# Working Hypotheses > Lightly illustrated essays by Rachel Pike about building startups and working with founders. > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. ## About - [About Rachel Pike](https://www.workinghypotheses.com/about.md): Operator and writer. Fifteen years working with startups; currently at Vercel, previously COO at Modern Treasury, earlier work in health tech, Angellist, VC, and flying laboratories. ## Essays - [Systems Create Integrity](https://www.workinghypotheses.com/essays/systems-create-integrity.md): More than a decade ago, I sat in an open plan office one row behind two women in their mid-twenties. - [How to Run a Good Startup Advisor Program](https://www.workinghypotheses.com/essays/how-to-run-a-good-startup-advisor-program.md): We ran an advisor program at Modern Treasury that I think every startup should copy. - [A COO's Unnamed Work Type #4: Janitor](https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-4-janitor.md): The last category of unnamed work was the most fun to name. - [A COO's Unnamed Work Type #3: Founder Vault](https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-3-founder-vault.md): Of the unnamed work that is emotional or interpersonal, there’s two types: the part that doesn't face founders, and the part that does. - [A COO's Unnamed Work Type #2: Onramps and Offramps](https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-2-onramps-and-offramps.md): Take three truths. - [A COO's Unnamed Work Type #1: TV Producer](https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-1-tv-producer.md): After about ten COO interviews, I started thinking about my last job differently. - [The Shape of Work is Changing](https://www.workinghypotheses.com/essays/the-shape-of-work-is-changing.md): If you draw a chart of effort applied in different phases of making things, the shape of the curve on that chart is changing. - [Anti-process leads to Sprezzatura](https://www.workinghypotheses.com/essays/anti-process-leads-to-sprezzatura.md): One common thing that happens in startups is founders renounce the whole idea of planning and process. - [The Three Types of COOs](https://www.workinghypotheses.com/essays/the-three-types-of-coos.md): About a year ago, I ran into a founder CEO friend of mine. - [Revenue is a Feeling](https://www.workinghypotheses.com/essays/revenue-is-a-feeling.md): One of the best conversations I had was with a COO who works for a solo founder. - [Have Wonder](https://www.workinghypotheses.com/essays/have-wonder.md): I sometimes get asked what to write about, and one hack I’ve used to explain it is: have wonder. - [Poka-Yoke & the Planning Fallacy](https://www.workinghypotheses.com/essays/poka-yoke-the-planning-fallacy.md): Most startups hate planning. Poka-yoke and how to do it right. - [Off-the-record COO Project](https://www.workinghypotheses.com/essays/off-the-record-coo-project.md): When I left Modern Treasury, I had years of questions built up. ## Resources - [Lean In Circle Opener](https://www.workinghypotheses.com/resources/lean-in-circle-opener.md) - [Questions from Michael Dearing](https://www.workinghypotheses.com/resources/michael-dearing-questions.md) - [Questions from Same as Ever](https://www.workinghypotheses.com/resources/morgan-housel-questions.md) - [Excerpt from The World of Yesterday](https://www.workinghypotheses.com/resources/world-of-yesterday-excerpt.md) ## Skills - [working-hypotheses](https://www.workinghypotheses.com/skill.md): How to discover, read, and correctly cite this site - [Skills index](https://www.workinghypotheses.com/.well-known/agent-skills/index.json): Machine-readable agent skills catalog ## Optional - [Subscribe on Substack](https://workinghypotheses.substack.com/subscribe): Free email subscription - [HTML site home](https://www.workinghypotheses.com): Human-readable site - [Essays index](https://www.workinghypotheses.com/essays.md): All essays - [Resources index](https://www.workinghypotheses.com/resources.md): All resources --- --- title: About Rachel Pike description: "Operator and writer. Fifteen years working with startups; currently at Vercel, previously COO at Modern Treasury, earlier work in health tech, Angellist, VC, and flying laboratories." url: "https://www.workinghypotheses.com/about" author: Rachel Pike type: about --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # About Rachel Pike I've spent the last fifteen years working with startups. My winding career journey started with flying laboratories, where I made sketches in lab notebooks. Then I spent time in VC, healthcare, building Modern Treasury as COO. I took a six month break to interview and to write these essays. Now I work at Vercel, running company strategy and operating. One way to describe my career meandering is that I "have a habit of leaving the golden ticket" (as used on the [Grit podcast](https://www.thetwentyminutevc.com/podcast/grit/)). ## What's written here & who is it for? I publish essays every few weeks. Half of the essays are generated from my off-the-record interview project. The other half I started writing over the last ten years. They're thoughts jotted down in the margins of day-to-day operating. Between the companies I've had the luck to be part of, and now the dozens of others I've glimpsed into from my interview project, I started tying things together into these essays. These essays are my current Working Hypotheses about how to build companies. They are written for readers interested in the inner workings of startups or for those who work closely with founders in some way. ## Illustrations I'd say I "art" at an approximately first grade level. But I am a highly visual thinker and often take to an oversized piece of white paper to work things out. Some of these images are redrawn from my notebooks, all of which I've kept. Some of them are hand drawn versions of things AI suggested to me, and some are ideas that come while writing. And sometimes I just let AI do the drawing. ## Subscription Subscription is managed on Substack. You can read posts there but the layout isn't as good. Every post is free. thanks for reading! Rachel --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/about --- --- title: Systems Create Integrity description: "More than a decade ago, I sat in an open plan office one row behind two women in their mid-twenties." date: 2026-06-17 url: "https://www.workinghypotheses.com/essays/systems-create-integrity" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Systems Create Integrity More than a decade ago, I sat in an open plan office one row behind two women in their mid-twenties. One of the women went to an Ivy League school. When she joined, she was confident and negotiated well for her salary. The other woman went to a state school; she had joined earlier and hadn’t negotiated as hard. Both women were at the manager level in the marketing department. At that time, the company didn’t have a comp system. We had two managers on the same team, both talented, and one was paid $50,000 more than the other. As they worked together over the years, they talked. And they found out how differently they were compensated. Not long after the truth got out, they both left. ![departing](https://substackcdn.com/image/fetch/$s_!ZjLM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc967f923-dfec-4a7f-97e3-486ff9364bea_767x547.png) “Act with Integrity” was one of the five values at that company. It was a fine value, even if the wording was flat. “Act as if your actions will always be known” would have been more concrete. (1 Once over dinner at home, we were talking about a tough situation my dad was in at work. My mom said, “the truth always gets out, so act as if it will always be known.” So she gets credit.) But the writing wasn’t the main problem. The company wasn’t living up to it. In a startup, systems, not individuals, scale integrity. Compensation bands don’t force you to pay the exact same amount to people, but they do create a container of integrity, because there’s a maximum differential in top to bottom pay. The range flexes (for experience, performance, specifics of the role, etc) but it doesn’t break, which means that when people talk, the gap is explainable. No matter where they come from or what role they play, people should be paid fairly. Fairly against the market, and fairly against each other. ![ladder](https://substackcdn.com/image/fetch/$s_!Ps0b!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd8bd91bf-3436-40df-be56-2d771ac9e675_395x949.png) One way to think about a system is that you do work upfront so your employees don’t have to. The system makes it easier and faster for your team to operate. And that’s how it works: integrity, if built into the path of least resistance, no longer depends on any individual’s behavior. When we had about 10 employees at Modern Treasury, I told the founders the story of these two women. I said we needed to start thinking about comp bands. They looked at me sideways. I told them I didn’t believe in making comp transparent, but I wanted to operate as if comp was transparent. So we made a simple, consistent comp system. Building on that, when we had 65 employees, we purchased Radford. The Radford sales guy told us that we were too early-stage to buy. “It’s too expensive for startups.” But I wanted the data because I wanted to create a system. So we bought. Systems create fairness through structure. It doesn’t have to be comp bands. (2 The AI market is changing comp ranges and bands. OpenAI and Anthropic moved to a single-title model for all their technical teams, “Member of the Technical Staff.” But that didn’t change the need for a system. It just evolved into a different kind of system: less variation in base salary for levels, and a new bonus system to reward performance.) Some other examples of systems in a startup (3 Outside of startups, one of my favorite examples is seating charts. Good seating charts, even for a 12-person dinner, take effort to get right. But they create a more enjoyable and equitable experience for guests.) that create integrity: - Onboarding programs - [Access to advisors](https://www.workinghypotheses.com/essays/how-to-run-a-good-startup-advisor-program) - Consistent interview questions per role - Transparent decision logs - On-call rotation schedules Each of these uses a system to programmatically equalize access, experience, or knowledge across employees. Individually, integrity is acting as if everything will be known. In a startup, it’s designing a fair system. ![thanks for reading](https://substackcdn.com/image/fetch/$s_!yn3Z!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b043b06-ea08-4b6b-8d3c-d322f853ac20_930x295.png) 1. Once over dinner at home, we were talking about a tough situation my dad was in at work. My mom said, “the truth always gets out, so act as if it will always be known.” So she gets credit. [↩︎](#fnref-rss-1) 2. The AI market is changing comp ranges and bands. OpenAI and Anthropic moved to a single-title model for all their technical teams, “Member of the Technical Staff.” But that didn’t change the need for a system. It just evolved into a different kind of system: less variation in base salary for levels, and a new bonus system to reward performance. [↩︎](#fnref-rss-2) 3. Outside of startups, one of my favorite examples is seating charts. Good seating charts, even for a 12-person dinner, take effort to get right. But they create a more enjoyable and equitable experience for guests. [↩︎](#fnref-rss-3) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/systems-create-integrity --- --- title: How to Run a Good Startup Advisor Program description: We ran an advisor program at Modern Treasury that I think every startup should copy. date: 2026-04-15 url: "https://www.workinghypotheses.com/essays/how-to-run-a-good-startup-advisor-program" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # How to Run a Good Startup Advisor Program We ran an advisor program at Modern Treasury that I think every startup should copy. The way we ran the program extended the reach and broadened the impact of our advisor program compared to any I’ve seen at other startups. ## What Advisors Did For Us Advisors did many things, including: emceeing panels at conferences, coming in for whiteboard sessions, introducing us to prospects, sharing product intel about competitors or past players in the market, or helping scope phenotypes for hires. In addition to these one-off things, there were three ongoing ways advisors contributed. ### 1. Provided Specific, Undiluted Advice One thing I hadn’t thought about before this program was the different level of advice directed at founders versus non-founder executives. Founders and CEOs get tons of advice. Non-founder executives get less. And filtering advice through a game of telephone (from advisor to CEO to relevant executive) is lossy. It doesn’t lead to the same level of detailed, specific, and productive conversations. For example, we had an incredible advisor who had been a VP People and CPO. She coached our Head of People on specific projects and programs, something only that advisor could have done directly. ![side by side comparison of advice.](https://substackcdn.com/image/fetch/$s_!srha!