How to be a culture mismatch at Amazon

This is my new most-upvoted answer on Quora. I’ve written popular answers about Amazon on Quora before, but it was interesting to do a take on Amazon’s culture by examining its inverse.

What are some cultural faux pas when working at Amazon?

Here are some from my time there:


  • Not acting like an owner. Amazon has an ownership culture; there are no excuses and nothing that’s “not your job”.
  • Prioritizing your own results over the broader interests of the company as a whole. This is the other half of ownership: you are expected act like an owner of the company and to put on your “Amazon hat” for all decisions.
  • “Optimizing for optics.” Amazon thinks and acts long-term; Bezos invests the company’s resources with a seven-year time horizon. Sacrificing the long term to make this quarter’s numbers is seen as simply wrong.
  • Being lax with spending. “Frugality” is touted as a company value.

Customer focus:

  • Making decisions that aren’t in the best interests of the user. Amazon aims to be the world’s most customer-centric company. That means always putting customers first—even when it means cannibalizing your own sales in the short term (don’t optimize for optics).
  • Caring only about technology. The engineers you’ll find at Amazon generally aren’t interested in technology for its own sake. They care about the product, the user experience, and the business. Being uninterested in these things is seen as not acting like an owner (see point 1).

Metrics and data:

  • Changing the site without an A/B test. Not a faux pas so much as a total taboo. Every change to the site goes through split testing, or “Weblab” as it’s called internally.
  • In general, not measuring your results and making data-driven decisions. Data is used to decide what to do. Data tells you how well you did. Data is used to predict ROI before the fact and measure it afterwards. If you’re not measuring when you could be, you’re not part of the culture.
  • Not knowing key metrics off the top of your head. Leaders, at least, are expected not only to measure but to have important numbers memorized and at their fingertips in any meeting or conversation. If you own SEO and someone asks how many sessions Amazon gets through search engines per month or what the conversion rate is, “I’ll look that up and get back to you” is not really an acceptable answer.

Operational excellence:

  • Complaining about oncall. Well, everyone complains about it, but in humblebrag fashion, the way kids at CMU or MIT complain about their homework. The faux pas is to take the attitude that it’s beneath you or that you shouldn’t have to do it—to care only about writing the code but not about whether it’s actually available for customers to use. That’s not acting like an owner (see point 1).
  • Breaking the website, especially during Q4. Kind of goes without saying. A story: When I was running Landing Page Optimization and helping out with SEO, we wanted to put a feature on the detail page to help foreign customers get to the right Amazon site for their country, if they landed on the wrong one. It was December and the middle of the Q4 site freeze, when any non-critical changes are postponed for the sake of uptime during the all-important holiday shopping season. But my manager encouraged me to get the feature live immediately, so I petitioned for an exception. Jon Jenkins, at the time director of a group called Website Operational Excellence (sometimes referred to as “We Own Everything”), gave me the go-ahead, but told me: “Whatever you do, don’t break the detail page. If you break the detail page for one hour during Q4, you will lose the company more money than your feature could ever hope to make up.” (We launched the feature. We didn’t break the detail page.)

See the original answer on Quora or check out my take on what allows Amazon to be so innovative. Or just for fun, see my list of what you can’t buy on Amazon.

Startup idea: Turnkey self-publishing service for authors

Scott Berkun recently commented on his experience self-publishing his latest book:

What was awesome about the process? Control. The cover, title and content for any published book involves rounds of discussions between the author and the publisher…. If you self-publish, you, the author, are in control….

What sucked about the process? Control :) If you have control over everything, you have to take care of everything. Every single task must be done by you, or you must hire someone to do it….

What was a pain in the neck? Some of these services are not designed for consumers, so their websites suck and take time to learn. The worst offenders are Lightningsource and (the service used to buy ISBN numbers). Once learned, most processes are easy….

