How We Make Decisions

Most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90 percent, in most cases, you’re probably being slow.


The practice we use for our decision-making process is largely based on how Amazon does it. That’s why we recommend reading and

How Jeff Bezos Makes Decisions - Type 1 And Type 2

What you should know after reading this:

  • The difference between type-I (a big, strategic decision that it’s hard to turn back from) and type-II (an everyday operating decision) decisions

  • It’s better to disagree and commit than get stuck in information paralysis and elevate true misalignment issues quickly. Put another way, it’s OK to question a decision up front but it’s not OK to fight it or ignore it during implementation.

  • Don’t make decisions by yourself, instead seek perspective. Unless nobody else is impacted upstream or downstream by your decision? Then just get it done.

  • Don’t wait for all the information. Most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90 percent, in most cases, you’re probably being slow.

  • Taking the time to seek perspective from others before making a decision will not slow the company down. On the contrary, if done correctly, this should increase overall execution speed significantly.

Our learnings

Information flow

Previously, the input for decisions, the reasoning, and the decisions themselves were often not clear to everyone in Awell. The challenge to create good information flow is a common one in organizations that are larger than a handful of people and this is a problem that will only scale as we grow.


When 2 people are discussing an issue, the need to be efficient is important. When a team is discussing an issue, the need to be efficient is paramount because each inefficient minute is multiplied by the number of people in the discussion.

Data vs opinions

We strive to make decisions based on facts, data and information from insights. While we are comfortable working with believable opinions if necessary, it is wise to plan for data when making decisions based on opinions.

In some cases, there may be different interpretations or perspectives on the same set of facts or data, which can lead to (healthy) debate. Therefore, while facts can provide a more solid foundation for knowledge and understanding, they may not always lead to absolute certainty or agreement.

Facts vs assumptions

Assumptions can be dangerous. When building your decision-making framework, stay cognizant (and wary) of assumptions. Sometimes, though, assumptions are unavoidable: Not all beliefs or claims can be easily reduced to facts or evidence.

Our decision-making process

Written by default

A list of all our written decisions: (internal link).

We use a written decision-making process, consisting of the following steps:

  1. The recommender (= person who comes up with the issue & solution) writes up a decision page. There is a Confluence template for this (= decision page).

  2. The recommender invites the below people to the page to get their input (async). Once it is received, the document is ready to be reviewed by the decision maker (D).

    1. As: people in the team that need to agree with the decision

    2. Ps: people in the team that need to perform the decision

    3. Is: people in the team that need to provide input on the decision

  3. The recommender schedules a decision meeting and invites the D, As, Is, and Ps. The decision can also be made in a weekly meeting (no need for a separate meeting if the issue is not urgent).

  4. At the decision meeting, D makes the decision. The next actions are logged with a DRI and a due date. If the D cannot decide based on the input in the document, one extra round of written comments and input is requested and the decision is deferred to the next meeting.

This strategy, though time-consuming for the Recommender, yields the most thoughtful decisions in a very short amount of time. The bonus is that well-documented decisions are available for everyone to see.

Move fast and inform for reversible decisions

Many decisions are reversible, two-way doors. Those decisions can use a light-weight process.

Not every decision needs a decision page. For everyday, low-impact decisions, or decisions where nobody else is impacted upstream or downstream, we have the following guidelines:

  • Can be executed with incomplete information.

  • Act fast, test the outcome, and improve.

  • Aim to get it right but consider other factors such as time, effort, and money.

  • Be prepared for the decision to go wrong.

  • Do not aim for perfection.

  • Take the decision, informs others, and answer questions