Operating Cadence & Cultural Norms

Founders often struggle with requiring high performance from their teams while staying empathetic and supportive.  It can be challenging to ask for high output from people, while keeping them motivated.

Yet, employee happiness typically stems from high levels of productivity, engagement, and meaning in work.  Some find working at a startup to be enjoyable, but this enjoyment should be a consequence of performing excellent work, and not an end in itself.

Twitch CEO Emmett Shear observed:

“Employee happiness is not the leading indicator. It’s a trailing indicator that you’ve done a good job in building a meaningful workplace where people are productive. I’ve never seen unhappy employees in a company where people are engaged with their work and producing at a high rate. And I’ve never seen happy employees in a company where employees are not highly engaged and productive. If you want people to be happy, focus on helping people be productive.”

In this post, we will focus on:

1. Tactics to scale company culture and operating cadence
2. How you should spend your time
3. Methods to evaluate your performance

I. Scaling Company Culture & Operating Cadence

What are the hallmarks of a great company culture?  High employee engagement & productivity.  

As founder/CEO, a key component of your job is to keep the pace and efficiency of your startup as it expands. 

Only the CEO can establish the operating rhythm of a company.  Many CEOs often discover that their employees would rather slow down to offer work-life balance, ensure employee satisfaction, and reduce potential turnover.  However, maintaining momentum is crucial as it is the sole consistent advantage that startups hold over established companies.

Rather than merely telling people to ship faster, a structured approach is essential to keeping employees inspired and efficient. Such systems differ across companies, yet they all have a common characteristic: They prioritize the urgency of resolving customer issues. This customer-centric approach is a common trait amongst successful companies, often reflected in their core values.

Examples: Tactics to move fast and maintain customer centricity

Parker Conrad has instituted this customer-centric approach. 

His objective is to amplify his team’s sense of urgency to address customer concerns promptly.  To achieve this, he has integrated Slack channels to circulate real-time customer feedback from various channels including in-product customer surveys, sales renewal discussions, customer support dialogues, and new sales interactions.  

For example, every customer has its own unique Slack channel for his teams to provide sufficient context.  Whenever feedback is received that Parker deems unclear or urgent, he interacts, tags relevant individuals, poses questions, and delves deep to find the answers.  The common understanding that Parker is actively involved prompts swift responses by his team to customer and employee questions.

The system Parker implemented offers two main advantages:

First, it enables Rippling to identify the underlying issues more efficiently than waiting for the monthly aggregation of data into dashboard for review:

“The channels contain anecdotal data that’s lost in dashboards. You see patterns or draw conclusions that are not always visible in data that's already rolled up. You lose a lot of information when you roll up data. Data is good at answering questions that you know to ask. What it doesn't tell you is the questions that you didn't know to ask. You see patterns in the ‘anec-data’ that you don't see in the actual data.”

Secondly, Parker’s methodology also helps catalyze company-wide urgency:

“When we started these channels, I was the only person in them for a long time. I looked at every piece of feedback and would relentlessly tag people.

‘What happened here, what happened here, what happened here?’ Engineers, product managers, whoever it was. As the number of people in the channel grew, I had to have a tough conversation with my leaders. I told them, ‘I need you to beat me to the punch. If someone gives a negative NPS, I want you to be in the ticket before me and your team to be in there before you.’

Once we set that up, suddenly people started getting more deeply involved and we started tagging the person responsible for that account. Over time, this became a thing that the team was doing. But I'll still jump in once in a while and go through the last 10 or 15, and I'll always find one or two to leave comments on or tag people on, so that people know that I'm still in the channel.”

This process is much more for identifying issues than it is for speed. That said, it can also improve speed if the team is ready to jump in and solve problems quickly as issues arise. For example:

• You notice that the conversion to demo rate is declining
• Upon looking at the anecdotal data (e.g., sales calls), you realize the SDRs do not understand how to handle objections
• This kicks off an urgent need to replace the leader of that team.

Rippling has codified the idea of evaluating anecdotal data into its leadership principles. Rippling leaders are expected to “go and see for themselves” when issues arise.

At Athelas, Tanay Tandon has built a differentiated process to drive urgency, setting aggressive deadlines that help improve shipping speeds:

“We have a really strong demo culture in the company. People sprint to get their thing shipped by Friday so that they can present it at the 4pm all-hands. We also have demo days which are more high production quality, every three weeks or so. Product features are built in sprints to get stuff out to show at all-hands or demo days. Both of these act as forcing functions to ship quickly.”

