SPEC CASE STUDY

'N' Crowd - Social recs for Netflix

Concept feature letting friends swap Netflix picks, cutting ‘what should I watch?’ fatigue.

Role:

UX Designer

Duration:

4 weeks

Tools:

Sketch, Proto.io

Problem

Netflix’s main priority is engagement. An endless catalog leaves viewers unsure what to watch next despite a recommendations algorithm.

Goal

These were the guiding objectives before any user research.

  • Add an in-app way for subscribers to recommend titles to friends, without changing core playback.
  • Help users decide what to watch faster by surfacing human recommendations inside Netflix.
  • Ship a clickable prototype in four weeks to validate the idea with no back-end work.

Research insights

Before sketching screens I needed to know two things:

  • How people really pick something to watch (sources, pain-points, deal-breakers).
  • What “social” could add without feeling intrusive or complicated.

Methodology

  • ‍Five semi-structured interviews with frequent Netflix users (24-35, mix of heavy-series bingers and casual movie-watchers).
  • ‍Rapid competitive teardown of three social-watch products: Teleparty and Sling (sync watch-party), Spotify (share playlists and view people's listening habits), and Tronko (share movie recommendations with friends).
  • Affinity-mapping of interview quotes.

Insights

  • Friends are the #1 trusted source.
    Users rely on specific friends far more than reviewers—or Netflix’s algorithm.  
  • No one uses the "rate titles" feature.
    Two feared skewing the algorithm, three never noticed it—so star ratings can’t drive discovery.  
  • Privacy is non-negotiable.
    ‍    Every participant said: “Don’t show what I watch unless I choose to share.”
  • Existing tools miss the mark.
    Watch-parties require schedule sync; public feeds overshare. Nothing offers a private, asynchronous “send recommendation” flow.

design

My goal was a friction-free, privacy-respecting way to share titles - something existing watch-party and activity-feed tools miss. The flows had to:  

  • Require no new accounts.
  • Keep users inside Netflix.
  • Never broadcast viewing history automatically.

I called the concept N Crowd - a name that reinforced privacy and discernment.

key flows

Invite someone

Netflix Crowd dashboard showing TV and movie recommendations from friends.

From the N crowd main page click the white "Invite Someone" button. Then...

Invite someone

Invite screen showing email input and crowd member list for adding friends to your Netflix Crowd.

... enter their email address and click "Send invitation". No sign-up, no schedule sync.

Respond to invitation

When a user receives an invitation email, they accept by clicking the red button. Then...

Respond to invitation

Post-acceptance screen asking user to recommend shows to a friend who joined their N Crowd.

... arrive at their new N crowd page. Or...

Respond to invitation

If they don't want to accept, they can click the grey "Start Your Own Crowd" button.

Browse recommendations

Netflix homepage showing carousels by category: TV Shows, Dramas, and Ncrowd. Each carousel scrolls horizontally with large show thumbnails.

A new row on the homepage allows easy access to the most recent recommendations

Browse recommendations

N Crowd page showing a selected connection. Shows carousels for TV Shows they recommend, Movies they recommend, and titles you recommend for them’ with horizontal scrolling.

While more granular controls are found in the "N" Crowd hub.

Make a recommendation

Animated flow showing a user recommending ‘Moneyball’ to friends. The movie is selected, a recommendation menu appears, and the user chooses recipients from a list.

Choose a show and select who you want to recommend it to. Done.

Design tradeoffs

  • Explicit “Send Rec” vs Automatic Feed Sharing
    Chose an intentional Send button; nothing posts unless the user acts, meeting the privacy line all five interviewees drew.
  • Home-page “N" Crowd Row vs Separate Page Only    
    Added a single friends row to Netflix Home for zero-click discovery, yet kept a full "N" Crowd hub so users can browse recs by connection. The hybrid keeps visibility high without overhauling the entire home layout.

Testing

Now that the screens were done, I turned them into an interactive prototype using Proto.io. Six users completed four core tasks without guidance.
Findings & fixes:

  • Users wanted to invite multiple friends at once → allow comma-separated emails in the field.
  • alf weren’t sure a rec was meant for them, specifically → added “for you” label to headings.
  • One tester feared mis-sending to all friends → distinct “Recommend to All” button + undo option.

Bonus:* “Delete Connection” button demoted to text link to avoid accidental drama.

Animated zoomed in flow showing a user recommending ‘Moneyball’ to friends. The movie is selected, a recommendation menu appears, and the user chooses recipients from a list.

Improved recommendation UI

Lessons

  • A smart algorithm still needs human context.
    Whenever the “cost of a bad pick” feels high, users lean on people they trust, even if the platform’s recommender is world-class.
  • Prototype small, test early, refine fast.
    Even lightweight studies (5–6 users) can uncover blind spots—like unclear audience labeling or risky bulk-send defaults—long before code is written.

Next step: Curated picks

A future release could surface collections from trusted critics and filmmakers—turning Netflix into a discovery hub, not just a catalogue.

Netflix Feeds landing page displaying a carousel of filmmakers (Paul Thomas Anderson, Jordan Peele, Greta Gerwig, Alexander Payne, Guillermo del Toro, Quentin Tarantino) whose curated recommendations users can browse. Below that is a Top Critics row featuring A.O. Scott, Richard Brody, Stephanie Zacharek, and others.

An assortment of lists curated by trusted sources.

Artist pages where they recommend the titles that affected and inspired them.