicks on their platforms and use that data to optimize ad campaigns on that basis.
The concept of the Content Fortress was born out of the new privacy landscape that is taking shape through platform policies like Apple’s App Tracking Transparency (ATT): because the stream of conversions data between advertisers and ad platforms — in what I call the hub-and-spoke model of advertising measurement — is severed as a policy, advertisers are incentivized to host their content on large platforms. When this happens, the platforms can observe content interactions, like purchases, in a first-party setting and use that data for ads targeting. To my mind, the canonical example of a Content Fortress is FB Shops, which allows retailers to host their storefronts within the Facebook Blue and Instagram apps. But FB Shops is not unique: the entire mobile retail experience is moving onto social platforms as a result of ATT (as I describe in more detail here). This new privacy reality helps to explain why everything is an ad network.
No company captures the promise of a Content Fortress as acutely as does Netflix. While the majority of Netflix’s engagement takes place on televisions — as of 2018, only 10% of view time took place on smartphones, with 70% taking place on televisions — Netflix has expanded its mobile product offering dramatically recently with the launch of its mobile gaming initiative. The allure of working with Netflix as a mobile games developer is clear: distribution. Netflix’s 222MM active subscribers worldwide can play its portfolio of mobile games, which is published on iOS and Android under the Netflix label, for free — but non-subscribers can’t, and they’re only offered the option of subscribing to Netflix as a native in-app purchase. I describe the dynamics of Netflix’s mobile games publishing business, and why Netflix is likely permitted to do things that are ostensibly prohibited in App Store policy, in The App Store has a ‘Too Big To Fail’ problem.
Instead of selling its personalization machinery and placements to advertisers, Netflix should offer them to content creators as a means of distribution through a publishing business. In other words, Netflix should become a Content Fortress. Netflix could use its smartphone app to distribute games and other types of content, powered by its personalization engine. And these third-party developed apps, while still published under the Netflix label, should be freemium and not gated by the Netflix paywall. As these products grow, so does Netflix’s data set of non-video streaming engagement data, allowing it to drive cross-promotion between published apps in a sort of first-party audience network. And these apps could also cross-promote users back to Netflix, potentially allowing the company to attenuate its subscriber growth slowdown.
Of course, building a Content Fortress and operating an ad network accessible by third-party advertisers are not mutually exclusive. Netflix could easily sell ad impressions in its television app and reserve the impressions in its mobile app for the exclusive use of its publishing business.
Whatever the choice, Netflix has the ability on mobile to compel developers to bring content into its first-party publishing walls, allowing it to use the power of its personalization engine to drive discovery in ways that aren’t economical for many advertisers in the post-ATT environment. Netflix should use that power to build a Content Fortress on mobile. Netflix’s strategy with its mobile games initiative seems to be to position mobile games as a value-add while it raises subscription prices. Instead, the company should utilize its audience and personalization technology to build a cross-promotion network between Netflix and its published games (and, potentially, other types of apps).