Skip to content
Share
Explore
Script
Below, you’ll find a script for my talk. The words may not match exactly, but it’s pretty close. And if you like the template, feel free to copy it and make it your own:
Copy template
Category
Section
Script
Category
Section
Script
28 mins 21 secs
28 mins 21 secs
000
2.8Section
Category
Search
Section
Script
Slide
Slide Link
⏰
Introduction
8
Hi everyone - my name is Shishir Mehrotra and I’m CEO / Co-founder of Coda. Before starting Coda, I spent a number of years at Google, responsible for running the tech side of YouTube, and before that I worked on a number of products at Microsoft, and started a company called Centrata.
And today, I’m very excited to talk with you all about Rituals for great decision making.
24 secs
Before talking about decision making, let’s take a small detour and talk about the first word: Rituals.
6 secs
Rituals is a topic I'm somewhat obsessed with. People have lots of normal hobbies, I have normal hobbies too, but I have this weird hobby, I like to collect rituals from different companies.
I'm turning them into a book called Rituals of Great Teams. And my whole inspiration started with a conversation with this guy Bing Gordon. So, Bing, if you don't know him, was the Chief Creative Officer at Electronic Arts. He's now a famous investor, Amazon, Zynga, a whole bunch of other great companies.
He and I sat on a board together, and he kept harassing the CEO with this question. He said, what are your golden rituals? And none of us really understood what he meant. And at some point we said, okay Bing, what the hell is a golden ritual?
And he said, well, great companies have a small list of golden rituals, and he had this really crisp definition.
He said they have three characteristics. They are named, every employee knows them by their first Friday, and they are templated.
And he rattled off his list. Amazon had six pagers, Salesforce had V2MOM. I just thought it was a really sticky way to think about the way the companies function. So over the last four years, I've taken this idea and made it my hobby, collecting up rituals, putting it into this book.
1 min 20 secs
By the way, I’m writing the book in the open — so if you want to follow along, you can sign up for the braintrust to get sneak peeks of each chapter and give me feedback. ritualsofgreatteams.com
13 secs
So, along the way I got to meet some really interesting people.
One of the most interesting conversations was with Dharmesh Shah, co-founder of Hubspot. Dharmesh told me about a ritual called flash tags, which I’ll talk about a little later.
But he also made an observation about rituals that I thought was really interesting. He said, as everybody here builds products, and for the most part, we build products for our customers. But as companies, we actually build two products. One for our customers, one for our employees. And if you ask people to describe the product for your employees, they'll call it culture. And if you ask them a little bit more about what that means, they'll start describing a set of rituals. And so if you think back to that, you know, 6-pagers, V2MOM, sometimes people refer to that as the operating system of a company, but he had this provocative way of thinking about it, that actually it's a two way mirror of company culture.
1 min
So my main message for you today is this: You all spend a lot of effort designing your products. But after this talk, I hope you will put as much thought into designing your rituals as you do in designing your products.
15 secs
Ok, so that gives you the context for today’s talk, which is specifically about the rituals for great decision making.
I’ve taken some time to collect together a whole bunch of interesting rituals for how interesting companies make decisions. And systematized it a little bit for you in a way that is interesting and so that you can borrow a great ritual from it as well and apply it to your own businesses.
We're going to talk about in four parts: how to ask the right question, follow the right decision process, in the right forum, with the right stakeholders.
35 secs
And two quicks caveats:
37 secs
4 mins 31 secs
1 Decision Question
10
I'm going to tell you a little bit of a story about eigenquestions. And we'll talk through the lens of a experience we had at YouTube.
10 secs
So before starting Coda, I worked at YouTube for a number of years. I joined in 2008.
In that time period, YouTube was far from a success. In fact, most people thought YouTube was going to be Google's first major failure. And we had a lot of challenges, we were losing hundreds of millions of dollars a year, we had crazy videos, we had a big lawsuit, with Viacom. But if you sat in a staff meeting on any given day, the thing you would probably hear us talk about is a thing we call the Modern Family question.
One of our top 5 search queries each week was for Modern Family, which happened to be the #1 show on TV, so people naturally came looking for it. There was one big problem though: we didn’t have Modern Family on YT.