Taking it a step further, Tanay is obsessive in reviewing customer feedback every weekend. Tanay analyzes Gong and SDR calls himself on Saturdays to gain firsthand insight.  Watching sales call recordings offers insight into which sales reps are performing the best, which pitches worked and which didn’t, and mitigates any issues in meeting sales quotas early. 

Tanay also answers customer support questions, peruses customer emails, and formulates lists of prevailing customer concerns. On Sunday afternoons, he convenes with his co-founders and key leaders to detail a new product or feature that remedies any significant customer issues he has noted.  He may also designate engineers to implement the product/feature solution outlined, and launch the fix in 1-2 weeks. Tanay frequently opts to collaborate with the same engineers who have demonstrated the capacity to complete their regular tasks in addition to the new product responsibilities.

“This is an extra forcing function on the people that I know are really good to solve a specific problem that one of our customers is struggling with. The point is to show the entire team that you can always up the pace.

Because the team works directly with me, it also has the positive effect of showing everyone how ingrained I am in the product and how much time I spend talking to customers. It’s critical for people to know that the CEO knows every screen in the product and understands the customer deeply. But more than anything, it shows people that things can be done very quickly, if you do a good job spec’ing them, do a good job scoping them in weekly chunks, and talk actively with customers.

It also gives me and the managers the confidence to call bullshit when people are like, 'We can't do this in a week or two.’ When we shipped a particularly ambitious project, people were like, holy shit, this was done in two weeks? They couldn't believe it and it upped the pace of everything in other pods.

"People are motivated by the opportunity to showcase their work to other employees. If you set aggressive deadlines around these showcases, you generate more productivity and excitement in the whole company. It also has the positive effect of turning engineers who demo a lot into mini-celebrities.

Creating a platform for your most productive people to become famous within the company is very impactful.”

Amazon maintains high velocity operating cadence even at its current scale by organizing its engineering and product teams into decentralized small teams. How does this work?

• Small, decentralized teams drive product development for Amazon. Amazon employees are organized into small self-contained teams focused on a single product or problem. These teams typically have 6-8 people and are called “two-pizza teams” because you can feed them with just 2 pizzas. The teams have all disciplines they need to execute (e.g., product, engineering, QA, etc).‍
• Single-threaded owners run each team. These people have full responsibility for decisions and for outcomes. It is the responsibility of the single-threaded owner to get support from other centralized functions that are not broken down into small teams. Functions like Sales, Finance, and Marketing are often centralized and resources are shared across small teams. The single-threaded owner is responsible for writing performance reviews for all employees on their team
  • Leaders don’t have to be engineers, but leaders need to be curious and willing to learn
  • Non-technical leaders usually are supported by an engineering manager/leader

• Amazon optimizes for speed over efficiency: Teams are expected to work as quickly as possible to reach their goal. They are allowed to build or borrow anything they need to get the job done. If they end up rebuilding something or duplicating the work of another team, the assumption is there is a good reason for doing this.
  • While Amazon optimizes for speed, they have built review systems to enable consistency. Within Amazon, there is a respected group of IC engineers called “principal engineers.” Outside of their normal work, they also organize for “PE reviews” to provide feedback to teams outside of their own 
  • When a small “two-pizza team” wants to make an important architectural decision, they are expected to talk to a principal engineer. Amazon prefers loose coordination over multi-step approval processes that can drag timelines  ‍
• Amazon builds using a service-driven architecture: Amazon breaks every external and internal product into a service that can be built and supported by a small team. When they want to build a new product, they “spin up a new service.” Teams are expected to maintain anything that they build, and internal services are held to the same standards as external services.‍
• Teams always work backwards from the customer: For new launches, teams must write a press release and a user FAQ before they start writing code. This ensures that users are at the forefront of every decision the team makes.
• We include a link to the book “Working Backwards” if you want to learn more about Amazon

Twilio, like Amazon, believes the key to maintaining speed is to recreate early startup conditions through customer-centric small teams. Employees are more motivated to do great work when they have a sense of autonomy, mastery, and purpose versus working purely for compensation. How does Twilio do this?

• Think in terms of ownership, and not tasks. Design the small teams with the purpose of owning a particular problem or component of the product. Twilio defines four attributes for each team:‍
  • Customer: Who is the customer your serving (e.g., external developers, internal tools, etc)‍
  • Mission: What is the mission of the team? What is the problem the team is solving?‍
  • Metrics: What is the key success metric? (I.e., How will you know if the team is successful)‍
  • Code Base: Code base ownership (ideally, the team owns the problem and the code)
• Team’s are self-contained groups of 7-10 people and have a single-threaded owner (like Amazon)
  • Each team includes everyone they need to solve the customer problem. In the case of Twilio, this mean engineers, designers, product managers, and a single-threaded owner

• Each team has a single-threaded owner responsible for the success of the mission. This person could be a product leader, an engineering leader, or even a business leader in a different function. Successful single-threaded owners have the following characteristics:
  • Entrepreneurial drive
  • Nuanced understanding of the customer
  • The ability to see the big picture and how the dots connect to your business plan

• Having a single-threaded owner creates opportunities for local decision making, and it also helps you identify a future bench of leaders for the company

• Owners can report to other owners. At the time, Twilio’s R&D team was made up of 800-900 small teams with 900 total owners. This enables you to get a lot more done as you empower a bench of leaders in the company.

