The Rituals Behind Great Decisions: Eigenquestions, Dory/Pulse, and the Forum That Replaces Standing Meetings

The speaker opens by describing an unusual hobby: collecting the rituals of great teams and turning them into a book called Rituals of Great Teams. The inspiration came from investor Bing Gordon, who asked a boardroom repeatedly, “What are your golden rituals?” and then supplied three tests for a golden ritual: it is named, every employee knows it by their first Friday, and it is templated. Gordon's examples included Amazon's six-page narratives, Google's OKRs, and Salesforce's V2MOM; his claim was that great companies have a small list of such rituals.

A conversation with the founder of HubSpot reframed rituals as a second product. Companies build for the end customer, but they also build for employees, and what people often call “culture” becomes concrete when they describe the rituals they use. The speaker argues that designing rituals is therefore designing a two-way mirror of culture, not just preserving an operating system that happened by accident.

The practical ask is to spend as much time designing the employee-facing product as the customer-facing one. Most people spend their days on the right-hand side of the picture—products for customers—while the rituals that run the company are left implicit. The speaker wants teams to treat those rituals as designed artifacts that shape how culture actually operates.

Before moving into decision-making, the speaker gives two caveats. The first is that the talk is stuffed with examples from Coda, Coda customers, and other companies, and those examples are meant to be inspirational rather than prescriptions; if one resonates, use it exactly as is, but the main point is to design your own. The second is that Coda is a good tool for running rituals, yet the rituals can be implemented in whatever toolkit a team already uses.

The emphasis is on taking control of how teams run their rituals rather than borrowing a famous company's process wholesale. The speaker frames the rest of the talk around four parts of decision-making: asking the right question, following the right decision process, using the right forum with the right stakeholders, and ensuring the people in the forum are the right ones. The examples that follow are therefore a library of patterns to adapt, not a mandate.

The first topic is asking the right question, illustrated by a 2008 YouTube dilemma known internally as the Modern Family question. YouTube was losing money, had grainy video, and faced a major lawsuit, yet staff meetings kept returning to the fact that “Modern Family” was a top-five search query every week even though YouTube did not have the show. ABC had put every episode of the top TV show online for free on abc.com, which was unusual at the time, and the team split between linking users out to abc.com and refusing to do so for fear that great content would never be uploaded to YouTube.

Every meeting repeated the same unresolved debate, so an offsite was called and the speaker was asked to frame the discussion with data. Participants were already locked into their positions, but the night before, the speaker read research from Google's shopping team explaining why users chose Amazon over Google Shopping: users knew Google indexed Amazon and the rest of the internet, yet they preferred Amazon because the experience was consistent even if it was not comprehensive. That led to a different opening question for the offsite: in ten years, would the online video market value consistency or comprehensiveness?

The group settled the consistency-versus-comprehensiveness question in about an hour, and looking back sixteen years later the speaker believes they were right: the online video market exploded and no single site became comprehensive. Answering that question also resolved the immediate Modern Family question and several others, including decisions not to take third-party players on YouTube and to take back control of the iOS app from Apple. The speaker calls this an eigenquestion—a made-up term borrowed from linear algebra's eigenvector—meaning the question that, if answered first, would answer most of the other questions.

The second topic is forming the right decision process, starting with a conversation with Colin Bryar, former chief of staff to Jeff Bezos and author of Working Backwards. Two things happened at Amazon in 2004: Bezos sent the famous memo banning PowerPoint, which Bryar apparently drafted, and a product called Writely—a predecessor to Google Docs—arrived and enabled collaborative writing with a right-hand comment rail. The result was a familiar failure mode: the most important decisions got jammed into a 100-pixel comment panel hanging off the side of a document.

The speaker enumerates what is wrong with that pattern. People watch avatars at the top of the doc, the CEO wanders in and leaves, and nobody knows what that means; comments must actually be left; and when the review starts, comments are read in order, so a spelling mistake, a grammar note, and a proposal to cancel the project arrive in the wrong sequence. Meanwhile a Slack thread runs in the background asking whether the team is serious, and that conversation happens outside the decision itself.

Coda's answer is a two-way write-up, sometimes shortened to Dory/Pulse. The driver adds a “done reading” button at points in the document, questions go into a table where people vote them up or down (Dory, named for the fish that asks all the questions), and the central step is Pulse: everyone goes on record as supportive or not supportive, with reasons. A toggle hides everyone else's pulses until all are written, which is how the ritual removes groupthink.

Employees describe this as a golden ritual because it reflects Coda's value that great ideas can come from anywhere. The speaker reports hearing people tell friends that their topic outvoted the CEO's, that they were asked for their opinion and could write it down thoughtfully rather than performatively, and that the input was incorporated into the discussion. The ritual is also one of Coda's most copied, with many adaptations across companies.