Interestingly, ABC had decided to put every episode of Modern Family online for free. That may seem obvious these days but it was groundbreaking then - no network had done that.
1 min 1 sec
So the Modern Family question was: should we link out to ABC.com?
The team divided on this question. Half felt that we should “do what the user wants” and “link out” to the site. But the other half felt that if we linked out, then we would lose our chance to host that type of content on YT.
After many rounds of discussion, we held an offsite to try to resolve this conflict and I was put in charge of framing the choice. The night before, as I was pulling it together, I read a writeup from a different team - the Google Shopping team.
They were in an all-out battle with Amazon, but were finding that they were losing. Google Shopping assumed they would win by being more comprehensive - why go to Amazon, when you can come to Google and find results for everything on Amazon and the rest of the web? But users kept telling them they valued consistency - Amazon’s results and pricing were clear, they understood the reviews and return policies, how shipping worked, etc.
1 min 4 secs
So I decided to reframe our Modern Family YT question. Instead of asking “should we link out to other sites or not”, I started with a question of “in 5-10 years, will the online video market value consistency or comprehensiveness?”
This frame led to a much more interesting discussion and we decided that the video market would value consistency over comprehensiveness. As a sidenote: I think we were right - 15 years later, the online video market is huge, but there’s no one site that has all the video.
This led to an obvious answer to Modern Family - we would not link out. But it also led to answers to many other questions. E.g. we decided to stop supporting 3rd party embeds. Most interestingly, this led to us taking back control of our mobile app from Apple - which is probably a separate fun story.
And this whole process became labeled the Eigenquestion process.
54 secs
The name Eigenquestion is completely made up. But for those of you that want the source, it comes from a math concept of eigenvectors. The math majors in the room will remember that an eigenvector is the “most discriminating vector in a multi-dimensional space”. But you don’t really need to know the math.
19 secs
A better definition is this: the Eigenquestion is the question where, if answered, it likely answers the subsequent questions as well.
I’m sure you can recognize this situation: you’re in a discussion where there are 10 questions. How are they ranked? Maybe by the order they came in? Or the severity? The reason this was the Modern Family question was we were trying to do a deal with ABC.
But in your list of 10 questions, there’s probably a question further down, perhaps by answering #6, you’ll answer all the questions as well.
33 secs
And the technique is simple. Instead of spending most of your time looking for the right answer…
6 secs
4 mins 19 secs
2 Decision Process
27
Ok, so that’s the first part of this framework. Let’s talk about the decision process.
How do you actually make decisions?
7 secs
Colin Brayer is the former Chief of Staff at Amazon. He wrote a book called Working Backwards, going through Amazon’s rituals.
8 secs
I had a chance to spend a bunch of time with Colin, and one of the stories he told me a bunch about was this very famous memo that Bezos sent out in 2004, called No More PowerPoint.
Actually a fun side note, Colin said he actually wrote the memo, Jeff sent it out, and nobody thought it was going to be a big deal. But it turned out not only did this memo change how Amazon operates, it changed how basically every other company operates.
So in 2004, total happenstance of history, actually two things happened that year.
Jeff sent out this memo.
And a new product came out called Writely—the original version of Google Docs.
These two forces changed how meetings ran basically everywhere. Obviously changed Amazon, but it changed most other companies as well. My guess is every company here has either adopted the Amazon way of run meetings, or anti-adopted this way. But your team probably has a strong opinion on this!
58 secs
With Writely, we could all co edit and co write with each other, but it also had one feature that everybody kind of underestimated: commenting in the right-hand panel. And all of a sudden our meetings looked like this.
14 secs
While researching for the book, we had a dinner where one of the participants asked:
It felt like the most important decisions in our company are jammed into the right hundred pixels of a Google Doc. Is that really the best we can do?
15 secs
The first one is, and I'm sure everybody here has had this experience, you put a lot of effort into a writeup, you send it out, and then you watch the avatars at the top.
13 secs
WTF does that mean?
Then you're ready to have the discussion, and how are we going to order the discussion? Well, we might as well go in order.
10 secs
So you start with the first comment, and it says, hey, you made a spelling mistake in the title. Okay, that's not so useful.
