Coming next: The Second Factory →

Back to Library
Adoption & Product✓ Principle

The Netflix Principle (Recommend, Don't Force)

The Problem

Your data catalog has 2,154 reports. A new analyst joins: "Where do I start?" You have no answer. They're overwhelmed. Or worse: they keep asking YOU for recommendations 50 times. "What report should I use for sales?" You become a human concierge. Doesn't scale. Meanwhile, great reports sit unused because new users don't know they exist. It's like a library with no recommended reading list—just 10,000 books on random shelves. Users need help navigating, not a giant catalog.

The Principle

Build a recommendation engine. Netflix doesn't show you 10,000 movies—it shows you 20 you'll probably like based on what others watched. Same with analytics. Track who uses what. Recommend reports based on: user's department, role, what similar users view, trending/popular reports. Surface the right 5 reports for each user, not a catalog of 2,000. The magic happens when users find great reports without asking you. Think retail: "Customers who viewed this also viewed..." That's how you scale discovery.

🎯

3 Steps to Take Action

  1. 1

    **Track Usage Patterns** Instrument your analytics portal: - Who views what reports - When (time of day, day of week) - Sequence (report A → report B) - Duration (time on report) - User attributes (department, role, location) Build a usage database. This is your recommendation engine fuel. Minimum data: user_id, report_id, timestamp, duration.

  2. 2

    **Build Simple Recommendations** Start with basic recommendations on portal homepage: **Personalized:** - 'Popular in [user's department]' - 'Your team frequently uses' - 'Based on your recent views' **General:** - 'Trending this week' (most views) - 'Recently updated' - 'Highest rated' **Collaborative Filtering:** - 'Users like you also viewed' (similar usage patterns) - 'Frequently viewed together' Don't need ML—simple rules work. Test with 5 recommendations per user.

  3. 3

    **Measure Discovery Impact** Track: - % of users who click recommendations - Reports discovered via recommendations vs search - Time to first useful report (for new users) - Hidden gem reports that get surfaced Goal: 40%+ of report discoveries via recommendations (not search/asking). Iterate based on click-through rates. Promote what works. Kill recommendations nobody clicks. Bonus: Surface your best reports to more users without building 50 more reports.

💡 Inspired By

Netflix recommendation engine, Amazon "Customers also bought," and A major bank's vision for collaborative BI discovery (2016).

This is what Analytics Pitfalls & Principles is about: people helping people.

In Practice

A financial services analytics team had 1,847 reports with terrible adoption. New hires spent 2 weeks just figuring out which reports to use. The team built "BI for BI"—usage analytics on their analytics. Tracked: who uses what, when, for how long. Built a recommendation engine: "Popular in Sales," "Your team uses these," "Trending this week," "Frequently viewed with this report." Added it to their portal homepage. Time-to-first-useful-report for new hires dropped from 2 weeks to 2 days. Report discovery jumped 310% (users finding reports via recommendations vs search). Bonus: recommendations surfaced hidden gems—great reports that were buried got traffic. Netflix-style discovery without hiring a concierge team.

More on Adoption & Product

From the book

In the book there is a demo that wows the room and helps nobody, and a rough mockup that changes everything. The story is the difference between them.

This card is one of 57 drawn from The Auction Block by Brian Cariveau — the story of an analytics leader with 90 days to fix what the cards describe.

Each leadership pattern falls into a few pitfalls most often. Find your pattern.The Analytics Leadership Mirror: free, 5 minutes, no account