Scott had to hire a designer and an editor himself. He set up his own account with a print-on-demand service and bought his own IBSN number. He used Kickstarter himself to take pre-orders.

Startup idea: Create a service that does all of this on behalf of the author, putting together all the pieces to let them self-publish. Create the service that any author who has been through Scott’s experience wishes existed.

In a way it’s like a traditional publisher, only it’s completely fee-for-service. There’s a flat, up-front fee to the author, who then keeps all the profits from the book itself.

In the pre-digital world, publishing a book required a huge capital investment in the form of the printing press. Because of this, each book was essentially a joint venture between author and publisher. Both parties made an investment: the author put in months or years of his time to write the book, and the publisher put in the capital to print and market it. Then, both parties received a share of the profits.

But now that the capital requirements have radically changed, this business relationship no longer makes sense. Many authors can now afford the entire cost of publishing. They have no need for an investment from the publisher, and so they have no need to give up an ownership stake in the outcome. As with startups, this means the author can keep not only the profits, but also control.

However, as Scott’s story illustrates, there’s still a need for all the other services that a publisher provides, such as design, editing, administrative tasks, and overall project management. That’s what this would provide.

Phase 2 of this idea: Create ways for each author to raise exactly the amount of money he needs, if he can’t fund the project out of his own savings. Maybe that’s only a few thousand dollars for publishing and marketing costs. Maybe it’s a few tens of thousands to spend 6 months finishing the book. Lessons from Kickstarter, Y Combinator, and AngelList could apply here.

To speak in one’s own voice

From a post titled The 10,000 Hour Rule by Steven Pressfield:

To speak in one’s own voice means to let go of all the other voices in our heads. Whose voices? The voices of what is expected of us. Yes, that means the voices of our parents, teachers, mentors. But it means something more elusive too. It means our own expectations of what we should be doing or ought to be thinking—what is “normal” or “right” or “the way it ought to be.” …

In terms of the aspiring writer, we sit down and try to write the way we think writers write. If we’re painting, we paint like painters paint—or dance like dancers dance. What this means of course is that we’re writing like somebody else writes and painting like somebody else paints and dancing like somebody else dances. …

How does the actor get past his own excruciating self-consciousness? How does the entrepreneur come up with an idea that’s really new? The answer is they both beat their heads against the wall over and over and over until finally, from pure exhaustion, they can’t “try” any more and they just “do.”

“All customers change from something”

Sean Ellis, my favorite blogger who never blogs, has written a rare post, “The Cult of ‘Great Product'”. A key excerpt on differentiation:

All customers change from something. Generally they either switch from a competitive solution or from just tolerating a problem without a solution. New products should decide on one of these markets. Trying to serve both markets generally leads to failure.

If you decide to target “greenfield” people (those without a current solution), then your product roadmap should be focused on simple, effective execution of their desired task. Simplicity is usually much more important for greenfield users than being feature rich. …

If you are targeting people who will be switching from another solution, then usually features are an important part of people’s decision to try it. … If you can differentiate on one of the key gripes of the competitive solution, there is a good chance you can be successful. … You’ll need to both message this differentiation and also deliver on the promise.

Go read the whole thing.

User-generated content? Bah!

An article has been making the rounds recently titled, “The Internet? Bah!” Written by Clifford Stoll (of Cuckoo’s Egg fame) in 1995, it explains how the Internet will never take over the world:

Visionaries see a future of telecommuting workers, interactive libraries and multimedia classrooms. They speak of electronic town meetings and virtual communities. Commerce and business will shift from offices and malls to networks and modems. And the freedom of digital networks will make government more democratic.

Baloney. Do our computer pundits lack all common sense? The truth in no online database will replace your daily newspaper, no CD-ROM can take the place of a competent teacher and no computer network will change the way government works.

Except, of course, that the Internet did take over the world, and virtually everything the article pooh-poohed has come to pass.

I have some sympathy for Stoll, though. His blindness to the revolution to come and the possibilities therein reminds me of my own attitude, circa 2001, toward Amazon allowing user reviews on its site.