• This doesn’t work for all teams. Some teams are better off being functionally organized. Sales is a good example as you don’t want a small sales team for every EPD small team. Sales usually stays largely functional with product-specific sales teams that are created over time. At Twilio, there is a generalist sales team with sub-teams for specific products

Below is a specific example of how this worked at Twilio.
Twilio’s first product was voice. The team quickly scaled to 20 people. At that point, they divided it into two self-contained teams of 10 people. One team was focused on building a set of APIs for customers to use and another team was focused on building connectivity to global carriers.

This enabled two clean code bases, two customers, two missions, and two sets of metrics. The API team built new APIs for customers, and the connectivity team connected to every carrier in the world so users could make calls in a low-cost, high-quality way.

When RunwayML was about 30-40 people, they organized into Ensembles. 

• Ensembles are small, fully functioning teams with everything they need to reach a goal
• For example: a designer, a full-stack engineer, two researchers, and a video editor
• There is no fixed structure. Main idea is they have everything they need to work
• Each team is given a mandate (i.e., their mission) that they must achieve as a team
  • For example: Make Model B perform X by Y date
• These teams are flexible. If they are not performing, they can reassemble and move to other pods
• New ensembles are created regularly (e.g., every one or two months, sometimes faster)
• This was done to enable better communication and coordination as the team grew over time

Recently, they adjusted the approach because they have more major initiatives across the company. Now teams are organized into small groups of 6-7 people. Each team has a defined mini roadmap and broader scope and structure versus what they did with Ensembles. Outside of this change, the principles are the same. In other words, teams are self-contained and small, which enables fast and coordinated work.

II. Where should you spend your time?

Your three jobs are to (1) build a great team, (2) set the company’s direction, and (3) scale the culture. We outline how other founders spend their time to maximize performance and output across all three jobs below.

Identifying bottlenecks and making sure the company is set up to deal with them.

‍You must ensure that each team knows where they’re going, they don’t have resource constraints, and that they have the right team in place to execute. Only the CEO/the founders have a full picture of the business, so only you can do this. Your job is to make sure the company is set up to appropriately deal with bottlenecks, so the company can keep moving quickly.

Examples: Brex's approach to bottlenecks

• Brex begins every planning cycle identifying the bottleneck by asking the question: "What is limiting our growth?"
• By definition, the bottleneck is the single point in the system where if you increase its throughput, the output of the system increases as well
• Therefore, Brex views "bottlenecks" as the single highest leverage point where a company can put resources
• Brex attempts to "overwhelm" their bottlenecks with resources to unlock growth in the entire system

The way we think about resource allocation at Brex is we start with a very simple question, which is what is limiting our rate of growth right now? That's how every planning cycle starts. And usually there's a bottleneck in the system, right? The definition of a bottleneck is this single point in the system, where if you increase the throughput, it increases the output of the entire system itself. Right? That's the definition. So by definition, the single highest leverage point where you can put resources in is alleviating your bottleneck, whatever it is, and we have this unusual habit of overwhelming the bottleneck with resources.

What we find over time is that when you have this sort of conviction to put resources into the very one thing that matters [the Bottleneck], you get to these much different outcomes and much faster.

‍

• This clarity needs to come from the founder. Pedro aligns the team around this concept and focuses leaders on bottlenecks in his regular leadership meetings.
• Most recently, the leadership team has been focused on deepening their understanding of customer segments. Historically, they've sold Brex the same way to three different customer segments. Going forward, they want to take a much more segmented approach to selling to drive growth.

The bottleneck mentality frees up a lot of anxiety. There's one thing that matters, and it's a little bit essentialist...The bottleneck for Brex right now is segmentation. We can't grow faster because we have a one size fits all approach to how we sell and market our products. That's the bottleneck. We're spending all of our time just talking about segmentation. That's my entire leadership team meeting for the past maybe one or two quarters. Now. It's a very complicated problem, but I think it puts a lot of emphasis on the few things that matter.

Building the right team.

