Netflix Watched Feature

Background

Part of Netflix’s recommendation engine relies on watch history to personalize the experience. When a user’s watch history is empty the platform has no information to work with.

While this makes sense for a brand new user, it is a frustrating experience for someone who has simply created a new account. This is especially true for the password-sharing displaced user: someone who has watched Netflix for years on a family member or friend’s account and now has their own subscription. They know the platform, have established taste preferences, and have likely already watched several series and movies. The result is a homepage full of recommendations for shows they have already finished and a cluttered, harder-to-navigate experience.

The Already Watched feature lets users manually mark content by movie, series, season, or individual episode as watched, giving Netflix the signal it needs to deliver a personalized experience. It not only benefits new users who may have an existing watch history on other platforms but also benefits long-time users who want to tidy their history.

Problem Statements

1. I am a user who was removed from a shared account after Netflix’s password-sharing crackdown and now have my own subscription. I am trying to get a personalized Netflix experience. But my recommendations are generic and show me movies and series I’ve already watched because Netflix has no record of my viewing history, which makes me feel frustrated and like I have to start over on a platform I have already been using for years.

1. I am a new user who watches a lot of content across multiple streaming platforms, some of which are duplicated on Netflix. I am trying to get relevant recommendations but the platform has no way of knowing what I have already watched elsewhere, which makes me feel like I have to manually dig through content I have no interest in before the platform starts to understand my taste.

Goals

• Get users to actually use the feature. We will measure success by tracking how many new users mark something as already watched within their first 30 days.
• Help new users get useful recommendations faster. The sooner Netflix knows what someone has already seen, the sooner it can show them things they actually want to watch.
• Reduce friction for users re-engaging with a long-running series by ensuring progress bars reflect their actual watch state.

Non Goals

• This feature does not import watch history from other streaming platforms like Hulu or Apple TV. Users mark content manually.
• Marking something as watched is not a rating. It does not affect a user’s thumbs up or thumbs down history.
• Watch history marked via Already Watched is private. It is not visible to other profiles on the same account.
• Bulk import via CSV or any other file upload is not part of V1.

Hypothesis

If we allow users to manually mark content as already watched, new users will receive relevant recommendations faster, leading to higher feature adoption and longer session times as users engage with content that actually matches their taste.

Vision Narrative

Sophia has been watching Netflix for six years, on her college roommate’s account, then her parents’, and most recently her ex-boyfriend’s. She knows the platform well, has strong taste preferences, and has already watched several series from start to finish.

Last month, Netflix locked her out of the shared account. She signed up for her own subscription, opened the app, and was greeted by a homepage full of shows she has already seen. The continue-watching row is empty. Nothing on the screen reflects who she actually is as a viewer.

During account setup she is prompted with a simple screen: “Tell us what you’ve already seen.” She taps through a grid of popular titles quickly. Stranger Things: all seasons. Peaky Blinders: all seasons. Bridgerton: Seasons 1 and 2 only. The whole process takes about 90 seconds and her homepage already feels more relevant.

A few weeks later she remembers she also finished Schitt’s Creek last year. She opens the show page, taps the Already Watched button, and marks all six seasons. Her recommendations update immediately. The feature is always there when she remembers something she forgot to add.

Rough Scoping & Timeline

V1 Core Experience

• Entry points: A checkmark button on the show or movie detail page alongside Play and More Info, and as an option inside the three-dot overflow menu.
• Selection flow for movies: Tapping Already Watched on a movie marks it as watched immediately with no additional steps needed.
• Selection flow for series: Bottom sheet on mobile, modal on desktop and TV. Users choose between All Seasons, Specific Seasons, or Specific Episodes. Selecting Specific Episodes closes the modal and activates a selection mode directly on the episode list.
• Visual feedback: Episode progress bars fill to 100% using the existing red progress bar. No new visual language introduced.
• Platforms: iOS, Android, Web, TV apps.
• Onboarding: Optional prompt during account setup for new accounts. Skippable.

V2 Enhancements

• Ability to remove or edit an Already Watched marking from the profile history page.
• Contextual prompts, for example when a user navigates to Season 2 of a show with no Season 1 history.

Rollout & Testing

• A/B test with a subset of new accounts. Control group sees standard onboarding, test group sees the Already Watched prompt during setup.
• A/B test with a subset of existing accounts to measure whether long-time users use the feature to tidy their history and whether it leads to improved engagement.
• Primary metric: feature adoption rate within 30 days for new accounts.
• Secondary metrics: home row engagement at Day 7 and Day 30.

Key Trade Offs & Decisions

• Episode selection is inline, not inside the modal. Users recognize episode thumbnails more reliably than titles alone, and interacting directly with the episode list leverages existing UI patterns rather than introducing a duplicate list inside a modal.
• Already Watched uses the existing red progress bar at 100% fill rather than introducing a new visual indicator. This keeps the visual language consistent and avoids adding new design complexity.
• Marking something as watched is not a rating. Combining watch state with sentiment in the same interaction would add friction and likely reduce how many users complete the flow.