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3e07d148-4718-4eec-8ba1-65837032a0ad_2000x1000.heic) ### 2. Showed Senior Leaders Our Commitment Bringing on top-notch advisors showed our senior leaders we were investing in them. It helped us with retention, especially at the seniormost levels, because it extended the expertise they were training under to be more than just me, or the founders, or the executive team. Practically for me, it took the pressure off. It meant I didn’t need to be the technical coach on every single topic my direct reports might need coaching on. The woman who ran marketing was superb, and I think she would say I coached her on plenty of topics. But she was also a more seasoned marketer than me, and I couldn’t out-market her on many topics. Plus there were topics that neither of us spiked on. So we sought help. That meant I could coach and mentor her on topics where I spiked, she ran with things she spiked on, and we used the expertise of advisors where we needed it. ### 3. Worked For Us That VP of marketing was originally an advisor! We hired four VPs or department heads from our advisor program that I can remember. This program really helped us find out who we wanted to hire. (1 I learned, yet again, the banal lesson that pedigree doesn’t necessarily mean excellence.) That was actually an interesting story. This advisor was an early customer success manager at a company that sells in a completely different market (SMBs) with a lower price point and a very different type of customer relationship (transactional). She ended up running all of CS there after working her way up for more than a decade. But the company wasn’t a hot tech company that you might name drop at a YC event by saying something like “oh yeah, my advisor is from ABC company.” Despite that, because she had played so many different roles as the company grew to over 10,000 people, her hands-on experience was really useful in coaching our CS leaders. Interestingly, this advisor held monthly coaching sessions with two leaders from our CS team for the entire year. She was the single exception to the no-structure structure I described above. She was so great we renewed her advisor agreement for a second year. And even now she’s in touch with our CS leaders personally. After her extended advisor period ran out, I recruited a new advisor to coach the CS team. This person had more senior roles at companies that were name-droppable. Her most recent one was a SaaS company that sold to medium and large companies with high ACV, so the parallels to Modern Treasury were clear. After she became an advisor she skipped calls and barely answered our emails. We regretted putting her in the program. Yet another reminder that name brands really don’t mean as much as specific experiences. ## The Mechanics For each department, (2 I used department very loosely here. Sometimes it was a team, sometimes it was a bigger group.) the program had two slots. Advisors signed a standard agreement for one year that we could optionally extend for a second. They earned shares monthly based on a simple calculation of hourly rate and hours per month. Most advisors had high rates and we agreed to work together 2 or 4 hours per month. So the math for an advisor with a rate of $1K/hour, for 4 hours, at a $10 share price, was simply: ![advisor math](https://substackcdn.com/image/fetch/$s_!Eqah!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a1f2e09-b472-44ed-a85f-8b8779e1512a_1956x1030.heic) Even though vesting was based on monthly hours, the engagements weren’t highly structured. We usually used those hours in chunks, for example, a sales advisor spending a day at our sales team offsite. We’d engage ad hoc in between. The program agreements became so standard that it was our Head of People who administered the paperwork, not our Head of Legal or Head of Finance. The agreement was just over one page and there was a one paragraph appendix for stock vesting. Even though each department had two spots, in practice, most departments only had one advisor at a time. Some had periods with zero. But the reason we set up the program with two is twofold: first, once in a while advisors needed to overlap, and we didn’t want some petty rule to get in the way. But more importantly, having two slots set up a psychological phenomenon where bringing on an advisor never felt like firing your one shot. So it took away the sense of always being able to do better because if a truly better foundations engineering advisor came along, we could bring them in too. In practice, that never happened. It was just a trick of setting up the program that worked well. ![012 illustration](https://substackcdn.com/image/fetch/$s_!dW-c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F456a0773-3651-4cd2-9554-94316de5898c_944x458.heic) ## Getting Advisors Advisors came into the program a few different ways. First, if the head of that department needed coaching in that area. They identified that need, or we did. For example, the woman who ran marketing for me was excellent. She was highly skilled in the communications, brand, and storytelling side of marketing. As we grew and needed to build more operational demand gen programs, we brought in an advisor who had run that engine at Mulesoft. That advisor helped us both. Second, to extend or deepen our network. One of our investors might suggest someone that we should get close to or learn from. We had partners suggest people, too, and we even had partners change jobs then become advisors once they no longer had a conflict of interest. Third, to capitalize on opportunistic relationships. The founders or I could meet someone that we thought would be useful, but we weren’t sure how yet. An example was my buyer at Cash App, who left Block a few years after they implemented Modern Treasury. As we were expanding our product into the workflow side of payment operations, I thought he could help us. I didn’t have a specific project, I just wanted to compensate him for being “on call” for our product, product marketing, and sales org as we brought that product to market. He ended up doing sessions with each of those teams that year. ## Nothing Burgers That Actually Work Out Sometimes advisors didn’t end up doing a ton during the year. But that turned out fine, too. Because we designed the compensation program to be annual with optional extensions, the downside was limited. We ended up with ~25 people who were senior in the ecosystem who were connected to us and rooting for us: folks we could text when we needed something without feeling guilty about it. Sometimes a reference call is all we needed to set a deal, close a candidate, or get a partnership over the line. In a way, it was like each department had a short list of angel investors. ![nothing burger](https://substackcdn.com/image/fetch/$s_!rnOy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa78f649f-f31d-406c-aa4f-60260da8b595_695x584.heic) Lastly, and I think most importantly, the way we structured the program took any story away from getting an advisor. It didn’t have any negative connotation. They were just joining a program we had done 20 other times. We didn’t run it perfectly. One thing I’d change is that not all departments used the program equally, and I would reward participation more strongly next time. But in any case, I will 100% implement this at the next company I join. It was really useful. ![: thanks for reading](https://substackcdn.com/image/fetch/$s_!3RDP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72384c8b-3ed8-4f2c-9007-25e084ac922d_1102x360.heic) 1. I learned, yet again, the banal lesson that pedigree doesn’t necessarily mean excellence. [↩︎](#fnref-rss-1) 2. I used department very loosely here. Sometimes it was a team, sometimes it was a bigger group. [↩︎](#fnref-rss-2) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/how-to-run-a-good-startup-advisor-program --- --- title: "A COO's Unnamed Work Type #4: Janitor" description: The last category of unnamed work was the most fun to name. date: 2026-04-04 url: "https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-4-janitor" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # A COO's Unnamed Work Type #4: Janitor The last category of unnamed work was the most fun to name. I’ve heard it called “cleaner upper,” “babysitter,” “caretaker,” and “fixer,” but I think the best name is “janitor.” Being a janitor means taking over a function for a discrete period of time, restructuring it, and then landing it in another part of the company. It starts with an ask that sounds like this example from my [interviews](/essays/off-the-record-coo-project): my CEO said to me “All we know is this broken. Marketing is not performing. Can you just go… figure it out?” (1 The next sentence was: “I have played that role many times.”) There’s many reasons why a company might need this. A function could be in disarray, leaderless, underperforming, or the remit needs to be adjusted. (2 Note that none of these are “a net new function needs to be built.” Although COOs do that, so do all leaders as their teams and companies grow. So I don’t view that as part of being a janitor specifically or as a part of unnamed work.) Or the [the org chart](https://bonkersworld.net/organizational-charts) might be wrong, or wrong for that startup at that moment. ![broom](https://substackcdn.com/image/fetch/$s_!s0vz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07fd4f6c-962c-47bc-b0c6-bc08ff28ab00_1065x1178.heic) To do this well, the COO needs an ability to get up to speed quickly, comfort flexing from macro to micro, ability to develop rapport and trust with a new team quickly, judgment, and lack of ego. The last one is particularly important. Otherwise it would be tempting not to place this team back in another part of the org. The spark for this essay was a conversation with a COO in healthcare. This person is a delivery and G&A [type of COO](/essays/the-three-types-of-coos), so her ultimate KPI to the company was cost. (3 For example, although CS teams are often considered in cost, at one point in this company the customer success org switched priorities and started to focus on revenue generation. At that time, this COO gave that team away.) The big surprise is that for awhile, this COO led all of engineering. She called it being that team’s “caretaker.” At the time, the main goals of the engineering team overlapped highly with her operations teams. They were shipping products to reduce cost and improve operational efficiency. There was some org and leader change, so she took it over until they found a new technical leader, then landed it back in its rightful place. Another COO described the work as care without bias: > The way that the product and engineering teams were working was really, really poor... Now you had this thing that could become very political, and to fix it, you need somebody who is completely unbiased. You need to be neutral. You need to only care about the company. Founders do this type of thing frequently. It’s even a way people describe “[founder mode](https://paulgraham.com/foundermode.html)” now. (4 This is not my interpretation of the original essay, but it is how I hear the term used a lot now.) It’s less commonly known how often COOs are running part of the company that doesn’t make any sense from the outside. That’s the point. A janitor works invisibly, cleaning up messes that other people make. Invisibility is what defines all four types of unnamed work. They’re also temporary, inscrutable, and a bigger part of the role than you might guess. ![thanks for reading](https://substackcdn.com/image/fetch/$s_!Dijb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc4570f60-5436-4fde-80cf-8930a477bb3c_1256x305.png) 1. The next sentence was: “I have played that role many times.” [↩︎](#fnref-rss-1) 2. Note that none of these are “a net new function needs to be built.” Although COOs do that, so do all leaders as their teams and companies grow. So I don’t view that as part of being a janitor specifically or as a part of unnamed work. [↩︎](#fnref-rss-2) 3. For example, although CS teams are often considered in cost, at one point in this company the customer success org switched priorities and started to focus on revenue generation. At that time, this COO gave that team away. [↩︎](#fnref-rss-3) 4. This is not my interpretation of the original essay, but it is how I hear the term used a lot now. [↩︎](#fnref-rss-4) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-4-janitor --- --- title: "A COO's Unnamed Work Type #3: Founder Vault" description: "Of the unnamed work that is