I thought it was a terrible idea. Most people don’t know how write, so you’ll mostly get terrible reviews. You need a professional for that sort of thing. Or so I reasoned.

Boy, was I wrong. Some key things I missed:

  • Voting can be used to sift the best content to the top.
  • There is value in even a simple user review like “it broke after a week”. Or even simpler, a review that adds extra factual data about the item, such as a table of contents.
  • More importantly: Even if, say, only 10% of the population can write a really good review, most of those 10% are not professional reviewers. In fact, the person best qualified to write a review of a given product is almost certainly not a professional reviewer.
  • Related, a sincere review from a layman with domain knowledge and passion is far superior to a trite one from a professional writer who tried a product for just long enough to review it. This goes triple for obscure/niche products.
  • The long tail means that you can’t afford professional reviews for every product in the catalog, and that you need user-generated content to get coverage (even a simple review being much better than nothing, especially online, where you don’t have the product right in front of you).

Fortunately for Amazon and its customers, Jeff Bezos had a much more benevolent view of the average user and the quality of content they would create. And fortunately for the world, no one gave me veto power over the massive trend toward user-generated content.

What allows Amazon to be so innovative?

This question was asked on Quora. My answer has become one of my most upvoted answers ever. Here it is:

Some key pieces:

  • Visionary leadership: Jeff Bezos built an online retailer before most people knew what the Internet was. He was talking about what he called the “machine-to-machine Web” before the term “web services” was even coined.
  • Customer focus: Amazon is driven by serving customers and creating value for them, not by beating competitors or by winning any piece of some pre-existing market. This helps keep them focused on a consistent strategy and prevents them from wasting resources on me-too or copycat products. Bezos also says that focusing on customers helps keep you from being constrained by your skills.
  • Long-term approach: Amazon explicitly takes a long-term view of its business. The now-famous 1997 letter to shareholders included a section titled “It’s All About the Long Term” in which Bezos wrote: “We will continue to make investment decisions in light of long-term market leadership considerations rather than short-term profitability considerations or short-term Wall Street reactions…. When forced to choose between optimizing the appearance of our GAAP accounting and maximizing the present value of future cash flows, we’ll take the cash flows.” Elsewhere, Bezos has said that big bets take about seven years to come to fruition. He is willing to be persistent for that long.
  • Entrepreneurial culture: Amazon’s culture attracts and retains entrepreneurial types who want to own a piece of the business and do something innovative. For instance, rather than a functional organization in which engineers report to engineering managers and PMs report to PM managers, Amazon organizes into cross-functional teams that each own a piece of the business. Teams whose performance can be measured by a single key business metric, or “fitness function”, are given a high degree of autonomy as long as their metrics keep improving.
  • “Willing to be misunderstood”: In the Q&A at a recent shareholder meeting, Bezos said: “A big piece of the story we tell ourselves about who we are, is that we are willing to invent. We are willing to think long-term. We start with the customer and work backwards. And, very importantly, we are willing to be misunderstood for long periods of time.”
  • “Still Day One”: Amazon never rests on its laurels, nor do they fall prey to the age-old idea that everything has been invented already. Bezos always has his eye on the future and sees an ever-expanding path ahead. Whenever asked where Amazon is, or where the industry is, or how far any of it has left to go, he replies: “It’s still day one.”

Original answer on Quora

Steve Jobs

Some men do great things in the prime of their lives, then retire to enjoy the fruits of their labors, and perhaps to sit on some boards or give some talks. Steve Jobs’s last decade was the most productive of his life—hell, of almost any man’s life. He burned everything there was to burn within him and went out in a blaze of glory.

When he stepped down as CEO of Apple, some said it was because he was close to death. Perhaps instead, his life had to end because he could no longer work.