Hiring is one of your highest leverage activities. This extends beyond executives into anyone you hire to solve a problem. If you hire the right person for a role, this frees you up to focus elsewhere while also advancing the company.

Nat Friedman at GitHub divides everything into low-scale and high-scale tasks. Low-scale means thereʼs an existing system or person to solve it, so it shouldn't take up too much of his time. High-scale means there isnʼt a structured solution, so he should spend his time and energy helping the organization adapt. 

For example, at one point his high-scale task was to figure out how to get MSFT’s field sales team to sell GitHub. GitHub's existing infrastructure wasnʼt set up to solve that, so he put in place the people and systems needed to do so after identifying a solution. Spend your time focusing on high-scale tasks where you don't have leaders.

"Hiring is undoubtedly the most high-leverage thing I do. The single most important thing I did one year is hire our COO. After that, writing documents and inspiring other leaders." — Emmett Shear, Twitch

"Recruiting is the highest-leverage. I spend lots of time recruiting. I hired a CFO after 168 interviews, 168 individuals that I actually talked to over 24 months. It was brutal but I was very happy with the outcome. The CEO generally has three jobs: Set the vision, make sure the company doesn't run out of money, hire the best team possible.” — Max Levchin, Affirm

Communicating with your company.

As a CEO, you’ll need to drive alignment and set the direction and strategy for the company. One of the best strategies for driving alignment is repetition. You will need to repeat the founding story and come up with memorable phrases so the company remembers them. Try to embed these key phrases in everyone’s day-to-day work. If you mention a principle or idea in many different ways, it won’t stick.

Examples: Tactics for communicating at scale

While Peter was at Segment, he spent 80–90% of his time on hiring and company communication

“The highest-value thing that I do is narrative construction: What are we going to do this quarter or year? How can I communicate that in a way that everyone at the company can rally behind and remember? What is the narrative for the market that we're taking over? What is the narrative for why we care about diversity and inclusion? Everything is narrative construction: how we think about X, such that people can remember it.” — Peter Reinhardt, Segment

Segment's 100 week Plan

• Peter believes narrative construction is key to communicating strategy. The right narrative drives urgency and alignment at scale

Peter on the importance of narrative building

‍"Strategy is the layer of the stack of mission value strategy where you have the most flexibility as CEO. If you try to constrain the strategy to any of the strategy frameworks that are out there It's transparently bullshit and not really the right way of approaching it. This is where your skill at synthesis as a CEO is tested the most. Can you create a narrative from scratch of what needs to happen and why that needs to happen? At the highest level that distills things down."

‍

• When Segment reached $70M ARR, Peter developed a "100 week" narrative to transition the company to the enterprise. Salesforce and Adobe wanted to launch competing products and were telling customers that everyone needed a CDP (Segment's product). They had distribution, but no product. Segment had a product, but needed to build enterprise distribution. Peter picked 100 weeks as an arbitrary timeline to build distribution before Adobe and Salesforce could build a new and competing product. He communicated this to the company.

Peter on the 100 week narrative

‍“The articulation was that we have a hundred weeks until Salesforce and Adobe have a product in the market. They already have the GTM team, and so we have a hundred weeks until they get a product on the market. And if we don't have a footprint in all the enterprise companies before they have a product in market, we're going to get locked out of all their accounts. So we have a hundred weeks to get a footprint in all the enterprises.”

‍

• This narrative aligned the company and created urgency as it clarified that Segment had 100 weeks to win the most attractive segment of the market.
• They broke down this high-level narrative into 1) building a fast growing commercial business (e.g., tier just below enterprise) and 2) an enterprise business

Peter on why the 100 week plan worked

‍“It was a rallying call, it had FOMO and all the things that you would hope for in a strategy. We broke it down into needing a highly efficient and fast growing commercial business and then secondarily an enterprise growth story.”

‍

• The 100-week plan was then broken into a financial plan with quarterly targets and input metrics that were owned by leaders and teams
• They would track progress against the plan through: 1) weekly executive team KPI reviews and 2) all hands (that happened every couple of weeks)
• All hands are an important communication forum as its your opportunity to highlight progress against the strategy and reiterate the initial narrative you outlined

Peter on breaking down strategy into metrics

‍“We broke down the strategy further into metrics. You can write these however you want, but for us this was OKRs. The important part of OKR is just that it keeps breaking things down into smaller and smaller and smaller pieces until you have things that people can actually go do this week. from an operating perspective, we had these weekly E-team KPI reviews, and all hands every couple of weeks. We [outlined] the structure for what was going to happen at each all hands at the company. Like when we were going to review the last quarter, when we were going to do the board recap, when we were going to talk about the next quarter. And we would show people this is how all these pieces of company infrastructure fit together into an execution model.”