emotional or interpersonal, there’s two types: the part that doesn't face founders, and the part that does." date: 2026-03-26 url: "https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-3-founder-vault" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # A COO's Unnamed Work Type #3: Founder Vault Of the unnamed work that is emotional or interpersonal, there’s two types: [the part that doesn't face founders](/essays/a-coos-unnamed-work-type-2-onramps), and the part that does. This is the category of unnamed work that is hardest to write about because its defining feature is privacy. This is the work that happens when only the COO and the CEO (or COO and founders) are in the room. It resembles therapy, but it’s more about being a context-aware vault. As a vault for confidential conversations, a founder gets a place to work things out by talking them through. They can talk about a personnel issue that can’t be discussed anywhere else, vent about another co-founder, or express personal doubts about the business or strategy. One version of this came up when I interviewed a COO who runs founder “unload” sessions. The prompt for these sessions is “tell me everything you don’t like about what we’re all doing to your company.” Then they talk through that prompt, debating why it’s happened, and comes out aligned about what to do about those things. ![lock](https://substackcdn.com/image/fetch/$s_!u6Mj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb199adb2-d862-4a1b-bd2e-86e9656cafed_601x1080.heic) Another thing that happens in private is positive feedback. It’s lonely to be a CEO and it’s exhausting to be a founder. You carry the weight of the company. People rarely say “good job.” (1 I wonder if that’s why some founders proxy their success onto things that aren’t really success: hype, valuation, event invitations, or other signals of cachet that aren’t always correlated with outcomes. They have to get signal from somewhere.) So often the COO takes on the responsibility of saying these things. I’ve seen and heard many versions of this, from specific “I see your improvement in X skill” (for example, public speaking), to general “I think that went really well” (for example, after a press interview, board meeting, or a big event) or even more simply: “I think you’re doing a good job.” ![star](https://substackcdn.com/image/fetch/$s_!9TLH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe798a06e-0a24-458a-a7ed-0259b82e3e94_1180x1255.heic) I have heard stories about a more dramatic type of “therapy” where a COO mediated founder disputes, but that didn’t come up directly in my interviews. Perhaps it’s so private it couldn’t be mentioned. Of course many other people can play the role of vault: board members, investors, executive coaches, peer CEOs, spouses, or advisors. But the difference is that all of these people are outside the company. So getting detailed, gritty advice requires a founder or CEO to provide a large amount of context, which can be exhausting. It’s easier to turn to a COO. Sometimes, what most defines the COO-CEO relationship is the work that is never visible, and never named. ![](https://substackcdn.com/image/fetch/$s_!1P-F!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feaa981e5-60ba-473c-8620-a6348f4fe0e2_1108x288.png) 1. I wonder if that’s why some founders proxy their success onto things that aren’t really success: hype, valuation, event invitations, or other signals of cachet that aren’t always correlated with outcomes. They have to get signal from somewhere. [↩︎](#fnref-rss-1) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-3-founder-vault --- --- title: "A COO's Unnamed Work Type #2: Onramps and Offramps" description: Take three truths. date: 2026-03-05 url: "https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-2-onramps-and-offramps" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # A COO's Unnamed Work Type #2: Onramps and Offramps Take three truths. First, COOs know their CEOs intimately. Second, many founder CEOs have rough edges. Third, it’s the CEO’s job to make decisions. So many people come before him or her to adjudicate or make a final call on something. The combination of these three truths leads to the second type of unnamed work: onramp and offramps, or emotional translation. The only one of these I think could be contentious is the first truth, knowing the CEO intimately. (1 Marriage is an overused analogy for these type of longstanding business relationships. We don’t have a good word for it in English, perhaps that’s why we use “work spouse” instead.) But it was present in almost every duo that I got to hear about through [my interviews](https://www.workinghypotheses.com/essays/off-the-record-coo-project). Of the nearly thirty people I’ve met so far, this type of close understanding was present for all but two. I was worried this was an extension of the fact that so many of the COOs I interviewed are women. But then it showed up when I interviewed men too. At one point one male COO put his head in his hands across from me over Zoom and paused. Then he said “[my CEO] just could not, almost pathologically, relinquish control. She couldn’t help herself.” (2 He also said “taking things to her was a red line factory,” which I thought was an evocative turn of phrase.) He continued, clearly demonstrating how well he know his CEO. ![highway exit sign](https://substackcdn.com/image/fetch/$s_!QrLF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d073c03-bafa-4390-b06a-787cf7742f1f_1216x1046.heic) In the onramp, high stakes decisions often get dress-rehearsed in front a COO. Can you help me figure out how to present this information so that the decision goes in my favor or the way I want? Or, I don’t care how this is decided but I don’t want to upset the CEO or get a bad reputation by presenting it badly. (3 For example, presenting information or data that shows weakness or nonperformance in their part of the business.) Can you help me figure out how to say it? Essentially, COOs are training other employees and leaders to to communicate in a way that will land with the CEO/founder. I wonder if CEOs think about how often a COO is shaping the stories that are shared with them. Not to a particular end, (4 Or if there is a particular motivation, perhaps they are bad at their job.) but because their advice is sought out. In the offramp, a COO cleans up meetings that go badly or communications that don’t land. In one conversation over my kitchen table where a COO told me that her CEO is “ruthless,” “unempathetic” and “prone to making whoever she speaks to cry.” One of my main jobs, this COO confessed, is steeling people to that likely outcome before they walk into the room (onramp) or making them feel better after they walk out (offramp). Another COO describe the offramp as translation combined with emotional support. “I often say something like ‘his is what [he or she] meant. This is the outcome they are trying to land. They didn’t mean it that way.’” A third COO describe it as “I’m their EQ foil.” (5 I learned that it’s from jewelry making where the foil makes the jewel shine more brilliantly.) ![gemstones](https://substackcdn.com/image/fetch/$s_!fGXJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3ff5243f-adcd-4e57-aff0-e97db12026ca_721x816.heic) A consequence of the onramps and offramps work is that many COOs end up influencing or determining what CEOs get presented to about. In these cases the work is notabout emotional shielding but about reducing thrash. At first blush, I suppose that could read like a bad thing. I could imagine a CEO wanting all the information about all of the things. But in actuality, that would slow down him or her and the company. As long as the right, relevant, and appropriate information is flowing to me, and what those things are is explicit and agreed upon, it will make a company move more quickly. “One of my biggest advantages as COO is knowing what my CEO cares to know,” said a COO to me last summer. “I know what he wants to be bothered about, and what he doesn’t.” (6 At Modern Treasury, I helped countless people prepare to share information or set up decisions with Dimitri, and I batted some things away from his proverbial “desk.” I almost always told him that I had redirected people, but only after it had happened.) What she’s describing might be why this work stays unnamed: the deep yet almost involuntary read on another person that informs so much of what a COO does. It’s the kind of thing a person doesn’t know they’re doing until someone points it out. 1. Marriage is an overused analogy for these type of longstanding business relationships. We don’t have a good word for it in English, perhaps that’s why we use “work spouse” instead. [↩︎](#fnref-rss-1) 2. He also said “taking things to her was a red line factory,” which I thought was an evocative turn of phrase. [↩︎](#fnref-rss-2) 3. For example, presenting information or data that shows weakness or nonperformance in their part of the business. [↩︎](#fnref-rss-3) 4. Or if there is a particular motivation, perhaps they are bad at their job. [↩︎](#fnref-rss-4) 5. I learned that it’s from jewelry making where the foil makes the jewel shine more brilliantly. [↩︎](#fnref-rss-5) 6. At Modern Treasury, I helped countless people prepare to share information or set up decisions with Dimitri, and I batted some things away from his proverbial “desk.” I almost always told him that I had redirected people, but only after it had happened. [↩︎](#fnref-rss-6) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-2-onramps-and-offramps --- --- title: "A COO's Unnamed Work Type #1: TV Producer" description: "After about ten COO interviews, I started thinking about my last job differently." date: 2026-02-21 url: "https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-1-tv-producer" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # A COO's Unnamed Work Type #1: TV Producer After about [ten COO interviews](/essays/off-the-record-coo-project), I started thinking about my last job differently. Many leaders, when asked what they do, respond with their title, eg “I’m COO of Acme Corp Startup.” And if you then ask what that means, they list the functions they manage. For example, that might be, “I’m responsible for the marketing, partnerships, post-sale, and analytics teams right now.” But in my last job, the truth is that was only half the work. Or less. The other half was a set of responsibilities that I now think of as unnamed work. I didn’t realize how much of this work I did, or how much others do, until I started my interview project. Once I had this realization, I made a list of all the types of unnamed work. It was all in a single category: “the stuff no one sees.” But I now realize there are different components, and different COOs look after different subsets. (1 Unnamed work is often the job of a COO, but not always. In all my interviews so far, a COO has done between one and three of the four categories.) The first category of unnamed work is fairly logistical. It includes: - Convening conversations that need to happen - Hosting or emceeing big public events (eg user conference, customer advisory board, partner summit) - Planning company offsites or kick offs - Facilitating executive meetings or offsites - Managing regular company cadences - Spinning up new teams - Ensuring north star goals or company strategy are properly communicated to the company I thought for months about what to call this category, and with the nudge of a good friend, I now think the most apt analogy is the role of TV producer. (2 This is an incredible essay on being a leader in TV.) In this case, the product is not a show, but the company itself. And the person who does the day-to-day work of making that happen is the Producer, the kind with “p.g.a.” after their name. ![: TV illustration](https://substackcdn.com/image/fetch/$s_!eEft!