The speaker says there are at least fourteen ways to Dory/Pulse and highlights adaptations tuned to different company cultures. At Zoom, the decision process is built around root-cause reasoning—credited to Eric—rather than eigenquestions: the template pushes teams to get to the root cause instead of fixing proximal problems. At Coinbase, the chief product officer at the time, Sergey, rejected the democratic voting element as contrary to Coinbase's culture, but liked forcing everyone to write things down without wasting time talking. Coinbase therefore kept a version of Pulse and mapped opinions to RAPID roles, making clear who is recommended, who agrees, who performs, and who is the decider; the decider's view is the one that matters.

YouTube had a different problem during “Big Rock Planning,” where every team naturally asked for resources. The ritual it created, $100 voting, asks everyone to pretend they are in charge of YouTube and allocate $100 across all the projects. The effect was that people took off their team hat and put on their company hat, which changed how the group thought about resource decisions.

At Stripe, David Singleton described “spin the wheel” as a secret to the company's 59s of reliability. Operational reviews used to go through outages from most severe to least severe; the new ritual has everyone prepare, then randomly selects whose outage will be discussed. The message is that if you really want 59s of reliability, you cannot only worry about the big problems—you have to worry about the small ones too, and Stripe applied the same idea to other rituals.

Zapier, an early distributed team, developed rituals for remote work, including “silent time”: people write what they are working on and how they are thinking about something, then read everyone else's thinking. In practice people spent all the time writing and no time reading, so Zapier added a specific writing phase and a feed of everyone else's pulses with a button to mark each item red; the meeting does not start until everyone has worked through the feed and marked it. The speaker's point is not to copy any single version but to match the decision process to what the company is trying to emphasize.

The third topic is the forum where decisions happen. Pixar's Brain Trust is a famous review ritual; the speaker notes that Creativity, Inc. describes many Pixar rituals and that the first eleven Pixar movies were all nominated for Academy Awards, which Pixar largely credits to its review process. Netflix calls its product reviews “Gladiator meetings,” held in a room that looks like a forum around gladiators, showing that there are many ways to design a decision forum.

The speaker then asks a diagnostic question: how many standing attendees are in the decision-making meeting? People often answer that the executive staff meeting has eight people, then admit it has grown to twenty, and finally reveal that the real decision meeting has sixty standing attendees. The speaker's provocative plea is to have no standing attendees in decision-making meetings, and to use a scalable design if that sounds impractical.

Coda's Catalyst ritual turns a single-threaded meeting into a multi-threaded one. The company blocks 1–2 p.m. every day, and people schedule topics into a doc by filling in the stage, the people needed, and roles that are Coda's version of DACI or RAPID; one role is called Brain Trust, inspired by Pixar. Each topic generates a unique calendar invite with only the relevant people and the topic as the title, plus Slack notifications that reinforce the role and the principle behind it. The change let Coda go from five or six topics a week to an average of fifty decisions a week, and the meeting can become a forcing function for async work: if everyone fills out Dory/Pulse in time, the driver can cancel the meeting.

The fourth topic is who to invite, and the speaker returns to the HubSpot founder's “flash tags” ritual for feedback severity. The founder noticed that feedback was often misread—sometimes it was a small thing, sometimes a big thing—so he added hashtags to his feedback and published a guide to interpreting them, using a dying-on-a-hill analogy: #FYI means he does not even see a hill, while #plea means dying on a hill is not on his bucket list but would be a good candidate if it were. Coda began adding flash tags to templates, but found that they often arrived too late, after the team was ready to ship.

The marketing team's fix was “proactive flash tags”: a list of upcoming decisions with four columns mapping to the four flash tags. Stakeholders pre-indicate whether they do not need to be involved and just want to know the decision, would like to give input, or may have strong opinions. The implications are obvious: “do not need to be involved” means do not invite me to the Catalyst, while “may have strong opinions” means invite me at the beginning so my concerns are heard before I am surprised later.

The adoption problem was solved by putting the proactive flash-tag list everywhere. A central decision tracker lives in the company hub, but the list is linked into the executive staff meeting, every team sync, and even one-on-ones, where people are asked which upcoming decisions they have not yet marked. The speaker says this changed how the company thought about who to invite to each forum, because even a great question, process, and forum fail if the right stakeholders are missing.

The speaker closes by returning to the four parts of the decision stack: ask the right question, use a decision process that reflects the culture, create a forum that accelerates progress, and find the right stakeholders scalably. Each part has rituals that can be borrowed from the examples, but the through-line is intentional design rather than imitation. The talk's broader invitation is to spend as much time designing the employee-facing product—the rituals that run the company—as the customer-facing product.

The speaker notes that everything discussed is published with links to the rituals, and invites listeners to join the Rituals Brain Trust, contribute a ritual, or borrow from others to improve the community's decision-making. The final message is that great decision-making is not a single meeting or template but a set of deliberate rituals that fit a company's culture. The examples are thus a starting point for designing a system that teams can actually use.