Why We Fail: What I Learned From 5 Years with “Friends” — Netflix’s Social Strategy

Netflix’s “Friends” (Born 2004, died 2010)

I write a lot about product strategy, using Netflix as an example so that others can learn from the company’s success and failure. I often highlight that half of Netflix’s high-level product strategies fail. I do this to help product leaders understand how hard it is to launch and grow startups. I also think it’s important to communicate how challenging it is to blend art and science, to effectively capture “lightning in a bottle,” which is how I describe the challenge of building products using consumer science techniques.

This essay focuses on one of Netflix’s mistakes and tries to answer the question: Why did Netflix fail so often? The simple answer:

1.) As humans, there are many biases that cloud our judgment, and

2.) Inventing the future is hard.

Given this second point, swiftly removing the losers and doubling down on the winners is critical and is highly dependent on good judgment. I wrote this essay so you will be more conscious of the things that cloud your judgment. With a little luck and increased discipline, these ideas should make building products a bit easier for you.

Netflix launched “Friends” in 2004. That year, Facebook grew from one million to six million members. Because of Facebook’s meteoric rise, the persistent question from Silicon Valley venture capitalists was, “What’s your social strategy?”

With Friends, we believed we could delight customers in hard-to-copy, margin-enhancing ways. Netflix members would delight in movie ideas from friends, we would build a hard-to-copy network effect, and movie suggestions would be cheaper as friends suggested “long-tail” movies to one another.

While we hoped to improve retention, our easier-to-move proxy metric was the percentage of members who connected to at least one friend. At launch, we engaged two percent of members, and after four years, we got to eight percent. But in 2010 — I was at Chegg by that time — Netflix killed Friends because they recognized they needed at least twenty percent of members to engage. Below this threshold, it’s hard to improve retention. After nearly six years of effort, Netflix was still far from that threshold.

Yes, Friends failed to get big enough to improve retention, but looking back, I think the real question is, “Why did we persist so long?” Here’s my answer:

  • CEO support for the project. Reed Hasting’s confidence in the idea contributed to our persistence. It’s hard to quit when the CEO is passionate about an idea.
  • As a strategy, social makes sense. Friends promised to fulfill the holy troika: to delight customers in hard-to-copy, margin-enhancing ways.
  • Small wins cloud judgment. Our proxy metric kept going up and to the right, but we failed to recognize early enough that it would never be big enough to matter.
  • Product leadership requires optimism. You need hope to overcome the inevitable challenges. Product leaders tend to see the positive in projects and to ignore the negative. It’s a natural, almost necessary bias for successful product leaders.
  • We assumed the failure was in our execution, not the idea. We initiated Friends before Facebook opened its social graph and we had many industry-specific challenges, so the project was hard to execute. For instance, the “U.S. Video Privacy Protection Act” required explicit, opt-in permission to share DVD rental history with others, creating friction as members tried to connect. Our focus on addressing challenges like this distracted us from a more careful evaluation of the social strategy itself.
  • Paradoxically, small success makes it hard to kill features. Our job was to delight members, and we worried that killing Friends would do the opposite. In 2009, we faced a revolt when we removed “Profiles,” a feature that enabled members to manage multiple queues, even though only two percent of members used this feature. As it turned out, Profiles users were highly passionate as they believed that merging their accounts would lead to divorce (really!). Netflix eventually reinstated Profiles. (Did I mention that most of our board members used Profiles?)