9 secs
Next one says, oh, this project should be cancelled. The next one says, oh, you should try this other idea. Next one says, another spelling mistake.
As you work through, you're clearly not going in the right order.
13 secs
In the meantime, everybody is sitting on Slack, messaging each other with what they really think. Is this for real? Are they really going to try to ship this thing?
Oh my god, this is the best idea ever.
That's all happening somewhere else.
15 secs
So, is there a better way?
We spent a bunch of time on this problem and ended up selecting a set of rituals that have become the Coda Golden Rituals. So if you took Bing's test, and ask any Coda employee on their first Friday what our Golden Ritual is...
18 secs
…they would almost certainly tell you about a thing we call two-way writeups. Sometimes referred to internally as Dory and Pulse.
So here's how it works. You'll send out a write up.
11 secs
Somewhere in that document is this done reading button—sometimes it's even in multiple places. The driver of the meeting will make a point of, “you may not have to comment anything, just tell me if you're with me so far.”
14 secs
There’s also a Dory, named for the fish who asks all the questions. And it's just a way to rank your discussion topics. Everybody adds their list of topics, then votes them up and down.
13 secs
But the most important part of this ritual is in the middle, so we call this the Pulse, the two-way part of this discussion.
You don't just get to come in and mark everything up. You have to go on the record on what you think: I'm supportive or not supportive of this decision. And we reduce bias with a toggle that hides everyone else’s Pulse until everybody's done.
24 secs
Here you can see all of those rituals put together—the done reading button to make sure everyone has the right context, the Pulse to see how everyone feels about the writeup, and the Dory to capture questions and prioritize the discussion.
So if you ask a Coda employee on the first Friday about Coda's Golden Ritual, they would almost certainly tell you about this. They'll say it because it's reflective of our culture. Something like, “yeah I joined this company Coda, and one of the values is great ideas can come from anywhere.”
And they'll talk about how in the last week “I was in this meeting, and I had this question, and I outvoted the CEO.” I don't know why outvoting the CEO is such a big deal, but that seems important to people. Or “I got asked my opinion on this topic. and it wasn't performative. I actually had to write down my real view, and people read it and interpreted and built it into the decision.”
So this is reflective back to Dharmesh's point, it's a two way mirror of our culture.
1 min 5 secs
Not only is it our most popular internal ritual, it turns out it’s one of the most copied and modified rituals from teams that use Coda. And since every company is different, the way it’s used is often quite different too.
So I went out and looked at all the different people doing variations of Dory and Pulse.
And I wrote about 14 different ways people have adapted this particular ritual to their culture in ways that may not be recognizable from ours.
I thought I’d give you a few examples.
32 secs
At Zoom, they have a very specific way of thinking about decisions. Eric Yuan, Zoom’s CEO, has an idea he calls Root Cause Reasoning, or RCR. Instead of focusing on asking questions to get the right questions, they focus on getting to the root cause of a problem.
Eric’s view is keep working your way down a problem until you get to the root cause, so they built that into their decision making structure, starting with the design team. Matt led this work, and his team actually turned this ritual into the RCR template you see here.
35 secs
At Coinbase, Surajit picked up a similar ritual. He and I talked about Dory Pulse, and his first reaction was, “That's like so anti-Coinbase culture. We're not going to vote on decisions. It's not a democracy. There's a decision maker, and it's super important who that is.”
But he took the same idea and ended up adapting it a slightly different way. He said, “Well I actually like the idea of everybody writing down their viewpoint—that's a good idea. But I want them to do it and be super clear who the decision maker is.”
So they use RAPID there. In their case, they have something very similar to Pulse, and they end up with writing down exactly what your role is. You can work through this, you can read everybody's input, but all that really matters is the decision makers input at the end.
51 secs
At YouTube one of our toughest decisions every planning cycle was picking a set of priorities, we called them Big Rocks. The problem was that everyone just wanted to advocate for what their team was responsible for.
So we came up with a new way to handle this called $100 voting.
It’s a simple idea: If you were responsible for YouTube and you have a fictitious $100 to spread between each of the candidate Big Rocks, how do you distribute them?
The key was that it forced them to step out of their own team, and think with their company hat on.
