The Courage to Take Risks


How an organization deals with ambiguity and views risk directly drives its ability to compete. Some leaders see a growing competitor and conclude it’s riskier to respond than to double down on its shrinking market. Other team members see this same market environment as a burning platform.

Why is risk-taking so tricky?

The common wisdom is that companies need help taking business risks. In reality, companies make big, unconventional bets—when requested or demanded by big customers or big markets. Clay Christiansen’s observed this in the Innovator’s Dilemma:

it is a company’s customers who effectively control what it can and cannot do. As we have seen in the disk drive industry, companies were willing to bet enormous amounts on technologically risky projects when it was clear that their customers needed the resulting products.

Christensen, Clayton. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (Management of Innovation and Change). Harvard Business Review Press.1

Even risk-averse companies are willing to take on certain business risks.2 However, companies and teams do consistently struggle with a type of risk. And that risk is social precisely who has the right to decide where there is no social consensus on the path forward.

Sidebar—Group vs. Individual

Quick. Who is more likely to arrive at the correct answer fastest—the group or the individual?

Before you answer, know that there is a logical answer to this. You can derive the answer to this question from logic.

How I like to frame it, what’s more likely?—that two or more people arrive at the same answer together, or that one person arrives at the right solution?

So, let’s assume there is an 80% chance any individual has the correct answer. The odds that one person will arrive at the right answer is 80%, and the odds that two people will agree on the correct solution is 80% x 80% or 64%.

As the group size increase, the advantage that an individual has the answer over the group only grows.

If you reflect on your experiences, this likely holds. When a group comes together, everyone has their idea about the problem, situation, and solutions. In an effective group, what happens next is information exchange. Group members share their reasoning and perspective. Next, someone has an epiphany, and they form a new perspective from the information and more information exchange—they share the idea.

Often someone else will recognize that this idea is a workable solution to the problem—and now takes on the role of convincing or explaining the concept back and why it is workable—next, one by one, each member of the group sees or tests the workability of the solution until the group arrives at a consensus.

In these examples, the individual team member arrives at the correct answer before the rest of the group. Even still, the group plays an important role—providing information that improves, validates, or identifies exceptions to the solution and helps improve how you communicate your answer, so your team gets on board.

But there is an inherent problem with this—while the individual will arrive at the right answer faster than the group—the question is which individual? In strategy contexts—people with the wrong answer often have even more conviction that they have the right answer—than those with the right answer.3

So consensus building in many organizations becomes a conflict resolution process when there are multiple solutions, each with its advocates, and no independently agreed upon approach for validating which answer is correct.

While this might help resolve conflict, it’s also a proven way to arrive at the wrong answer in a novel situation.

Instead, I recognize organizations focus on finding “who has the right answer,” not “who should have the right answer” or whose job is it to have the correct answer—it’s learning to recognize who has the correct answer and that person may exist inside or outside of an organization. Uncovering who has the right answer is job one for most problems.

Social risk

Again the risk isn’t business risk—it’s social, specifically the right to decide and move forward when people disagree on the solution.

The  A16Z podcast: Innovating in Bets with Thinking in Bets author Annie Duke highlights this social risk as the difference between consensus and non-consensus decision making.

Here is a summary way to think about it:

  • If your decision result has consensus and the outcome proves the decision to be right, you are a leader.
  • If the decision result has consensus and the outcome is wrong, you will be forgiven—because everyone thought it was the right answer.
  • If the decision result does not have a consensus and you can see it through to fruition and proves to be right—you’re viewed as L’enfant Terrible (a prickly genius).
  • If the decision result does not have a consensus and you can see it through, and it proves wrong—you’ll be viewed as a fool.

Why consensus results in lower quality decision making

Decision Quality

Let’s assume you have a strong leadership team for a product. Ninety-five percent of the time, the answer your leadership thinks is the right answer is the correct answer. So, in short, there is only a five percent chance the leader is wrong.

Let’s see what the odds are that the decision that everyone aligns on is the right decision:

  • Two decision makers: 90% chance
  • Three: 85% chance
  • Five: 77% chance
  • Ten: 60% chance
  • Fifteen: 46% chance

You can map it out: 95%^(# of decision makers).