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F48b76394-3e28-4adf-a3ce-54d6dd3a8105_1119x889.heic) I like this analogy because it’s easy to picture making a TV show. The Producer is not the creator of the story, nor are they the director. (3 Although of course they sometimes steer or nudge these things in one direction or another.) They pick up the logistical, team, and communication pieces that creators and directors (in our analogy, founders) [aren’t well suited to](https://mollyg.substack.com/p/founder-mode-a-liars-perspective). “I guess, [the CEO] is like, all vision, very little execution. I am, like, all execution, very little vision,” one COO told me. Another COO shared with me that she writes a document every week of what each exec should be sharing with their direct reports. She is not in the room for those conversations, and she acknowledged that this is a strange thing to do. But it’s a type of production her company needs so she does it. This is an example of the first item in the list above and an obvious example of unnamed work. If you asked this COO about her job, she would never say “I tell each exec what hard and important conversations they need to have every week so they don’t procrastinate it and we stay on track and grow.” But she does do that, and it takes a lot of her time and effort. The skills used in this type of unnamed work are: communication, public speaking, one-to-many storytelling, facilitation, influence, and project management (or management of project managers). ![: on off switch illustration](https://substackcdn.com/image/fetch/$s_!cwp1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76a2e3ee-30db-4fd7-99ee-d2152a050433_869x1005.heic) One thing that’s nuanced about the unnamed work of production is knowing when to step in and step out of the role. To be a good facilitator, sometimes it’s more effective not to have an opinion, but to elicit quiet voices and guide the conversation toward the thorniest topics. So how do you facilitate and simultaneously get your views represented, without each of those things stepping on each other? One COO-turned-CEO shared with me that she sometimes asks another leader in her company who shares her view to represent them for her, so she can fully be in the facilitation role. Meaning, she explicitly steps into the Producer role for certain meetings, then steps back out of it for another meeting where she’s making a decision. Every COO does some version of this invisible work, and not one of them leads with it when they describe their job. The unnamed work stays unnamed. But it’s often the work that determines whether a company runs well, or just looks like it does. ![thanks for reading illustration](https://substackcdn.com/image/fetch/$s_!QUGo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2e3ce4bd-83e0-4321-8cbb-67caecb940e9_1108x288.png) 1. Unnamed work is often the job of a COO, but not always. In all my interviews so far, a COO has done between one and three of the four categories. [↩︎](#fnref-rss-1) 2. This is [an incredible essay](https://okbjgm.weebly.com/uploads/3/1/5/0/31506003/11_laws_of_showrunning_nice_version.pdf) on being a leader in TV. [↩︎](#fnref-rss-2) 3. Although of course they sometimes steer or nudge these things in one direction or another. [↩︎](#fnref-rss-3) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/a-coos-unnamed-work-type-1-tv-producer --- --- title: The Shape of Work is Changing description: "If you draw a chart of effort applied in different phases of making things, the shape of the curve on that chart is changing." date: 2026-02-05 url: "https://www.workinghypotheses.com/essays/the-shape-of-work-is-changing" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # The Shape of Work is Changing If you draw a chart of effort applied in different phases of making things, the shape of the curve on that chart is changing. For the purposes of this essay, it’s useful to think of work in three stages: planning, producing, and perfecting. (1 While these reductive phases happen roughly sequentially, you can jump back and forth if needed. Also this is more suited to some types of work than others. It maps to writing a memo, for example, but less to managing a team.) Planning is all the stuff you do before you start to produce any particular work artifact. It can include gathering data, putting together a strategy, researching, reviewing, deciding what to work on. It’s doing the thinking. Producing is making the artifacts of knowledge work: memos, deals, contracts, presentations, templates, code, designs, agendas, plans, betas. You are making an actual thing a colleague can see, touch, interact with, and react to. Perfecting is the most obvious. It’s editing, refining, revising, adjusting. ![phases of work](https://substackcdn.com/image/fetch/$s_!ig2o!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcc488b3-1618-4d45-a2d5-6c437618a321_1815x407.heic) Before AI, I believe most work followed a chart that roughly looks like a shifted sine curve. For a typical work product, the level of effort would be small, large, then small again. (2 This has been how work felt to me for my entire career, across academia, investing, and operating. It’s the stage where I am most likely to get into flow state. I can plan all I want but until I’m creating (producing), I’m not really in it in the same way.) ![phases sine curve](https://substackcdn.com/image/fetch/$s_!aa99!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F39aa2e0d-3cc1-4f85-b8f0-19255735ffa7_1336x1073.heic) Now, the three phases are inverted. With AI tools, production is asymptotically becoming zero effort. A deck, code, a memo: they can be made nearly for free now. But the other two phases require more effort than before. Gathering or producing novel data, or slurping it up into a way that an AI tool can use, that takes a lot more time and effort than it used to. (3 Or said another way: you still need to think.) Planning out what to produce, and then prompting, is a new part of this phase that didn’t exist before. (4 There’s an argument that PRDs are the planning phase for engineering work. But I think because PMs aren’t engineers, the PRD is the end of the producing phase for that person, and the code is the product for the engineer. So this model still applies.) And perfecting! As anyone who has reviewed AI output knows, this phase takes longer than it used to. The shape of the curve looks like a shifted cosine now. ![phases cosine](https://substackcdn.com/image/fetch/$s_!8w8Q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff5f79c6b-d59a-43ae-a031-af1aa5f823c4_1336x1073.heic) One strange consequence of this is that what used to look like work doesn’t look like work anymore. Thinking, walking, reading, interviewing, staring off into space. This is the biggest part of “productivity” now. (5 Arguably it always has been, but when producing took more effort, the percentage has shifted.) I wonder how long it will take us to catch up to this. It could be a very long time, because what the work curve used to look like has been consistent for so long. (6 I know for myself that writing and rewriting a prompt, or prompting an agent 137 times, doesn’t put me into flow state the same way writing does.) I spoke with a CEO friend of mine last year. She complained about AI memos. Her executive team kept sending her memos for this, memos for that. Now that writing a memo takes near-zero effort, they could be more productive! But they hadn’t caught up. Thinking and perfecting are the work now. But I feel for them; the main thing they used to bring to their boss is now free to make, so they are making more of it. Last year I went to a vibe coding hackathon and experienced this first hand. I spent most of the day banging my head against the proverbial design wall. I don’t usually write memoir-style essays, but I’m including it below as it describes the feeling of expecting a sine curve and experiencing a cosine one. ![: thanks for reading illustration](https://substackcdn.com/image/fetch/$s_!1PJH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62eabb92-8e74-4e93-b2ba-1a9aef3d7e59_1011x293.heic) ### A Vibe Coding Nadir I went to build a personal app: a recipe keeper with version control. (7 I like to follow the scientific method of cook > taste > tinker > adjust recipe. I want software to re-home my messy recipe notes that currently live in post-its and Google Docs, including one titled “Neverending Experimental Perfection Bagel Recipe.”) Although it stopped after Covid, I used to have a habit of collecting cookbooks and reading them cover-to-cover. I imagined an app that evoked those afternoons reading a cookbook purchased that day, marking the recipes I wanted to make first. ![recipes](https://substackcdn.com/image/fetch/$s_!tG-9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe6503a75-562c-475f-91af-94d8c5eaa79c_868x1056.heic) Using ChatGPT and Claude, I wrote a PRD, then uploaded that PRD into the vibe coding tool. I tinkered with the functionality on the edges, nipping and tucking in a few places. Watching the agent work was pleasant enough, evoking the lumbering but forward-moving feeling of a large rock rolling along the bottom of a stream. Then v1 was done, and it felt good. This is it: from idea to clickable thing in less than an hour! Predictably, pizza arrived. As my fellow participants vacated our communal table, I looked down at the app I had created. It was hideous. It was white with accents of construction-sign orange: neither vibrant nor neon, squint-inducing yet with putrid undertones. It didn’t invite an appetite. Buttons had the rounded corners of 2019, the font was indiscernible muck, and the recipe creation form was less appealing than a form on the California DMV website. Why was it a form anyway? I hated it. In my mind, I wanted a bookish, spacious place to tinker with the most sensual thing: food. I wanted to combine the expansive white space of writing, the minutiae of scribbling recipe changes in book margins, and the hearth and beauty of home cooked food. My mother-in-law used to say, “food works all the senses. First you hear it, then you see it, then you smell it, then you taste it.” This app repulsed all my senses. For the next two hours, I fought to get the agent to design something I could stand. I started by going back to planning, and writing down in more detail what I thought I wanted. But I hit the boundaries of my vocabulary almost immediately. Although I have been managing creatives and creative processes for over ten years, until this moment I hadn’t realized how much my design counterparts had been doing to translate what I thought I was saying into what I was really saying. If I said “make it more bookish” while waving my hands in the air, then my Creative Director might say to himself “let’s switch to a softer serif, like Iowan Old Style, reduce that line height, use spacing to create hierarchy rather than big swings in font weight and size, and maybe have asymmetrical margins to make that bookish feel.” (8 I actually had to email Frank to get this quote here.) If I said “I want this to feel more vibrant and warm,” he might mentally translate that into contrast, spacing, or color combinations. I railed against the limits of my language for about thirty minutes, getting nowhere with the single exception of removing the putrid orange. Now I had a functional app in black and white. I decided to use an example to explain what I meant. Dear agent, here is what I’m looking for: here is a sample site designed by a friend. Isn’t it great? See the fonts and lines? It feels literary. Make mine like that. But the agent couldn’t parse what characteristics of that model website I was attempting to replicate, even when I told it in my halting designglish, instead it implemented the shape of the grid system and left fonts, spacing, and lines out of it. Okay! I will give it more examples that are similarly bookish. But there’s one problem. I can’t think of any. I am not a designer in lexicon nor in visual memory. I don’t have a mental library of brands, websites, images, concepts, or artists. I imagine a designer’s mind as a neural net of visual nodes. With a stimulus prompt, they can reach into their neural net and find associated other visual nodes that branch off along varying patterns or characteristics. Bookish like this news site? Or this site with images of libraries? Do you mean bookish like the books that this publisher creates? Or by bookish do you mean large serif font like this brand uses, which is not actually how books are printed at all? Or do you mean the types of fonts this foundry makes? Et cetera. I went to chat with my friend who works at the vibe coding company. He’s a designer, so I asked him, “how do you give the agent you design design input?” “Oh it’s really hard,” he said. “The best thing to do is come up with example websites, around four or five. And if that doesn’t work sometimes I just ask it to build the app three different times, then adjust from whichever one is the starting point that most looks like I wanted it to. It never builds the same thing twice.” (9 Funnily enough, this is exactly what designers do with clients. They present three concepts, then play “Marco Polo” as my friend Frank calls it. Even though no one wants to admit it, the first round is just for calibration, then it goes from there.) It had been nearly three hours of vibe coding. The prizes were going to be awarded in an hour, and my black and white, DMV-inspired recipe keeper showed no signs of change. My best hope was to start over, three more times, in parallel. I packed up my laptop and left 30 minutes before anyone showed their work. 1. While these reductive phases happen roughly sequentially, you can jump back and forth if needed. Also this is more suited to some types of work than others. It maps to writing a memo, for example, but less to managing a team. [↩︎](#fnref-rss-1) 2. This