37 secs
This one is from David Singleton at Stripe and he told me about one of their favorite rituals at Stripe called Spin the Wheel. They actually extended a similar ritual from Amazon, and he says it’s the secret to why Stripe has 5 9’s of reliability.
It started with incident reviews. At most companies there is a meeting where you get together to talk about the top ops incidents and outages from the previous week. In most cases, you go through them in order of severity. But at Stripe, they would ask everyone to prepare and then spin a wheel, and pick someone at random. The idea was to send a clear message: if you want to achieve 5 9’s — you can’t just focus on the big things, you have to focus on the small things too.
That ritual has been adapted into many other forums as well — any time you want to inject a bit of forced randomness into a discussion, start by spinning a wheel.
1 min
One last example comes from Wade Foster at Zapier. Zapier, if people don't know, is the OG distributed company—like before everybody else thought it was cool. And they’ve adapted their rituals to fit that culture.
For example, every meeting starts with silence. Everybody writes their updates, and then reads everyone else’s.
Wade told me that they were having a problem where everybody was happy to write their own updates, but nobody wanted to read anybody else's updates. And so they ended up adapting the ritual a little bit differently. They have something kind of like Pulse, where everybody writes their update, but they added a done reading button that turns into a feed. So as you go, and you click what you’ve read.
43 secs
9 mins 47 secs
3 Decision Forum
18
There's lots of great examples of famous forums. Top left there, that's a picture of Pixar's famous forum called Braintrust.
Pixar, I love Pixar the company in many ways. I think the first 11 movies all got Best Picture nominations. It's kind of an incredible track record. And if you read Creativity Inc. and how they think about Pixar, it all starts with a very famous decision making meeting called Braintrust.
Netflix does all their product reviews in a room that's shaped like that—built in a roundroom. They call it the Gladiator-style room.
Everybody has these sort of iconic ways to think about their decision making. So, I spent a lot of time thinking about all the different decision forums, and there was one interesting observation about these forums.
46 secs
If you have a hard decision to make, and you've got to bring it to the main meeting where you're going to make a decision, how many standing attendees are in that meeting?
As few as possible. Go look at the invite. I've heard as few as four, I've heard as many as sixty. And the interesting thing about this is, there's a natural incentive, right? You have a decision making meeting every week. What does everybody do? They say, “well I might need to be there.”
And because that person is going to be there, I kind of need to be there too.
36 secs
I have a provocative view for you.
You should have no standing attendees in your decision making meeting. I know that sounds crazy, but I would actually aim at zero. How?
We ended crafted a ritual to make that practical, and it's a variation of what we found other companies do. At Coda, we call it Catalyst.
20 secs
Here's how our version of this works.
First off, just think about what's wrong with the standing meeting that has standing attendees. Well, you have a standing meeting, you have all these different people there that might be needed. But, you discuss one topic at a time. Inevitably, there's a set of people that really care about that topic, and everybody else doesn't really care.
So you've got all these unnecessary attendees, which is a waste of time. But more importantly, it's also a set of people that bring disruptive disruptive. They’re not really the important part of this discussion.
35 secs
The goal is to get as many topics at once, with the right attendees. No more, no less. Try to add lift, not drag.
9 secs
So we hold from 1 to 2 p. m. every single day on every person's calendar across the whole company. You're not supposed to schedule anything over that time. Then there's a little system called Catalyst.
13 secs
Everybody comes into a doc to schedule their topic, and they fill in exactly who they need for their topic. This is our version of DACI or RAPID. You’ll notice that we blend a bunch of rituals here — we actually renamed our equivalent of Approvers to Braintrust - just like Pixar.
19 secs
… an independent calendar invite is created for each. This turned out to be really critical.
It’s the hack that makes the whole process actually work!
10 secs
Saying, hey, you've been invited to this thing, this is our culture for this, this meeting type, this is what your role in this meeting is, and so on. It turned out that little piece of creating a unique calendar event and reminder system for each one completely changed the forum.
18 secs
We went from having stacks of meetings that looked like this to a much more multi-threaded set of meetings.
I've done the analytics on it. We used to make something like five-ish decisions a week. We now make something closer to 50 decisions a week. It also had another interesting impact.