So with each decision-maker, we’ve drastically reduced the probability of making the right decision. And this is just one level of decision-making in practice:

There are strategic decisions// few

There are tactical decisions// many

There are execution decisions // most

We make decisions every day at every level, and the quality goes down for the number of decision-makers.


You can apply the same logic to the likelihood that everyone can meet—for a decision today—the odds go down as you add decision makers.

Instead, you delay the decision until everyone gets in a room—which reduces your decision velocity and speed in finding out if a decision is right or wrong.

And finally, for each decision maker, the number of communication paths increases nonlinearly (n^2-n)/2.

The Mission

We’ve defined our mission—our mission is to get better at managing the social risks associated with non-consensus decision-making (which I’ll refer to as decision-making without consensus).

Amazon manages this with a culture of “disagree and commit,” which covers only a small portion of the story.

Examining the risks of making decisions without consensus

I believe in the power of starting with empathy and examining non-consensus decision-making not from the decision-makers perspective but from the perspective of the team and leadership.

From that framing, there are two core dimensions:

  • Was the decision maker informed or uniformed
  • Was the decision process unifying or isolating

Here is where we have an attribution error. I believe the reason why consensus decisions require less risk is primarily because of the decision process. By attempting to manage towards consensus, the decision maker becomes more informed by surfacing concerns across the team. Because each decision-maker has a voice, the impact is often unifying.

In short, a lot of what we internalize as decision-making risk is the rightful pushback against decision-makers making uninformed decisions through a process that leaves everyone feeling isolated.

Courage to Take Risks: An Action Plan

As product managers, we disproportionately impact the value we create for our business and our customers. And at times, the right decision will be the decision that does not have a consensus, and we’ll need to take accountability to move forward.

Here’s how we’ll begin to institutionalize that muscle:

  1. Directly Responsible Individual. A decision has a directly responsible individual, and she or he is the most informed party, regardless of title or seniority, who is accountable for that decision.
  2. Will make informed decisions. The Directly Responsible Individual (DRI) takes it upon herself to seek out additional perspectives, and team supports by providing open access to information.
  3. At a high velocity. Code takes time; it takes; design takes time it takes—the most significant input we have as an organization to increase the value we deliver to our customers is by increasing our decision throughput. As an organization, you’re building the capacity to make higher-quality decisions with less information than competitors.
  4. Through a unifying process. The first step in unifying is clarity—who is the directly responsible information we can all rally around, and the second is communication. We need to publish our decisions transparently and the decisions that went into them, and finally, we need to have more time for coffee. I’ve found nothing more unifying for a team than a 1:1 opportunity to hear the thinking explored further.

Expectations of DRIs

Avoid task mitosis. A collaborative organization tends to divide a task into equal parts to parallelize work or ensure everyone contributes to the success. While winning together is critical. Most information worker work is constructivist—to make a good decision in Y, you must have X information first. So parallelization without being intentional about whether things need to be sequential results in lower-quality decisions.

Minimize the number of decisions made at any one point in time. We’re all fallible, and our products should inherit our beliefs, not our biases. A way of minimizing the influence of our bias as product leaders is to reduce the number of decisions you directly make. Ask who is the most informed party (which is not the same as who is supposed to be the most informed party or who is the most informed job title)—if there is someone more informed than you on the decision—consider delegating them the Directly Responsible Individual responsibility for that decision. Smart delegation diversifies the perspective embedded into the product. The second element is time—don’t lock yourself tomorrow into the decision a dumber version of you made. Take active steps to minimize the number of decisions required today—because the answer reveals itself with time.

Reduce the cost of decisions. To thrive in the long run, we must win big and lose small. The way to do this is twofold: (1) only increasing the amount of exposure you have to a decision directly related to the quality and clarity of information. In short—we need to get good at validating ideas with the least cost possible. (2) we will prioritize organizational agility by ensuring that most decisions are reversible with new information and that we have the organizational deft to change tacts as we acquire richer information.