has been how work felt to me for my entire career, across academia, investing, and operating. It’s the stage where I am most likely to get into flow state. I can plan all I want but until I’m creating (producing), I’m not really in it in the same way. [↩︎](#fnref-rss-2) 3. Or said another way: [you still need to think](https://calv.info/you-still-need-to-think). [↩︎](#fnref-rss-3) 4. There’s an argument that PRDs are the planning phase for engineering work. But I think because PMs aren’t engineers, the PRD is the end of the producing phase for that person, and the code is the product for the engineer. So this model still applies. [↩︎](#fnref-rss-4) 5. Arguably it always has been, but when producing took more effort, the percentage has shifted. [↩︎](#fnref-rss-5) 6. I know for myself that writing and rewriting a prompt, or prompting an agent 137 times, doesn’t put me into flow state the same way writing does. [↩︎](#fnref-rss-6) 7. I like to follow the scientific method of cook > taste > tinker > adjust recipe. I want software to re-home my messy recipe notes that currently live in post-its and Google Docs, including one titled “Neverending Experimental Perfection Bagel Recipe.” [↩︎](#fnref-rss-7) 8. I actually had to email Frank to get this quote here. [↩︎](#fnref-rss-8) 9. Funnily enough, this is exactly what designers do with clients. They present three concepts, then play “Marco Polo” as my friend Frank calls it. Even though no one wants to admit it, the first round is just for calibration, then it goes from there. [↩︎](#fnref-rss-9) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/the-shape-of-work-is-changing --- --- title: Anti-process leads to Sprezzatura description: One common thing that happens in startups is founders renounce the whole idea of planning and process. date: 2026-01-13 url: "https://www.workinghypotheses.com/essays/anti-process-leads-to-sprezzatura" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Anti-process leads to Sprezzatura One common thing that happens in startups is founders renounce the whole idea of [planning](/essays/sprezzatura-and-the-planning-fallacy) and process. I’m not sure if this is because coordination isn’t needed when startups are very small or if it’s because planning so often fails. It turns out that renouncing process creates a new set of problems. If a founder has renounced process, but getting to some kind of answer (product, data, strategy, etc) requires it, what happens? The person responsible for that thing doesn’t want the founder to know what’s going on behind the scenes to come up with a good answer, because the founder announced that process is bad. But they also don’t want to reduce quality of their work. So their response has to look spontaneous while also being rigorous, well-reasoned, and cogent. What’s happening is sprezzatura. Sprezzatura is an italian word that refers to the work and effort it takes to appear nonchalant or carefree in the moment, while also being excellent. It’s the work it takes to appear both excellent **and **effortless. The latter is what makes sprezzatura different than other forms of preparation. ![tangled illustration](https://substackcdn.com/image/fetch/$s_!C-6u!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc23fdfbf-923d-4f2c-ba82-1db67f30e07f_1787x678.heic) Sprezzatura is found in fine [dining](https://www.newyorker.com/culture/the-food-scene/kisa-a-pitch-perfect-ode-to-korean-drivers-restaurants), [hotel hospitality](https://www.ft.com/content/3d2437b1-cb44-4bfb-a43e-8a7d580e4a2b), [fashion](https://x.com/dieworkwear/status/1814761123570458809), and originally [fine art](https://www.theartnewspaper.com/2022/04/20/the-big-review-donatello-in-florence). It’s attention to detail that’s so refined the creator introduces a bit of mess, and it presents as ease or grace. And that little bit of mess is what sets the other party at ease. Their need is not a burden; it’s weightless. This is how a[host](https://www.amazon.com/Unreasonable-Hospitality-Remarkable-Giving-People/dp/0593418573/) at a 3-star Michelin restaurant is striving to make their guests feel. Startup leaders are striving to make their zero-planning founders feel the same way about planning. The right answer appeared effortlessly out of nowhere. The part about sprezzatura that’s confounding is that the work to appear* *effortless increases the total amount of work enormously. Introducing the right mess requires a high level of mastery. One obvious component of sprezzatura is secrecy. Hiding work is more effort than doing something transparently, because you have to remember when and where to say or not say things, and who can and can’t be included. It’s an increase in cognitive overhead, which distracts from more useful, company-building tasks. One tip off that this dynamic might be happening is when a leader asks a founder, “hey, can I have a little more time to prepare xyz thing, so I can talk to my teams about it?” The telling part is actually the response: “No. Just don’t do anything to prepare.” (1 There are obviously countless reasons that meeting urgently could be the right thing to do. The point here is the founder is ignoring signs of sprezzatura.) What that really means is: prepare a lot, because you are representing tens or hundreds of people and their work as it relates to the new, top priority of the company. But then, spend extra effort to appear careless and ruffled, just enough so that it convinces me that you haven’t worked hard, even though what you show me still has to be excellent. That’s sprezzatura. It’s a lot of pantomime and pageantry to avoid reality. And it’s not worth it. ![thanks for reading illustration](https://substackcdn.com/image/fetch/$s_!8lJf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F414369a2-30ae-406b-b2d6-bab8a42a4e4d_1041x326.png) *Thanks to CR, Chris, Perry, Ami, Rob, and Dimitri for reading drafts of this essay. And thanks to Sanjay for teaching me the word sprezzatura years and years ago. It’s a good one.* 1. There are obviously countless reasons that meeting urgently could be the right thing to do. The point here is the founder is ignoring signs of sprezzatura. [↩︎](#fnref-rss-1) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/anti-process-leads-to-sprezzatura --- --- title: The Three Types of COOs description: "About a year ago, I ran into a founder CEO friend of mine." date: 2025-12-21 url: "https://www.workinghypotheses.com/essays/the-three-types-of-coos" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # The Three Types of COOs About a year ago, I ran into a founder CEO friend of mine. His company was in the middle of rapid scaling and he mentioned he was thinking about hiring a COO. “Oh, what kind?” I said. There was a long pause. “What do you mean, what kind? I just have the feeling that it would be good to have help.” The role of COO can mean many different things. (1 Nate Bennett and Stephen A. Miles wrote about the role with seven archetypes. Here’s the book version.) As I’ve said before: no one quite knows what a COO does. So as I started [to interview COOs](http://I know for myself that writing and rewriting a prompt, or prompting an agent 137 times, doesn’t put me into flow state the same way writing does.), I took notes on their core responsibilities. Over time it became clear that there are three main COO archetypes. ![three types of COOs with stars illustration](https://substackcdn.com/image/fetch/$s_!ofOO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe4c10563-1ad7-4d2e-8758-186713f625ed_1583x1178.png) ### Operations The first archetype is operations. This may seem like a silly thing to say, that a COO is responsible for operations. But depending on what the company makes, it’s not. Many companies don’t have operations at all. What is operations for a Saas platform? And even for those that do, what that person is responsible for can vary a lot. For example, I interviewed a COO of a satellite company. His teams have included spacecraft production, supply chain quality, manufacturing (including their in-house machine shop), and the information systems behind those. That’s a very specific set of operations for that particular business. (2 I was worried when I interviewed this person that his experience would be so different than mine that we wouldn’t be able to relate. About halfway through the interview we started talking about unnamed work, and it turns out he does almost all of this for their executive team, so it turned out our roles were much more similar than I had imagined.) Of those I interviewed, the majority of COOs that fell into this archetype work in healthcare. The reason why is that a lot of work in healthcare is mediated through humans. These COOs jobs were to hire and scale these high headcount teams, in call centers or physical delivery locations. Their teams are often comprised of hourly workers and the workforce is notably different than the rest of the company. From a headcount perspective, sometimes these COOs manage 60% to 80% of a company. ![relative team size illustration](https://substackcdn.com/image/fetch/$s_!5q_Y!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff96f17ff-b45e-4692-906d-c5c66eeeca89_2585x1542.png) One thing multiple of these healthcare COOs mentioned is that they were responsible for carrying the separate culture of their teams. An employee in a health care call center in a rural or exurban area is not going to immediately relate to a lifetime software engineer in an urban office. These COOs sometimes run separate all hands, value reward systems, and rituals from the rest of the company. They serve as a sort of cultural crosswalk. Sometimes, these folks even have a different title: Chief Operations Officer instead of Chief Operating Officer. For this category of COO, the core metric that these leaders are responsible for is gross margin, or something adjacent to it: COGs, efficiency, cost reductions, etc. ### G&A This is the least frequent type. G&A COOs manage a portfolio of functions that could include: IT, Workplace Facilities or Real Estate, Finance, HR and People, Government or Regulatory Affairs, Procurement, Compliance, RevOps, BizOps, Legal, Data and Analytics. This version of COO is sort of a descendent of the title CAO, Chief Administrative Officer. Surprisingly, this title has [little written about it](https://en.wikipedia.org/wiki/Chief_administrative_officer). From what I can find, it came to prominence in corporate America in the 1990’s, starting with companies like Gap. It’s most common in banking and then healthcare, perhaps because of regulatory burden in those areas. But there are a few notable examples in tech from semiconductor companies (NVIDIA, AMD, Intel). It seems that the CAO title has mostly dropped out of favor, and maybe that’s why these COOs have the title they do. This type is responsible for resource utilization, broadly interpreted, meaning ensuring total overhead costs are optimized. So it’s essentially operating margin. (3 This is distinct from COGs as it’s the supporting functions, which impact operating margin instead of gross margin.) ### GTM The most familiar pattern in the valley is a COO who is essentially a CRO+. This model was made famous by Sheryl Sandberg (Facebook), George Hu (Twilio), and later Claire Hughes Johnson (Stripe). These folks built revenue organizations while their founders focused on product and engineering. Perhaps this is most common because many founders and CEOs are product focused or otherwise technical, and they pair up with someone who can help them scale the machinery of revenue. This type’s core metric is obvious: revenue and revenue growth. They are responsible for all or most of the five functions of go to market: marketing, sales, customer success, partnerships, revops. Sometimes they had customer support, but only about half the time. ### All Together Some of these categories get blended in actual startups. That’s where the Venn diagram overlaps come in. (4 I had a COO role that was half G&A-type and half GTM-type.) In the messy real world of startups, these “overlap” jobs mean leaders manage different teams to different core metric for their business. Here’s how I see the COO landscape now. ![Three types of COOs with metrics illustration](https://substackcdn.com/image/fetch/$s_!rtLl!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F87c293b5-68c1-42d1-b5ac-84a9a2cef470_1583x1178.png) I’ve drawn this diagram for quite a few CEOs and recruiters now, and it’s been useful. Starting a hiring process without the type in mind is how you end up with a very capable person who is used to solving the wrong problem. ![thanks for reading illustration](https://substackcdn.com/image/fetch/$s_!63pd!