What's your favorite Amazon leadership principle?

The importance of being right a lot and ideas to improve your judgment

A couple of weeks ago, I left Amazon after eight years. I’ve spent the last couple of weeks unwinding, sleeping in, going for hikes, and reflecting on my time. This is my first time in the past 14 years of working that I’ve taken more than two weeks off. And it’s amazingly refreshing without the worry of checking in on email, Slack, or worrying about work piling up. As for this blog, over the next few months, I’ll be alternating from re-posting content I published internally to Amazon to this new blog and posting fresh content reflecting on my time and lessons at Amazon like today’s post. 

On my second to last day at Amazon, someone asked: “You’ve been at Amazon for so long. Which leadership principle is your favorite that you’ll take with you?” I love this question because Amazon’s leadership principles (LPs) have played such a central part in my life over the past eight years. I’m taking them all with me and will continue to embody them in my professional and personal life. 

On your first day at Amazon, you’re taught “everyone at Amazon is a leader” and then shared the LPs. The LPs codify the values of how Amazonians are expected to operate, how Amazon evaluates performance, how individuals are assessed when they interview, and provide a common language for the culture. They’re used to express appreciation like “This is a great example of invent and simplify” or to encourage someone to grow like “I’d love to see you stretch yourself to think bigger.” They’re used to challenge the status quo like, “Are we really being customer-obsessed with this decision?”

So, which leadership principle is my favorite? Are right, a lot.

Are right, a lot

Here’s how Amazon describes are right, a lot:

“Leaders are right a lot. They have strong judgment and good instincts. They seek diverse perspectives and work to disconfirm their beliefs.”

“Leaders are right a lot.”

This LP plays a central role in how leaders operate and make decisions. Are right, a lot is about an individual’s judgment reflected in decisions they make. To demonstrate the other LPs, you have to use strong judgment to decide when and how to act. The LPs often have tension between them, like earning trust while having a backbone – how can you disagree while maintaining trust with someone? Or moving fast (bias for action) while insisting on the highest standards? Your judgment is how you choose between tradeoffs.

One reason this LP is my favorite is I love the phrase “a lot.” The LP could’ve just been “are right” or “are always right,” but instead, the “a lot” gives permission to be wrong. It gives freedom to make mistakes, and it acknowledges that leaders won’t always have the correct answers all the time. “A lot” implies we should have good judgment most of the time but that making decisions quickly or with partial data means that we’ll make wrong assumptions and go off course.

We run a lot of A/B tests at Amazon. Thousands of experiments are happening at any given time across the Amazon ecosystem in which we are testing variations of the website, new features, and optimizing the customer experience. My teams ran a lot of A/B tests and not all of them won. It’s essential to take risks to learn and see what works. Whether an A/B test wins or loses is not the point because in either case, the team learns from it and their judgment improves. Over time, teams can build upon the learnings of these prior experiments. Yet, I would also warn about over-optimizing for wins. As a team matures with insight about the customer, it’s vital to take bolder risks in experimentation to continue learning. Don’t just optimize to be right all the time but also take risks you’ll learn from.

“They have strong judgment and good instincts.” 

Judgment is a soft skill that is infrequently discussed yet critical to how leaders operate. Like any soft skill, judgment is improved with time, practice, reflection, focused learning, and mentorship. We use our judgment to make individual decisions – how we operate, spend our time, choice of words in an email, etc. – and output decisions – product requirements, design details, how a team operates, etc. Judgment can show up in micro-decisions we make every day, like choosing to read an email like this or attending a meeting.

If we think of judgment as a skill, we can then look at ways to measure, gain input, and improve over time. By permitting yourself to be wrong, you can take more risks and evaluate the outcome to determine how you will improve next time.

“They seek diverse perspectives and work to disconfirm their beliefs.”

Let’s talk for a moment about how we make decisions. People make decisions using a combination of what data they know and their instinct, which is based on an individual’s experience. Seeking diverse perspectives gives us a broader range of data. As a designer, this can mean seeking views from a diverse range of groups (e.g., legal, privacy, accessibility, ML scientists) and individuals with diverse backgrounds — those with different disabilities, races, cultural and religious values, body types, gender/sexual identities, socioeconomic statuses, or education levels.

The phrase “disconfirm their beliefs” is vital. To demonstrate strong judgment, you cannot simply form a point of view and treat it as the only way of thinking. In Amazonian-speak, don’t over-index on having a backbone on your beliefs. I got this wrong earlier in my career. I thought to be right a lot was to state my point of view and stand by it. Doing so lost a lot of trust with those I worked with and looking back, I was far less right than I thought I was. Over time, I learned how to frame my beliefs as beliefs, pick my battles, and use others to disconfirm my beliefs. The more inputs we get, the more we can validate or invalidate our assumptions and improve our judgment.

Making decisions on partial data

I mentioned earlier that we often make decisions with partial data. If you wait until you have all the data to make a decision, you are likely moving too slowly, not taking enough risks, and losing out on delivering value sooner. Amazon taught me the importance of making calculated decisions quickly and with partial data.