Here’s what I hope you will take away from Netflix’s misplaced persistence in the face of long-term failure:

  • Behold the idea and not the source. Friends got extra time and investment because our CEO stood behind the idea. It’s common to give ideas extra weight when the CEO believes in them. But Reed was wrong. That happens. In this case, we needed to do a better job of evaluating the idea on its own merit.
  • Beware conventional wisdom. Thought leaders in Silicon Valley reinforced the value of a social strategy. And “social” worked well for music, which didn’t seem too different from movies. But, as we slowly learned, no one wants to reveal all the movies they watch. (“Did you really watch Adam Sandler’s The Ridiculous Six!?”) But even more critical, movie tastes are amazingly unique. Given this inherent diversity, reality eventually came into focus: your friends have sucky movie taste.
  • Temper your pride in ownership. As builders, we love to build stuff. And we don’t like to kill projects. Passion and hope cloud our judgment.
  • Clear objectives can help. We had a proxy metric, but no timeline for when we hoped to achieve a specific goal, something like “ten percent member engagement within two years.” Setting a goal guards against “we’ll figure it out next quarter” youthful enthusiasm. I liken it to the “Noontime Turnaround Rule” for summit day climbers on Everest. A climber is required to turn around at noon, even if they are one hundred yards from the summit at that time. The rule exists because a climber’s judgment becomes clouded by lack of oxygen.
  • Take your blinders off. As product leaders, we so strongly focus on the goal that we become blind to new data that should cause us to change course. Successful projects develop momentum quickly. Customers enthusiastically embrace good ideas despite early shortcomings. Tunnel vision prevented us from accepting failure earlier.
  • Ignore sunk cost. “We’ve invested so much, how can we give up now?” was a frequent refrain. Don’t let past investment in an effort inform future investment. Ask yourself, “Given what we know today, how much should we invest going forward?”
  • Be open to better ideas. In the end, we diverted the Friends team to another project. It took the judgment of an “unclouded” executive to encourage us to move on.
  • No one likes to be the “heavy.” For those of us in executive roles, our job is to provide unbiased judgment, free of pride in ownership. But no one wants to be the person who kills a five-year effort. But for many reading this essay, that’s your job: embrace it.
  • Learn to kill projects. Killing projects is a critical, rarely-practiced discipline. Netflix eventually “euthanized” Friends. Netflix let members know far in advance that it was stopping the effort, explained the “why,” gave customers recourse, and over-communicated this context, again and again. There was no member revolt.
  • Scrape the barnacle. One mistake Netflix did not make: when they killed Friends, they “scraped it” from the site. Like boats with barnacle-infested bottoms, companies that fail to “scrape” failed projects move slower and slower as they confront edge cases caused by needless, enduring complexity. This concept of a barnacle extends to “middling” success, as well. Creating a simple overall experience requires thoughtful sunsetting of these efforts, too.

Netflix made other mistakes — a long-term investment in the hypothesis that a more entertaining experience would improve retention, that unique movie-finding tools would create both customer and shareholder value, and that building our own “Netflix-Ready” TV-connected hardware device would do the same. You can trace these failures back to many of the themes above, too.

2010: The end of Netflix Friends. What barnacle should you scrape?

Skillful product leadership depends on rare judgment about product, people, and business. Consumer science — the process of forming and evaluating hypotheses through existing data, qualitative research, surveys, and A/B tests — requires finely-tuned intuition. And in the end, building successful consumer tech products is wicked hard.

You are likely engaged in a project today where your humanity clouds a disciplined evaluation of the project’s merit. Take a moment to reevaluate, discounting both executive-level support and conventional wisdom. Remove your “blinders” by setting an objective goal, taking stock in your pride of ownership, and asking, “What should we invest in today, irrespective of past investment decisions?” Last, if it’s clear you are working on a failed experiment, dare to communicate this and then “scrape the barnacle.” With intuition fine-tuned by recent failure, you will dramatically improve your odds for future success.

Thanks for reading, and if you enjoyed this essay, please “clap” as it makes it easier for others to find my writing.

As always, I appreciate your feedback. Please click here to complete the Net Promoter Score survey. (It only takes one minute.)

Again, thanks!


Gibson Biddle

P.S. Here are my other product management essays:

Many thanks to John McMahon, Michael Hart, Michael Rubin, Tom Willerer, Ashita Achuthan, Florian Fischetti, Dan Olsen, Nir Eyal, and Ha Nguyen for feedback and editing help.