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5afb9c3b-f1a7-40db-9c27-b70bc3831771_1256x305.png) 1. Nate Bennett and Stephen A. Miles [wrote about the role](https://hbr.org/2006/05/second-in-command-the-misunderstood-role-of-the-chief-operating-officer) with seven archetypes. [Here’s the book](https://www.amazon.com/Riding-Shotgun-Role-COO-Updated/dp/1503600386/) version. [↩︎](#fnref-rss-1) 2. I was worried when I interviewed this person that his experience would be so different than mine that we wouldn’t be able to relate. About halfway through the interview we started talking about unnamed work, and it turns out he does almost all of this for their executive team, so it turned out our roles were much more similar than I had imagined. [↩︎](#fnref-rss-2) 3. This is distinct from COGs as it’s the supporting functions, which impact operating margin instead of gross margin. [↩︎](#fnref-rss-3) 4. I had a COO role that was half G&A-type and half GTM-type. [↩︎](#fnref-rss-4) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/the-three-types-of-coos --- --- title: Revenue is a Feeling description: One of the best conversations I had was with a COO who works for a solo founder. date: 2025-12-15 url: "https://www.workinghypotheses.com/essays/revenue-is-a-feeling" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Revenue is a Feeling One of the best [conversations](https://www.workinghypotheses.com/essays/off-the-record-coo-project) I had was with a COO who works for a solo founder. Partway through I asked what it was like to work for his founder. After warmly describing his vision and creativity, he sighed, leaned forward, and said “you know what? We had two years in there where my founder kept telling me one thing over and over again. He kept saying ‘revenue is a feeling.’” On the recording, you can hear me laugh, awkwardly and loudly. “But revenue is the exact opposite of a feeling!” “I know,” he said. “I went nuts. I was like, I think I need to quit. I talked to my wife about it.” But this COO stuck with it. He didn’t ridicule or debate. “One of the tricks to my role and working with a founder like I do is [to] firmly believe that a COO’s job is to bring a founder’s vision to life. Everything they want to do is real, useful. And I knew in this case there was something behind it.” So after years of cautiously cajoling and calmly responding, he began to understand what his founder was saying. It’s not that revenue itself is a feeling. It’s that a spreadsheet won’t tell the founder if things are growing or working. It’s that he knows it before the numbers on the spreadsheet are ever shown to him. ![: spreadsheet](https://substackcdn.com/image/fetch/$s_!hSyD!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5424bf28-65ea-4179-91c7-b275cf45c007_1209x857.png) Framed like that, it feels obvious. Anyone can walk into a startup and feel it if it’s happening. Support tickets are overwhelming the team, ops teams can’t keep up, engineering has to retrench and focus entirely on bugs because scale is breaking things, or all of the above. It’s palpable to every employee. This founder was saying is that he doesn’t need to know the revenue number to know that things are working. He’s seeking that feeling, because it’s the predecessor to the spreadsheet number going up. Another COO I interviewed shared a shockingly similar story. “When I first joined,” she shared, “our co-founder wouldn’t even say the word ‘revenue.’ He’d say, “sales is just an experiment we’re running. Can you imagine being an AE on the ground who’s constantly being told, wait, I’m an experiment?” ![: thought bubble illustration](https://substackcdn.com/image/fetch/$s_!YhFD!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe3d28008-54ab-49b7-bc6c-aef62914987d_843x845.png) This COO also stuck with it, and eventually figured out what the founder was saying. He wanted the company to grow through product-led growth only, without enterprise sales. As it turned out, that wasn’t a good fit for their product. So the COO led the founders and board through a process to figure that out. Both sentences are laughable at first, and both COOs dug underneath that to uncover what was really being said. And despite the difference (one was expressing an insight, the other a preference), in both cases the COO’s role was to decode. Decoding takes patience. After that, you have to decide what to do with what you’ve learned. ![thanks for reading illustration](https://substackcdn.com/image/fetch/$s_!NWRc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc5a9e9e5-1efe-48b5-b610-9a979afa723e_1102x360.png) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/revenue-is-a-feeling --- --- title: Have Wonder description: "I sometimes get asked what to write about, and one hack I’ve used to explain it is: have wonder." date: 2025-11-24 url: "https://www.workinghypotheses.com/essays/have-wonder" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Have Wonder I sometimes get asked what to write about, (1 In my case that usually means for startups.) and one hack I’ve used to explain it is: have wonder. One way to get to wonder into what you write is to explain the making of. After years in any industry, company, or function, you become inured to the magic of producing whatever you produce. You’ve been around it for too long, so it’s hard to see it as magical. But to an outsider, it is. ![magnifying glass](https://substackcdn.com/image/fetch/$s_!hnvK!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9fe47460-2ec8-4a78-bfe7-d60c254d6080_1089x1008.heic) Many aspects of making something can be interesting to an outsider. Making sets for a broadway play, what it takes to give mice cancer in animal studies, how makeup is formulated. Day-to-day, these things are someone’s job that they get bored with just like anyone gets bored. But to someone who doesn’t do that, it can be interesting. Here are some non-branded examples that have stuck with me over the years: - [Making perfumes](https://www.newyorker.com/magazine/2021/02/01/how-to-make-sense-of-scents) - [The linguistics of texting](https://www.youtube.com/watch?v=UmvOgW6iV2s) (2 I believe that this type of “have wonder” content is what made TED so successful in its heyday. This was in the time when most of the talks were more about how things work and less about changing the world. Another favorite is Lessons from an Ad Man.) - [What happens to a stolen bicycle](https://priceonomics.com/what-happens-to-stolen-bicycles/) - [Airlines are just banks now](https://www.theatlantic.com/ideas/archive/2023/09/airlines-banks-mileage-programs/675374/) - [How sound effects are made](https://www.newyorker.com/magazine/2022/07/04/the-weird-analog-delights-of-foley-sound-effects) The “making of” trick works because it plays on two things simultaneously that seemingly conflict, and the tension is memorable. When you see something done at an exceptionally high level, it inspires awe. Imagine watching a ballet dancer who has trained for twenty five years performing steps exquisitely with a body you could never have, or reading the cover paper in Nature about how a scientist brought a new framework to a field you didn’t know existed, or reading the prose of a writer who brings together words you didn’t know with such precision you have to reread and reread. To behold extraordinary expertise is inspiring. But unless you are still young (or feel young), it also can feel impossible to achieve: “I could never do that!” On the other hand, learning how awe-inspiring products (art, writing, technologies, startups) are made is eminently relatable. There’s always weird shortcuts, tricks, mistakes, and quirks. And those feel similar to the weird quirks in our own lives. So we end up relating: “I could totally do that!” Which is cool, because we are also in awe of it. For me, I know I will never write prose like David Brooks or Anne Lamott, but I do feel like I [could crawl around on the floor](https://conversationswithtyler.com/episodes/david-brooks/) compiling scribbles, notes, and papers or teach basic writing [in Sunday school](https://annelamott.substack.com/p/what-is-this-all-about) to a local community. That’s the trick doing its work. ![bookshelf](https://substackcdn.com/image/fetch/$s_!DhPJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3fae987b-38b9-48c3-9622-e98b378c370d_1577x865.heic) To write (or shoot, take, draw) this type of content, you have to be an anthropologist. You just [landed on planet startup](https://science.nasa.gov/planetary-science/programs/mars-exploration/mission-timeline/how-we-land-on-mars/), and you are trying to figure out how it works. You are watching yourself do what you do. Once you see something ordinary to you through an outsider’s eyes, the magic creates itself. That’s because in order to explain it, and explain why it’s interesting, you have to give a lot of context. And that context is storytelling. Ben and David, probably two of the best storytellers on the planet right now, do this implicitly. Less than 20% of Acquired episodes are analysis or facts. More importantly, they play on the trick: they tell the history of the best companies in the world (I could never build Nike!) but they explain the shortcuts, messy mistakes, and quirks of the companies while doing so (but I have had random dreams with random words in them!). (3 According to Acquired, this is how Nike was named.) Another way to write have wonder content is to explain technical concepts, with awe, at an approachable level. I’m not an engineer but I want to [understand how AI works](https://writings.stephenwolfram.com/2024/08/whats-really-going-on-in-machine-learning-some-minimal-models/), I’m not a climate scientist but I want to understand [how field data is gathered](https://www.ted.com/talks/rachel_pike_the_science_behind_a_climate_headline), I’m not an academic but I want to [understand language](https://lexiconvalley.supportingcast.fm/) as if I’m still a college linguistics student. This type of content is hard to produce at first. It takes technical expertise (mastery) and a sense of what is and isn’t interesting to an outsider. (4 It’s easy to identify content that isn’t written this way. Explaining things without awe (a manual), or wonder but for an insider (Feynman), or content that’s clearly meant to sell (much of the internet).) One helpful thing though: it’s what you know. If you work at a startup, write about what your company does, with the single twist of an altered perspective. Insider to outsider. ![window](https://substackcdn.com/image/fetch/$s_!h09y!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F850fd18a-e0d4-4c26-a4ca-be854fa5daed_994x1183.heic) Some branded examples I like are [Cloudflare post-mortems](https://blog.cloudflare.com/18-november-2025-outage/) or the famous (and controversial) [OKTrends blog](https://web.archive.org/web/20101016072306/http://blog.okcupid.com/index.php/the-biggest-lies-in-online-dating/) from OKCupid circa ~2009. The tone that pulls the outsider in is wonder. Just ask anyone who has young children. Wonder is contagious. And the thing that makes it stick is the trick of layering the impossible with the relatable. ![thanks for reading](https://substackcdn.com/image/fetch/$s_!-IQF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc5a31cbf-32d1-4795-8e21-8c1c083749c9_1102x360.heic) 1. In my case that usually means for startups. [↩︎](#fnref-rss-1) 2. I believe that this type of “have wonder” content is what made TED so successful in its heyday. This was in the time when most of the talks were more about how things work and less about changing the world. Another favorite is [Lessons from an Ad Man](https://www.youtube.com/watch?v=audakxABYUc). [↩︎](#fnref-rss-2) 3. According to Acquired, this is [how Nike was named](https://www.acquired.fm/episodes/nike). [↩︎](#fnref-rss-3) 4. It’s easy to identify content that isn’t written this way. Explaining things without awe (a manual), or wonder but for an insider (Feynman), or content that’s clearly meant to sell (much of the internet). [↩︎](#fnref-rss-4) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/have-wonder --- --- title: "Poka-Yoke & the Planning Fallacy" description: Most startups hate planning. Poka-yoke and how to do it right. date: 2025-11-11 url: "https://www.workinghypotheses.com/essays/poka-yoke-the-planning-fallacy" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Poka-Yoke & the Planning Fallacy Most startups hate planning. Or rather, most founders do. When I hear a founder say “planning,” they say it the way people say “dentist.” Their voice slows and their limbs slacken as they succumb slightly more to gravity. Here are some [some things I've heard](/essays/off-the-record-coo-project): - “By definition, future plans are wrong.” - “All plans are artificial precision.” - “Don’t do work about the work. Just do the work.” - “Planning is a waste of time.” There’s some truth to these, because most planning processes fail when startups are learning to do them. The good news is that they fail for predictable and avoidable reasons. And startups should want to get this right, because planning is a form of coordination, and startups can’t scale without some coordination. ## The Bad Ways How coordination happens matters a lot. In many startups, it happens badly. The first predictable failure mode is drag. This happens when each individual coordination point is too complicated: too many people (20-person Zooms (1 Never better depicted than here.)), too many review loops, unmanaged long meetings, or talking around real issues. Dogmatic adherence to structure is another failure mode. A few I’ve lived through include esoteric plan structures, forced themes or groupings, and grammar strictures. In one planning process certain verbs were allowed and others weren’t. If structure and form usurp usefulness, coordination fails. The third common failure mode is participation trophy planning, where every team gets every activity recognized. Effective planning is prioritization, where teams decide what needs to happen, and what doesn’t**. **I spoke with one COO who was hired into a company of less than forty people then found out they had 160 OKRs. I imagine there was a significant participation trophy culture at that startup at that time. (2 This COO redid the planning structure and culture at that startup, and that company brings in >$100M in revenue now.) It’s not effective. ![drag illustration](https://substackcdn.com/image/fetch/$s_!xuYi!