In 2016, I remember having long debates with a product manager about using “and” or ampersand (&) in our UI text string. We were looking at adding a small line of text to describe a new “Settings” link in the Amazon App menu to explain what a customer could find there. He argued that an ampersand was shorter and mine was that our UI Text best practices at Amazon were to spell out “and” when there was room, which there was room in our case. We went back and forth for well over a week on this. We ended up running an A/B test using “and” to measure the impact of adding the UI text string. We were cautious about introducing any new complexity to the menu if it didn’t demonstrate value. Well, the A/B test lost. We wasted time debating this and I don’t have any conclusive answer on whether “and” or ampersand is better. Thinking back on this, I held onto my belief too strong and didn’t look to get a broader range of perspectives to inform this decision. This was a low-cost, reversible decision (what we call a two-way door decision).

Here are some ideas on making decisions with partial data:

1. Avoid over-thinking easily reversible decisions: If a decision is inexpensive and can easily be changed, don’t over-debate or over-think it. In my example earlier, we didn’t have data on “and” vs ampersand, and the decision was ultimately reasonably inconsequential. If you find yourself debating these types of decisions, step back from the debate and discuss your team’s process on how decisions are made and who makes the final decisions. Consider bringing in another individual to help make the final call.

2. Understand how you’ll evaluate the decision: When we take a risk to make decisions with incomplete data, we should consider how we’ll know if we made the right decision. When you have a feedback loop, you can improve your judgment in the future and move forward with stronger confidence in your decision.

3. Rely on other’s judgment when you lack experience: I’ve learned to avoid over-stating my opinion as something I think we must do and instead rely on other’s judgments over my own at times. One signal I pay attention to is whether I’m the only individual on my team who has a particular belief. This doesn’t necessarily mean I’m wrong, but I listen to this signal to see that my belief is the outlier and consider how I’ve arrived at a different belief than everyone else on my team.

Be willing to change your mind

Here’s a secret to being right a lot: you cannot be right a lot if you aren’t willing to change your mind when presented with new evidence. I heard Jeff Bezos shared this at an Amazon all-hands around 2014 and also with Jason Fried, co-founder of Basecamp, that people who are right a lot of the time often changed their minds. Jason writes, “He’s observed that the smartest people are constantly revising their understanding, reconsidering a problem they thought they’d already solved. They’re open to new points of view, new information, new ideas, contradictions, and challenges to their way of thinking. This doesn’t mean you shouldn’t have a well-formed point of view, but it means you should consider your point of view as temporary.” 

If you want to be wrong a lot, don’t change your point of view. You cannot be right a lot without also being willing to change your mind.

How to improve your judgment

I mentioned earlier that judgment is a skill we can improve with time and practice. Here are some ideas you can use to improve your judgment starting today:

1. Avoid overstating your opinion: Others evaluate our judgment by the strength of our stance in the discussion. When we become insistent on our opinion without clear rationale and data, we risk losing trust. Match the strength of your stance to your confidence in your opinion. When in doubt, start by under-stating your idea (“I think X could be true”) and seek to gather more input to disconfirm your beliefs.

2. Study topics outside your comfort zone: We strengthen our perspective as we gather a broader range of inputs. As a designer, I also study politics, science, history, economics, finance, etc. Similarly, seek mentorship from those outside your discipline.

3. Be the last to speak: As the first person to speak in a meeting (or reply to an email thread), you initiate the discussion on your opinion. However, you also lack data that you may gather from hearing others contribute to the conversation first. When I’m in a larger meeting and don’t have all of the necessary contexts to quickly weigh-in, I start by listening intently first. Then, I contribute after others have gone.

4. Ask for more time to provide input: In meeting-heavy cultures, it can be easy to feel pressured to make decisions in the moment. Someone asks you for your input and you may feel social pressure to give a response right away. I think it’s always appropriate to say, “I’d like a bit more time to consider this. Can I follow up in a few hours?” Buy yourself more time to gather more input.

5. Take risks and learn: I’ve found it helpful to be self-aware when sharing input when I only have partial data. I consider these calculated risks to see how others react and use these as opportunities to learn. However, if you provide your input without considering how much you know or don’t know, you aren’t taking a calculated risk that you can learn from the outcome. Start by trying to be self-aware of these moments so you can learn afterward.

6. Document your judgment and reflect: In her book Thinking in Bets, Annie Duke describes how poker players don’t evaluate their plays by the outcome (did they win or lose the hand) but by the quality of their decision-making for the play itself. What inputs did they consider? What details did they miss? Often, outcomes are beyond our control. Instead, reflect on the inputs you used to make your decision. Spending a little bit of time every day or week to reflect can help you look at opportunities to improve and become right a lot of the time.

---

That’s all I have for today. Thanks for taking the time to read. I’d love to hear from you: What’s your favorite leadership principle and why? You can comment publicly or reply directly to this email. If you found this email helpful, I’m deeply grateful if you share this post with others.

I don’t yet have a way for you to vote or provide input on future posts. That is coming soon. If you have any topics you’d like to read in the meantime, drop me an email.