“Rest in peace” seems inappropriate for such a man, like an insult or a curse. If there were a Heaven, it would be for Jobs a continuation of his life on earth, thinking about the future and working to make it real.

Creative lessons from Sondheim

Last Christmas I got a book of lyrics by Broadway composer and lyricist Stephen Sondheim, titled Finishing the Hat. The book is only for die-hard Sondheim fans, but the preface contains this gem:

There are only three principles necessary for a lyric writer, all of them familiar truisms. They were not immediately apparent to me when I started writing, but have come into focus via Oscar Hammerstein’s tutoring, Strunk and White’s huge little book The Elements of Style and my own sixty-some years of practicing the craft. I have not always been skilled or diligent enough to follow them as faithfully as I’d like, but they underlie everything I’ve ever written. In no particular order, and to be written in stone:

Content Dictates Form

Less Is More

God Is in the Details

all in the service of


without which nothing else matters.

If a lyric writer observes this mantra rigorously, he can turn out a respectable lyric. If he also has a feeling for music and rhythm, a sense of theater and something to say, he can turn out an interesting one. If in addition he has qualities such as humor, style, imagination and the numerous other gifts every writer could use, he might even turn out a good one, and with an understanding composer and a stimulating book writer, the sky’s the limit.

Good advice for writing lyrics that (mutatis mutandis) goes equally well for writing code or designing a product.

Remembering Danny Lewin, first victim of Sept. 11

The breaking news of Osama bin Laden’s death is bringing back fresh memories of the terrorist attacks of Sept. 11 for many people. It’s a time to remember again some of the great lives lost that day. For my part, I remember Danny Lewin.

Danny was co-founder and CTO of Akamai Technologies. I met him on a few occasions during an internship I did at Akamai in the summer of 2000. He was smart and confident, and I looked up to him. I still remember how he would frown and say, “Unclear to me”, when he wasn’t yet sure what he thought of an issue. That made an impression on me, and still stands for me as an illustration of intellectual honesty.

Danny was tall and imposing, built like a linebacker. He was also Israeli, and like many Israelis, he had served in the Israel Defense Forces—in fact, he was in Sayeret Matkal, an elite special-forces unit specializing in counter-terrorism. So when I heard that he had been on one of the planes that was hijacked and flown into the World Trade Center, I had to believe that he had fought them. Danny—strong, confident, a man of action—wouldn’t have taken a direct attack sitting down, and wouldn’t have been afraid to confront the danger head-on. Re-reading his bio this evening, I found my confirmation. Based on the testimony of a stewardess, who made a phone call from the plane in flight, it is reported that he was stabbed by hijacker Satam al-Suqami as he tried to foil the hijacking—making him the first victim of September 11.

I moved to NYC in June 2001, so I had been there less than three months when the attacks hit. I was only 21, and the events turned my world upside down. I was stunned the entire day, in shock, unable to process everything that was happening. But that evening, when I read the news of Danny’s death, everything hit me all at once. For the first time that day, I cried.

Danny, I shed another tear for you today. We remember you and we miss you.

Startups don’t hedge their bets

Recently I was giving advice to someone who is looking to move to the Bay Area and break into the startup world. He’s looking for a job at a startup here, and his plan is that if he doesn’t find one soon, he’s just going to quit his job and move here anyway while he continues the search.

He asked me whether that would look bad on his resume, to have been out of work for a few months. I told him no, and that in fact, as a hiring manager, I would admire his decision. He has the guts to take a risk by quitting his job in order to take a bold step toward his goals. That’s the right mentality for a startup.

Startups don’t hedge their bets. The very nature of a startup is to take a big risk for a huge reward. The challenge is so great, and the resources so small, that everything you have must be concentrated on a single objective. Startups have one bullet in the chamber, one chance to hit the target. You can’t afford to hold anything in reserve for a backup plan. (And besides, compared to an established, profitable company, you have nothing to lose.)

Part of the founder mentality is being willing, indeed eager, to take on this kind of risk.