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F46d2d005-df3e-45f6-b44b-1b0211c0d20c_1038x948.heic) One friend of mine called them “success theater,” which is exactly what they are. Obviously, I don’t advocate for those. But some planning and coordination is necessary to build and grow a startup. ## The Planning Fallacy Coordination in companies has been written about for a long time. (3 The earliest source I could find is “The Nature of the Firm” by RH Coase from 1937. This is the paper. The section that’s relevant starts with “first, as a firm gets larger, there may be decreasing returns to the entrepreneur function, that is, the costs of organising additional transactions within the firm may rise.”) To illustrate it: imagine a 40-person startup where planning happens in a relatively simple way: 100% top down, where every single person at every level has three direct reports. You could visualize their planning process this way, where each arrow is a coordination point: ![planning illustration 1](https://substackcdn.com/image/fetch/$s_!1GfT!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc61fee4d-066e-42a9-a6d0-e65877cde7e2_1456x833.png) *Planning top down in a 40-person startup.* That’s a start. But it’s wrong, of course. At the lowest level, managers will bring their teams together for a group discussion. So the number of coordination “arrows” is more likely: ![](https://substackcdn.com/image/fetch/$s_!uQwV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F86afb29a-8209-47c4-8874-ec3ed85f0aa7_1456x924.png) *Planning top down, v2. The fact that the CEO resembles a Christmas tree topper and teams are ellipses only hit me after editing this essay a dozen times.* But then, planning isn’t (usually) one hundred percent top down. There is a loop where priorities are also set bottoms up. (4 The best description of this loop is the W-shaped planning process, as described by my Nels Gilbreth and Lenny Rachitsky.) So that could be visualized this way. ![planning illustration 3](https://substackcdn.com/image/fetch/$s_!66Rs!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5ab584ce-5ec4-464f-a99f-0a1b93682a9f_1456x924.png) *Two way planning.* I think when a founder sees that they think: holy smokes, that’s a lot of arrows. (It’s 42.) The fallacy is comparing this against no planning. This is how they want it to work, or maybe how it used to work. No coordination! Look at all that free time. ![magic! illustration](https://substackcdn.com/image/fetch/$s_!N9ng!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F22dae1d3-0250-436b-8465-4277cd579beb_1138x1078.png) But you can’t build a startup without some level of coordination. How does any one person work on the right thing? So without some kind of planning process, employees coordinate without a system. Which means they all coordinate with each other. (5 A K40 network is beyond my drawing skills. This and the next network drawing are courtesy of ChatGPT.) ![k40 network illustration](https://substackcdn.com/image/fetch/$s_!C3eS!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf11fbae-c3eb-489f-9510-a3db9df351b7_1456x1419.png) *K40 network. Beautiful, isn’t it?* The most striking thing about this is that the number of arrows is much higher. For a 40-person startup, it’s 780: ![network math 1](https://substackcdn.com/image/fetch/$s_!5zIB!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Facbc6b2b-2473-4721-b3ba-2c3f4c4cb157_1146x457.png) For some reason founders often think they are choosing between zero and 42, instead of 42 and 780. This is the planning fallacy. (6 This is a variant of Kahneman and Tversky’s Planning Fallacy from 1979. In this version, founders overestimate one version of the future (the coordination version) and underestimate another (the “magic” no plan version).) In truth, employees won’t coordinate with every other employee. There are more accurate representations of a company as a network, where some teams are fringe and some are core, and the total number of arrows (in network terms, edges) comes down. But even if half are fringe and set into teams of four, the number of arrows is 270. That’s still a lot more than 42. ![270 math](https://substackcdn.com/image/fetch/$s_!puoH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4ead9042-7b2c-47bc-853c-127cfe330ece_1263x946.heic) There are two consequences of the planning fallacy. The first is simple: 780, or 270, is a lot more than 42. That means moving slower. The pernicious problem is that there is a negative feedback loop between the planning fallacy and believing planning is bad. The more a founder pushes for zero coordination, the more the planning fallacy takes over, and things become chaotic. And then, the more a founder may end up with the wrongly held view that their push for zero coordination is working. ## The Good Ways Alignment does not have to equal drag. A better model is poka-yoke, a term from the Toyota Production System which means, roughly, a behavior-shaping constraint. (7 Here’s wikipedia’s primer on poka-yoke. The example I like best is labeling boxes on all sides, so no matter how you absentmindedly put them back on shelves, a label will be facing out.) Instead of imaging process, imagine designing a system that makes it easier to make the “right” decisions and harder to make the “wrong” ones. Some useful principles in approaching planning: 1. Don’t do it too early. Wait until the math of the planning fallacy means it’s time to centralize coordination, then implement planning the right way. 2. First, set top-down direction. This prevents burning cycles on areas the company won’t pursue. 3. Then, limit the number of priorities or plans each team can have. Reward focus and alignment to the top direction, and be explicit about non-goals. Don’t be great at things that don’t matter, which means you have to be clear and in some cases, do the sensitivity analyses. 4. Acknowledge that this work is an ask. Map onto existing pivot points and remove work where possible. Run cadences according to real life and use hard constraints. (8 For example, don’t ask teams to submit work ahead of time. Just do heads down reading in the meeting. Multiple COOs complained to me “if only people would follow the process I laid out for them, then planning would be easier!” To me, that means the process is badly planned. Poka-yoke means planning for real life.) 5. Ruthlessly weed out participation trophy behavior. Communicate clearly that plans are not performance reviews, not everyone’s work will be recognized, and that’s how it’s supposed to be. 6. Don’t overrotate on form. Accept activity-based goals at junior levels, and even mid levels if necessary. However, do not flex on this: achievement of a goal or plan should always be objective. 7. Remember that measurement doesn’t come for free. Align resources (Eng, Data, Analytics, etc) OR change the objective measure/metric to be measurable today, not tomorrow. 8. Time-box steps to a few days each. Design a process that does not eat up more than 25% of any individual’s time during that period, and less for most employees. 9. Ask that teams check with each other about dependencies and work it out on their own. Checking with each other should be quick Slacks or 5 minute phone calls, not big meetings. In cases where dependencies can’t be worked out, give a clear method of how, when, and who to bring that to. And when they bring it, decide and respond quickly (1-2 days). 10. The most important thing is to do all of this in the open. Renouncing planning doesn’t remove coordination, it pushes it into more meetings, secrets, and even [sprezzatura](/essays/what-happens-in-anti-process-sprezzatura). Run a simple, explicit process instead. ## Impacts If you do all this, a light planning process can have meaningful impact in both growth and speed of company operations. It leads to faster resolution of trade-offs, fewer re-litigated decisions, reduced need for exec arbitration, clarified internal and external company narrative, easier cross-functional collaboration (less thrash), improved accuracy of forecast, reduced spend on initiatives that don’t drive revenue. Who wouldn’t want all that? ![: thanks for reading](https://substackcdn.com/image/fetch/$s_!8lJf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F414369a2-30ae-406b-b2d6-bab8a42a4e4d_1041x326.png) *Thanks to CR, Chris, Perry, Julie, Rob, Ami, and Dimitri for reading drafts of this essay. And thanks to Sanjay for teaching me about poka-yoke. * 1. Never better depicted than [here](https://www.newyorker.com/magazine/2016/05/16/hang-up-the-conference-call). [↩︎](#fnref-rss-1) 2. This COO redid the planning structure and culture at that startup, and that company brings in >$100M in revenue now. [↩︎](#fnref-rss-2) 3. The earliest source I could find is “The Nature of the Firm” by RH Coase from 1937. [This is the paper](https://onlinelibrary.wiley.com/doi/10.1111/j.1468-0335.1937.tb00002.x). The section that’s relevant starts with “first, as a firm gets larger, there may be decreasing returns to the entrepreneur function, that is, the costs of organising additional transactions within the firm may rise.” [↩︎](#fnref-rss-3) 4. The best description of this loop is the [W-shaped planning process, as described by my Nels Gilbreth and Lenny Rachitsky](https://review.firstround.com/the-secret-to-a-great-planning-process-lessons-from-airbnb-and-eventbrite/). [↩︎](#fnref-rss-4) 5. A K40 network is beyond my drawing skills. This and the next network drawing are courtesy of ChatGPT. [↩︎](#fnref-rss-5) 6. This is a variant of Kahneman and Tversky’s [Planning Fallacy from 1979](https://apps.dtic.mil/sti/tr/pdf/ADA047747.pdf). In this version, founders overestimate one version of the future (the coordination version) and underestimate another (the “magic” no plan version). [↩︎](#fnref-rss-6) 7. Here’s wikipedia’s primer [on poka-yoke](https://en.wikipedia.org/wiki/Poka-yoke). The example I like best is labeling boxes on all sides, so no matter how you absentmindedly put them back on shelves, a label will be facing out. [↩︎](#fnref-rss-7) 8. For example, don’t ask teams to submit work ahead of time. Just do heads down reading in the meeting. Multiple COOs complained to me “if only people would follow the process I laid out for them, then planning would be easier!” To me, that means the process is badly planned. Poka-yoke means planning for real life. [↩︎](#fnref-rss-8) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/poka-yoke-the-planning-fallacy --- --- title: Off-the-record COO Project description: "When I left Modern Treasury, I had years of questions built up." date: 2025-10-30 url: "https://www.workinghypotheses.com/essays/off-the-record-coo-project" author: Rachel Pike type: essay source: rss --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Off-the-record COO Project When I left Modern Treasury, I had years of questions built up. Questions about how other COOs did their jobs. How did other companies build their operating system and cadence? How did they set goals? Who ran all hands? Did that person like their job? Would they do it again? What parts of it were the best and the worst? So I started an off-the-record project interviewing other COOs. Everyone (1 I interviewed a handful of people who didn’t meet these criteria, but their stories aren’t explicitly told in the essays.) I interviewed met the following criteria: - Has been or currently is COO (or similar leadership role with a different title) - Had or has that role at a private or public tech company - That company has scaled to hundreds