We need to know our signing authority. Organizations have the context of signing authority, which is the dollar amount based on your position in an organization; you are empowered to make a contractual obligation on behalf of the company. Similarly, individuals have their signing authority in an organization—Gibson Biddle highlights that Netflix’s Head of Product has $100M in signing authority. Your signing authority will likely be a lot less—but you need to have a visceral knowledge of what your signing authority is—and for any decisions approaching the limits of your signing authority—you need to inform your manager before you lock in your decision.4

Expect to be audited Expect to be asked who you talked to, what informed your perspective, and how you’re dealing with risks. I wouldn’t expect a tome—but be willing and ready to discuss and discuss how you arrived at a decision and what you believe the risks to be. The best processes leave an audit trail as a natural outcome.

Frame the decision. Be able to explain the context surrounding the decision, whether the decision is easily reversible, how you’ve reduced the cost of the decision (by running an experiment), and what types of decisions will be impacted by this decision.

Make a decision. When you’re DRI, it’s wise to present feedback options and gather more input in your thinking. But options are not a decision—make a decision based on the factors you know. The team will audit your decision to ensure it’s factoring in the most pertinent insights.

Expectations of the team

Share information. Share information openly and without strings attached.

Provide your articulated perspective. We have a lot of opinions, but we’ll progress through clearly articulated viewpoints supported by an insight—“I don’t like that color” isn’t a perspective. “I’m concerned that the call to action will suffer from banner blindness” is. If you have an opinion without insight—take the time to research and develop your supporting understanding—then share your perspective.

Ask questions—respectively and within reason If you’d like to know the thinking behind a decision, ask. But ensure you don’t use questioning to require the Directly Responsible Individual to “justify” their decision until you’re, please.

Disagree and commit When the DRI has made a decision—shift 100% of your focus towards making that decision successful.

Expectations of leaders

Build one product Our products must be cohesive. They should not see the seams of different teams making different decisions. So the product leader must ensure product integrity (the sum of different product decisions feel and accrue to one product vision and set of product values) in that each decision does not result in a pied product. The product leader is the DRI for product integrity, and the UX leader is the DRI for experience integrity.

Delegate decisions. The team is closer to the problem than you are—push every decision that the team can make to the team. But ask honestly if a problem area has the right person or experience to serve as the DRI. If the answer is no, become more involved but develop a plan for how you can untangle yourself.

Audit decisions. Audit decisions through soft inquiry, usually in standing meetings. Ask questions to ensure the immaterial decisions are truly immaterial, that material decisions are reversible, and that the DRI is making an informed decision.

Trust the DRI. When you delegate a decision, and the team is gathering the necessary approach, bias towards the team—even if their approach differs from your own—they are closer to the problem.

Intervene when it’s not working When a DRI is not developing an informed perspective and has a habit of isolating and not unifying. Become directly involved first, discuss with the team and individuals to ensure they are meeting the expectations of team members—and finally, delegate a different DRI and work on a development plan with the individual to handle the situation better in the future.

Living it

Embracing social risk is an organizational challenge with organizational benefit. To appropriately respond to new situations in a novel context, you need to take an approach that is not consensus and be right.5 It’s the not consensus that most trips organizations up, and as a result, they don’t get to reap the learnings of being wrong or the benefits of being right.

There are steps you can do as an individual to help normalize the non-consensus. Focus on being informed and the unifying process. People often skip ensuring their decision-making process is unifying. The excuse is that you don’t have time. As an individual—unless you have sponsorship from an executive team to operate differently—the fastest you can move is ten to fifteen percent faster than the organization. Moving faster might feel right in time. Still, it will not be a unifying process, and you will find yourself lacking the organizational support needed to execute your solution.

Most importantly, if you want people to support your non-consensus decision, you, too, have to support non-consensus decisions. Supporting other teams needs to learn through doing (provided the risk isn’t material to the organization).

  1. Christensen, Clayton. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (Management of Innovation and Change). Harvard Business Review Press. Kindle Edition.
  2. To go deeper, when we’re taking a chance—we’re responding to uncertainty—with action. More in a future post.
  3. This is due to the difference between valid and reliable answers. Often people use the reliable answer, which has been successful before–as the answer for today’s problems, where the valid answer often doesn’t have any data to suggest it’s the right answer. For example, if you flip a coin and get heads twice, a reliable answer is to bet on heads—it has documented evidence it works, while the valid answer is tails.
  4. Gibson Biddle Case Study.
  5. Sam Hinkie Resignation

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a website or blog at

%d bloggers like this: