The Drum Awards Festival - Extended Deadline

-d -h -min -sec

Future of TV Media Netflix

As Netflix cracks down on password sharing, its ad profiles will also improve


By Eli Heath, Head of identity

September 19, 2023 | 9 min read

Eli Heath, head of identity at Lotame offers an alternative theory as to why Netflix is cracking down on password sharing - it can never truly know who its ads reach if it doesn't.

Netflix screen

/ Pexels

In the face of third-party cookie deprecation, industry leaders have devoted particular attention to email-based solutions. The reasoning seems straightforward on the surface. With an email login, you can draw a line between the browser or device and the consumer profile of that logged-in user. Also, there's a mass perception that email is fully interoperable with consumer data platforms (CDPs) and data clean rooms, which would align with myriad business goals in capturing first-party data.

But what happens when 43% of your real, paying audience shares logins?

That’s the situation Netflix found itself in – one where a significant share of logins and subsequent program viewing activity was blended across paying subscribers and friends and family masquerading behind the account holder's email.

So often, it didn't actually know who it was reaching.

Powered by AI

Explore frequently asked questions

Netflix’s latest earnings report revealed how it used a password-sharing crackdown program to identify non-paying customers and present a low-cost, ad-supported option to those engaged non-subscribers, which helped fuel 5.9m new subscriptions in Q2. Essentially, Netflix employed predictive, probabilistic methods to solve its identity challenges and fuel growth. Leaning into these same machine learning (ML) technologies to complement email-based solutions will solve addressability challenges in our ecosystem, and drive positive outcomes for both advertisers and publishers.

Along the road to the inevitable deprecation of third-party cookies, the industry has come to focus on deterministic email-based solutions to make connections across domains – the natural extension of the first-party data capture movement.

But email is not the Holy Grail of identity.

Netflix’s efforts to quash the common practice of password-sharing (79% of US consumers do it on one platform or another) have showcased one key flaw.

Email availability is lacking across the open web, with only 10-20% of the content on the open web using an authentication layer, such as a registration gateway or paywall. Email can’t be the be-all and end-all of identity.

This discussion has only furthered a perceived bifurcation of data – deterministic versus probabilistic. Let’s emphasize perceived. It’s a false dichotomy, and the two classes of identifiers aren’t adversaries of each other.

In reality, programmatic has always relied on a combination of deterministic and probabilistic methods – from targeting to frequency capping and measurement – and it will continue to do so. Layering the two together delivers addressability, scale, and interoperability more efficiently and effectively than either on its own.

Embracing these methodologies as complementary will be key to solving for identity in an environment where multiple identifiers are angling for a piece of the digital market — and it’s anyone’s guess which will survive the next decade.

Logins are only the beginning of solving for identity

As an industry, we need to move away from two pervasive myths: first, the misconception that deterministic data is sufficient on its own as an identity solution, and second, the fiction that a probabilistic identity solution or hybrid approach sacrifices accuracy or consumer privacy.

Email data indeed serves as a powerful foundation for identity, and where it lacks in scale, it trains and validates machine-learning models to deliver an expanded view of the consumer, increases addressability to engage with prospective and existing customers, and, importantly, ensures healthy publisher monetization of open web content.

Requiring consumers to disclose emails to websites changes the fundamental internet experience, turning everything into a quid pro quo – personal information in exchange for content.

Can we expect consumers to trade their email addresses for a watermelon-feta salad recipe?

Embracing probabilistic methods that can generate a pseudonymous identity to publisher traffic and assign audience attributes (e.g., 35 years old, New Jersey resident, in-market for SUVs) creates buyer demand for ad inventory, and feeds a virtuous cycle of healthy yield for the publisher’s business.

Probabilistic methods fill in the consumer profile

Remember the Netflix login quandary.

When multiple users share the same email login, deterministic-focused solutions pool together all of those users’ attributes into a single, inaccurate profile, which could lead to a diluted view of the account owner, in terms of interest, purchase behavior, and other important factors marketers rely on to analyze consumers, build audience segments, and run targeted advertising. We can all understand how this dilution plays out: Imagine a 60-something educator in Cleveland with a penchant for Korean dramas shares a streaming password with his daughter, a 30-something engineer and armchair detective in Portland.

Deprecated seed data leads to irrelevant targeting and wasted spend. While probabilistic and deterministic each has advantages and disadvantages, the advantages of one help mitigate the disadvantages of the other. Let’s not downplay the role of machine learning in refining deterministic data sets. Probabilistic signals – such as IP address, geolocation, timestamps, and user agent – combined with machine learning will root out false positives and help maintain profile accuracy.

How Netflix solved its own problem

There is indeed a “next chapter” in Netflix’s shared-login saga. We discussed how the company observed subscriber growth plateauing and began implementing a password-sharing crackdown. To kick start growth once again, it enforced new policies by combining and looking at IP addresses, device IDs, and signed-in device-level account activity to determine if an account is being used in the primary account holder’s household. It worked, and Netflix began gaining new subscriptions.

For Netflix, probabilistic methods solved a business challenge and helped drive growth. While I don’t mean to conflate the nuances across different identity-related use cases, our ecosystem can apply these same tools to tackle addressability challenges and to drive growth in the long run.

Industry leaders need to take a hybrid approach to drive the best value from the reliable and compliant data available to them. Doing so will point to a better future not only for identity, but for the entire digital landscape. Marketers will reach their goals of running personalized ads at scale. Publishers will reach their goals of monetizing the quality content necessary for fueling the healthy, free ad-supported internet users depend on.

Suggested newsletters for you

Daily Briefing


Catch up on the most important stories of the day, curated by our editorial team.

Ads of the Week


See the best ads of the last week - all in one place.

The Drum Insider

Once a month

Learn how to pitch to our editors and get published on The Drum.

Future of TV Media Netflix

More from Future of TV

View all


Industry insights

View all
Add your own content +