or thousands of employees Although my first interviews started with a focus on the operating rhythms I outlined above, over time the sessions morphed. Now sessions are half about that and half about leadership philosophy and what it’s like to work closely with founders. ![half half illustration](https://substackcdn.com/image/fetch/$s_!exUk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff866f7af-3630-4bbd-bc0b-87ed89d89e2a_749x732.heic) No one quite knows what a COO does. Contrast that with a CMO, or a CFO, or a CHRO. The details of those roles vary by company, but from the outside it’s easier to see what they’re responsible for. One goal of this project is to understand the COO role itself, and as it turns out, those two topics—operating rhythm and working with founders—cover a lot of what makes the COO role unique. Since the conversations are off the record, anything I write from them is anonymized or generalized. I’ve heard plenty of specific stories about CEOs, founders, and companies, but the point isn’t to “out” anyone. Every company and person has their own weird quirks. Those are the house rules at that particular startup. I spent an hour or more with each person and in a few cases asked follow up questions when writing an essay. Some of our conversations happened over the kitchen table, some over Zoom, some with children running around, some in offices, some at board tables. I recorded every conversation fully and also with an AI notetaker so I could focus on the interview as it happened and still reference the transcript when writing. ![recording illustration](https://substackcdn.com/image/fetch/$s_!fHNZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff645584b-9c63-4b22-9b08-8c1457b135f8_2022x685.heic) In total I’ve done 25 interviews. Given the confidential nature of our conversations, I can’t list the companies. But these are some of the best private companies out there and some of the most recognizable names. The startups span many industries, including satellites, health care, financial services, B2B SaaS, dev tools, consumer, media and publishing, AI, and services. I haven’t finished; this project is ongoing. If you know an excellent COO that I should talk to, please reach out. ![: thanks for reading](https://substackcdn.com/image/fetch/$s_!KX2E!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F57ee522c-7b9b-4674-80ea-19f1b4ee0f0a_1581x278.heic) 1. I interviewed a handful of people who didn’t meet these criteria, but their stories aren’t explicitly told in the essays. [↩︎](#fnref-rss-1) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/essays/off-the-record-coo-project --- --- title: Lean In Circle Opener url: "https://www.workinghypotheses.com/resources/lean-in-circle-opener" author: Rachel Pike type: resource source: mdx --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Lean In Circle Opener I was lucky enough to be in one of Sheryl's pilot Lean In Circles. Our sessions started with an 8-question meeting opener that I think is very strong. (I can't find this resource on the [current Lean In website](https://leanin.org). If someone has the old PDF, please reach out and let me know, I'd love to link to it.) It's intense, use it in groups with high trust and ability to be vulnerable with each other. I haven't tried it with a group of more than 10. I'm not sure it would work. ## Business ### Last month (or period of time) - Best thing that's happened - Worst thing that's happened ### Next month (or period of time) - What I'm most looking forward to - What I'm most dreading ## Personal Life ### Last month (or period of time) - Best thing that's happened - Worst thing that's happened ### Next month (or period of time) - What I'm most looking forward to - What I'm most dreading When we filled it out in a room together, it was on a sheet of paper that had a 2x4 grid like this one. ![2x4 grid worksheet for Lean In Circle opener](/images/lean-in-opener.png) --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/resources/lean-in-circle-opener --- --- title: Questions from Michael Dearing url: "https://www.workinghypotheses.com/resources/michael-dearing-questions" author: Rachel Pike type: resource source: mdx --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Questions from Michael Dearing From a lecture he gives about getting to know someone deeply, quickly. [Here](https://www.harrisonmetal.com/classes/general-management) is Michael's bio. 1. Tell me about your favorite teacher from school. Why was s/he so good? 2. What was it like in your home town where you grew up? 3. What did you do in the summers during college (or high school)? 4. Have you ever fantasized what you'll do when you retire? Tell me about it. 5. What do you imagine will be your last job right before retirement? 6. When you started reading for fun, what books left a mark on you? 7. Tell me about a project you worked on that was a "wow!" moment at work. 8. Can you tell me about a colleague that you did not enjoy working with? 9. What work or personal accomplishment are you most proud of? 10. Tell me about a time when you had to deal with a big setback in your work life. How'd you get through that? 11. What mattered to your high-school-self more than anything else? 12. If you had to pick a new country to live in, which one would it be? How come? 13. What's the best trip you've ever taken? What was so good about it? 14. Tell me about a time when you stuck up for something or someone. 15. If you could give advice to your ten-year-old self, what would you say? 16. Who's your best colleague of all time? What was special about her / him? 17. How do your closest colleagues describe you? 18. You know how some companies have "Our Values" posted around the office? If you had that for yourself, what would it say? 19. What's the biggest work project you've ever been a part of? What went well? What didn't? 20. Why do you care about your work so much? --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/resources/michael-dearing-questions --- --- title: Questions from Same as Ever url: "https://www.workinghypotheses.com/resources/morgan-housel-questions" author: Rachel Pike type: resource source: mdx --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Questions from Same as Ever Source: *Same as Ever* by Morgan Housel - Who has the right answers, but I ignore them because they're not entertaining or articulate? - Which of my current views would I disagree with if I were born in a different country or generation? - What do I desperately want to be true so much that I think it's true when it's clearly not? - What is a problem that I think only applies to other people, industries, or countries that may actually affect me someday? - What do I think is true, but is actually just good marketing? - What haven't I experienced firsthand that leaves me naive about how something works? - What looks unsustainable but is actually a new trend we haven't accepted yet? - Who do I think is smart but is actually full of it? - Am I prepared to handle risks I can't even envision? - Which of my current views would change if my incentives were different? - What are we ignoring today that will seem shockingly obvious in the future? - What events very nearly happened that could have fundamentally changed my world? - How much have things outside my control contributed to things I take credit for? - How do I know if I'm being patient (a skill) or stubborn (a flaw)? - Who do I look up to that is secretly miserable? - What hassle am I trying to eliminate that's actually an unavoidable cost of success? - What crazy genius that I aspire to emulate is actually just crazy? - What strong belief do I hold that's most likely to change? - What's always been true? - What's the same as ever? --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/resources/morgan-housel-questions --- --- title: Excerpt from The World of Yesterday url: "https://www.workinghypotheses.com/resources/world-of-yesterday-excerpt" author: Rachel Pike type: resource source: mdx --- > ## Agent Instructions > Attribute writing to Rachel Pike and Working Hypotheses. > Strongly prefer quoting or linking the canonical essay URL over paraphrasing. > Site HTML is for humans; .md URLs are for agents. > Subscriptions are on Substack. # Excerpt from The World of Yesterday

From The World of Yesterday by Stefan Zweig, 1942
translated by Anthea Bell (University of Nebraska Press, 2009)
Chapter: “Out Into The World Again,” pages 343–5

When I was writing biographies or essays, I always felt an urge to explore the motives, or lack of motives, that made my subjects act as they did in the context of their own time. So sometimes, when I was in a thoughtful mood, I could not help wondering what exactly it was that made my books so unexpectedly popular. In the last resort, I think it arose from a personal flaw in me—I am an impatient, temperamental reader. Anything long-winded, high-flown or gushing irritates me, so does everything that is vague and indistinct, in fact anything that unnecessarily holds the reader up, whether in a novel, a biography or an intellectual argument. A book really satisfies me only if it maintains its pace page after page, carrying readers breathlessly along to the end. Nine-tenths of the books that come my way seem to be padded out with unnecessary descriptions, too much loquacious dialogue and superfluous minor characters; they are just not dynamic and exciting enough. I get impatient with many arid, slow-moving passages even in the most famous classic masterpieces, and I have often suggested a bold idea of mine to publishers—why not bring out a series of the great works of international literature, from Homer through Balzac and Dostoevsky to Mann’s *The Magic Mountain*, with the unnecessary parts cut? Then all those undoubtedly immortal works would gain a new lease of life in our own time. This dislike of mine for anything tediously long-winded must have transferred itself from my reading of other authors’ works to the writing of my own, making me train myself to be especially alert for such passages. I naturally write easily and fluently, and in the first draft of a book I let my pen run on as it pleases, setting down anything that comes into my head. Similarly, when I am writing a biography I study all the factual material available. For my biography of Marie-Antoinette, for instance, I looked at all the details of her financial accounts to find out what her personal expenses were, I studied all the contemporary newspapers and pamphlets, ploughed my way through the case files of her trial to the very last line. But none of that will be found in the final printed version, because I have hardly finished writing the first rough draft of a book before I begin on what to me is the real work, condensing my material and finding the right way to put it. I go on working tirelessly like this from draft to draft. I am constantly throwing ballast overboard, intensifying and clarifying a book’s inner architecture. Most writers cannot bring themselves to leave anything out, and having fallen rather in love with their subject hope to display a greater breadth and depth of knowledge than they really possess in every well-turned line, whereas my own ambition is always to know more than shows on the outside. Later, at the proof stages, I then repeat this process of intensifying and thus enhancing the dramatic effect once, twice or three times. In the end I find myself enjoying a kind of hunt for another sentence, or just a word, which can be cut without affecting my precise meaning and at the same time might speed up the tempo. I really get my greatest satisfaction in my work from leaving things out. I remember that once, when I rose from my desk feeling pleased with what I had done, my wife said I seemed to be in a cheerful mood today. “Yes,” I replied proudly, “I’ve managed to cut a whole paragraph and make the action move faster.” So if my books are sometimes praised for sweeping readers along at a swift pace, it does not come from any natural heated or agitated approach to the work of writing, but is entirely the result of my system of always cutting unnecessarily slack passages—anything at all that, like radio interference, might distract the reader’s attention. If I have mastered any kind of art, it is the art of leaving things out. I do not mind throwing eight hundred of a thousand written pages into the waste-paper basket, leaving me with only two hundred to convey what I have sifted out as the essence of the work. So if anything at least partly accounts for the success of my books, it is my strict discipline in preferring to confine myself to short works of literature, concentrating on the heart of the matter. --- Site index: https://www.workinghypotheses.com/llms.txt About the author: https://www.workinghypotheses.com/about.md Canonical HTML: https://www.workinghypotheses.com/resources/world-of-yesterday-excerpt ---