‍

‍Example Q1 KPIs that came out of the financial plan

1. ARR and Gross Retention
2. Productive Capacity [capacity of sales team]
3. Pipeline Coverage [pipeline over target ARR]
4. % Healthy Accounts [based on customer indicators]
5. Enterprise customers “promoters” [enterprise customers ready to advocate for Segment]
6. A Segment specific metric that captured Gross Margin
7. Rule of 40

Kevin Systrom believes communicating well is the highest leverage activity for any CEO

"Communication is the highest-leverage thing you can do: both deciding what you're going to communicate, and also communicating it over and over again. When you're small, you can say, 'You know what, team? I've decided we're going after market X,' and everyone's like, 'Alright, great.' When you get big, you can't do that. You need to make sure you're pointed in the right direction because you can't change directions every week. Spending the time to get it right with your senior team is super high leverage. Then you have to figure out how to communicate it clearly. 'Why' is the most important word in your vocabulary. You can't just say, 'We're going after X.' You have to explain why. Then everyone can say, 'I know what we're doing, and I believe in it.' You can tell people to move left, but if they don't believe in left, you're not going to get results." — Kevin Systrom, Instagram

Auditing function/team health.

As we mentioned in the last pre-read, it becomes harder to have a complete picture of every function. As the CEO, you’ll need to build mechanisms to pulse check and dive deep into all of the functions at regular intervals. We outline tactics and strategies in the wiki document on Leading Leaders.

Staying close to customers.

As you scale, it is important that everyone at the company stays close to the customer. As the CEO, you set the tone. Spending time with customers or using your product keeps you close to customer needs. It also shows everyone in the company that it's important.

Examples: Tactics to stay close to customers

Every week, Nat Friedman has a two-hour video call with open-source maintainers, and goes to five enterprise customer meetings.

Emmett Shear talked to customers in batches. He shared what the company is up to, asked what theyʼd like Twitch to do, and asked what products theyʼre using in other facets of their lives.

Peter Reinhardt spent 10–20% of time every week in customer calls or meetings.

Tracy Young spent 25% of her time with customers.

Tony Xu still regularly completes deliveries to stay close to customers (e.g., Dashers, Restaurants, Users)

Aadit Palicha and KV Vohra (co-founders) still regularly pack orders, spend time in warehouses/stores, and complete deliveries

Vojta spends 3.5 hours per day with customers each week. The discussions fall into 3 buckets:

• Calls with strategic customers in target verticals or users of specific aspects of the product they are trying to grow.
• Calls with customers who are having issues and are not happy.
• Feedback calls on new and upcoming product launches. This helps inform the roadmap and releases.

III. Assessing Your Performance

You know you’ve done a good job if (1) you’ve hired a great team, (2) teams are hitting their goals, and (3) employees are motivated and happy. We outline some tactics to assess yourself below.

‍Evaluate yourself with feedback from reports and board members.‍

The best mechanism is a formal 360. Below are examples of how CEOs have collected feedback and communicated it with internal and external stakeholders.

Examples: Tactics to assess and communicate your performance

David Rusenko shares his results with his reports to let them know what he’s working on. It also helps leaders understand why he makes changes when he implements something new.

Marco Zappacosta shares his 360 action items with the whole company.

Emmett Shear asked his reports twice a year how heʼs doing and what he could be doing better

Pedro and Henrique shared feedback the board gave them to the entire company driving accountability

Keep a list of permanent KPIs to assess yourself.

• Rank yourself against other CEOs you know on a 0–10 scale.
• List the drivers behind your ranking (e.g., youʼve hired an A+ team of executives, moving fast)
• Once a quarter, evaluate yourself on each of those drivers: “If I were to rate everyone on my team 1–5, do I score at least a 4.5 on average. If not, what am I doing about that?”
• If you are constantly too busy, something is wrong. The key is building a great team and delegating

“If youʼre always busy, you are not successful, because you havenʼt set up the right framework for the company. Your job is to design a system that can operate with minimal involvement.” — David Rusenko, Weebly

• Building a system with “minimal involvement” will take time and you’ll go through multiple iterations to get this done. Regardless, the best CEOs will always have tools in place to hold their teams accountable and to have command of the business
• Tony Xu has done this by building a system of “rigorous measurement” at DoorDash that enables him and his team to know where they stand. He reviews these by himself, in the weekly business review, and the monthly business review. This helps him figure out where to spend time and with what leaders

Get an executive coach.

• Theyʼre expensive, but if they make you incrementally better at your job, itʼs worth it
• To find one, ask founders you admire whom theyʼve worked with. We can also help