18 secs
So the way it works is this thing gets scheduled, you get a Slack invitation saying you're invited to this catalyst. When scheduling your Catalyst, you can also say that the meeting is default async…
13 secs
…which means please fill out your Dory-Pulse before, you know, a day before the meeting.
And assuming everybody does that, and the driver gets a chance to interpret everybody's feedback, you're going to get another Slack message saying…
13 secs
…Hey, great news, feedback's been received, the driver has decided to cancel the meeting.
All of a sudden this becomes the forcing function, where the meeting ends up almost being like the punishment for not making the decision.
And it ends up re changing how you think about the whole process.
18 secs
So that's our way of doing it. There's lots of others. You can pick the braintrust model, the gladiator model, the catalyst model, or something in between.
But the main thing I would ask is if does your decision forum accelerating decision making? Or does it decelerate how everything operates?
18 secs
5 mins 7 secs
4 Decision Stakeholders
7
So you have this great forum, but who do you invite?
And in particular, do you have to invite the founder to everything? And obviously, parallelizing things doesn't work. The same person has to be in the middle of everything. This turned out to be a hard topic where we found inspiration from another ritual.
20 secs
So Dharmesh, I mentioned earlier, has a ritual he calls Flash Tags. And his observation was, as the founder, whenever he gave feedback, people tended to misinterpret it.
Sometimes he meant it to be taken very literally, sometimes he meant it as a suggestion. So he decided to come up with a thing he calls Flashtags. Now, every time he gives a piece of feedback, he adds one of these four Flashtags to the end of his feedback.
And he had this analogy for how to think about it, which was, hill dying status. So if he says the Flashtag “FYI”, it means I don't even see a hill, and if he says Flashtag “plea,” it means dying on my hill. Dying on that hill is not on my bucket list, but if it were, then this would be a pretty good one.
51 secs
We started using Flashtags as a way to prioritize feedback.
Someone comes in and say “I have all of this feedback.” And then you ask people to start writing how strongly they feel about it. That’s a good step, but usually too late.
Now somebody comes in and says, “I really, really think this is a bad idea. #plea”
21 secs
One of our marketing teams had this idea to turn this into a proactive process, so we call these Proactive Flashtags. The way it works is pretty simple, there's a list of upcoming decisions, and everybody gets asked, or all the stakeholders get asked, and you need to pick where you're at on the spectrum.
I don't need to be involved, just let me know the decision, I'd like to give input, or I may have strong opinions.
It's very effective for a bunch of reasons. From my perspective as an often stakeholder, I have a view of here's what's coming up, and I can render my opinion.
From their perspective, they know exactly how to interpret that. If I say I don't need to be involved, you don't need to invite me to the Catalyst. If I say I may have strong opinions, that means not only do I have to come to the Catalyst when you're ready to launch. I need to come to the catalyst when you're ready to brainstorm, and you should include me in that part of the process.
1 min 5 secs
The obvious problem with this is, how do you get people to fill that thing in? The way we ended up doing it is just by embedding it everywhere. So we created what we call a decision tracker, and we put it first off into our big company hub. Here's all of the different decisions.
Then we linked it everywhere, in the executive staff meeting. Every team has a sync meeting, so we put it in there. We even put it in people's 1-1 docs, and said, this is the list of upcoming decisions, please say where you're going to be on this, not whether you're going to be for or against a particular thing, just how much are you going to care. And we're going to use that to get to the right set of statements.
49 secs
3 mins 43 secs
Conclusion
3
Alright, so that’s the framework—four sets of rituals for great decision making. Can you find a way to encourage people to ask the right question, not just the right answer? Is there a decision process that actually matches your culture and your unique way?
You have a decision format forum that accelerates decision making. And you have a way for people to find the right, the right group of people for every decision.
26 secs
Everything that I just talked about is available to you, in the hopes that they’ll help you find your own decision-making rituals.
Just go to my doc at this link.
It will also invite you to join my braintrust for the rituals of great teams book. This room is full of amazing thinkers and I’d love your help in making the book great as well - so please sign up!
24 secs
53 secs
28 mins 21 secs
Want to print your doc?
This is not the way.
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
CtrlP
) instead.