Strategy execution starts with naming who decides: Nathaniel Manning of Kettle

The most useful artifact in this episode is a spreadsheet. Nathaniel Manning has built one at every company he has worked at, it runs to four or five hundred lines, and its entire purpose is answering one question: whose decision is this. It is the most concrete piece of strategy execution machinery anyone has described on this show.

Manning is co-founder and COO of Kettle, a reinsurance company using deep learning to price catastrophic climate risk. Before that he founded and led companies including BRCK, the largest provider of public Wi-Fi in Africa, and Ushahidi, the open source platform for crisis response. He worked on clean energy at the Clinton Climate Initiative, and was part of the first class of Presidential Innovation Fellows, serving as special advisor on open data at USAID and later as its chief data officer.

Three months on a cushion

Asked where the drive comes from, Manning traces it further back than most people would.

He studied philosophy and religious studies at university, focusing on Eastern traditions, and spent three months living in a Zen monastery in Japan. He describes himself as influenced by Buddhism, and the thing that stuck was simple. The point is to reduce suffering. Everything since has been a version of asking how to do more of that.

The practical consequence is that he finds it hard to get up in the morning for work that is only about money. He says he gets bored.

Reinsurance, explained without jargon

Manning's description of his industry is the clearest I have heard.

Reinsurance is the safety net below the safety net. The second parachute. It is insurance for insurance companies, covering catastrophic risk.

The logic follows from the household version. You buy home insurance because it makes no sense to keep the value of your house sitting in a bank account. Your insurer covers thousands of things that could go wrong with it. But it makes no sense for that insurer to hold the value of every house it covers against the possibility that a thousand of them burn in one wildfire, or tens of thousands are hit by one hurricane. So they buy a policy too, from the reinsurance industry, against what the industry calls tail risk. And then reinsurers buy insurance of their own, which is where it gets interesting.

The number that explains why this is now a growth industry: there has been a threefold increase in billion dollar catastrophes over the last fifteen years due to climate change.

The long tail of a crisis

Manning's route into insurance came through disaster work, and the insight is worth sitting with.

He spent the first three or four years of his career trying to reduce carbon emissions, having concluded after An Inconvenient Truth and some environmental studies classes that climate was his generation's problem. Then came a shift in outlook. It felt to him like the species had failed at prevention, which meant people were going to get hurt, so he moved to crisis response software.

Doing that work, he kept meeting insurance companies, who turned out to be the parties most interested in crisis response data. The same thing happened in government, where the humanitarian datasets he was opening at USAID attracted insurers and reinsurers.

Then came the observation. If you graph suffering over time in a disaster, there is a sharp spike at the beginning. Triage, hospitals, food and water, donations, media attention, roughly two weeks of it. Then the coverage stops and the spike tails off, and the immediate medical need has largely been met.

What remains is a very long tail of people rebuilding their lives. And on that stretch there is no media and no donation surge. There are people, their communities, and their insurance companies.

His thought was that the area under the long tail might be as large as the area under the spike. He had been building software for the spike.

The second thing that drew him in was what insurance actually is underneath the frustrating customer experience. Everyone pools their money and agrees that if your house is the one that burns, everyone else contributes a fraction of a percent toward rebuilding it, on the understanding that you would do the same next year. He calls it a beautiful arrangement, and notes that everything in the middle, working out what each person should pay, is a data science problem.

Putting your own money behind your model

Kettle does not sell software to insurers, and Manning's explanation of why is a good argument for anyone weighing a SaaS business against a principal one.

The company operates more like a dealmaker, transferring risk to whoever is best positioned to hold it. His analogy for a startup audience is a venture firm. A VC puts in a general partner commitment, but ninety percent or more of what it invests belongs to someone else. Kettle's capital comes from endowments, family offices and pension funds looking for diversification, and what it offers them is a return uncorrelated with equity markets, since a stock market crash has no bearing on how many wildfires occur.

Kettle puts its own capital in alongside, which aligns the incentives.

His objection to the software model has two parts. The market is smaller. And more pointedly: if you genuinely built a model that predicted gold prices, why would you license it to hedge funds for ten thousand dollars a month rather than trading on it yourself. Kettle's version is putting its money where its models are and taking the losses alongside its investors when it is wrong.

Which makes accuracy the entire competitive advantage. Being more right than the competition allows better returns and lower prices, and insurance is a category where, all else equal, people buy the cheapest option.

How you test a model when being wrong is expensive

Michael asks the obvious question about iterating toward an MVP when the stakes are catastrophic risk. Manning's answer is a staged one.

They built the product plus a visual layer so the output could be communicated, then ran real portfolios of homes through it to calculate what losses would have been.

The wildfire statistic underneath the whole business is remarkable. California sees tens of thousands of wildfires a year, and fourteen of them cause 98% of the damage. As Manning points out, humans are generally good at putting fires out, and most never become the ones on the news. The question is which ones will, and where.

In their first year of testing, Kettle's model placed all fourteen of the largest fires inside the riskiest 20% of the state. Their underwriting guidelines exclude the top 25% by risk, so they would not have been exposed to any of them.

The following year came real quotes, some of which were bound and went into production with real customers and real money. That produced the evidence that matters: actual returns for investors, actual losses, actual accuracy.

His conclusion about credibility in this industry is unglamorous. You have to keep being right, and it takes years of taking people's money and returning it before anyone concludes you have it. In the meantime, the strategy is transparency about how the risk is being assessed.

The COO as negative space

Manning has one of the better descriptions of the role I have collected.

He thinks about a team as a puzzle, where each new person's skills form a shape. The COO, in his framing, is the piece that fits the negative space left by the rest of the executive team. Another way of saying executive jack of all trades, which he thinks is accurate and which suits how his brain works.

The practical consequence is that the job keeps handing pieces away. He did all the finance work early on, and Kettle now has a CFO who was CFO of a public reinsurance company, whose models he cheerfully notes look considerably better than the ones he built in the first few years. He wrote the regulatory policy to get the company's managing general agent approved in Bermuda, and now that work sits with people who have done it many times.

What has stayed with him is four things: telling the story, fundraising, decision-making, and building the systems that let the team scale.

On that last one he points to a piece called Old Guard, New Guard, which describes the tension he lives with. You have the energy of the founding days, and then you hire people who were the CFO of a public company or a senior staff engineer at Google for a decade, arriving with a view about how it should be done. His framing for the resolution is that they are building the plane while flying it, so the answer is to implement the new thing together rather than either defer to it or dismiss it.

The Responsibilities Decision-Making Doc

This is the part to steal.

Some decisions at Kettle are formally gated. To break their own underwriting guidelines requires the CEO, the chief underwriting officer and the executive committee chairman, between them holding decades of underwriting experience, to all agree. That is deliberate friction around the one thing that could sink the company.

Everything else lives in a spreadsheet. Manning's chief product officer gave him the cleanest description of what it is for: when someone asks whose job is that, or whose decision is that, the answer should be written down explicitly. It is equally an answer to whose fault it is when something does not happen.

His example is payroll. Payroll going out on time has his name next to it. Most of the time the process runs and he does not think about it. If it failed, it would be his fault.

Four hundred to five hundred lines, one name against each.

Three ideas hold it together.

Evolution by design. Every quarter, every team revisits the document and updates it. Someone can say they no longer want to own the company's Twitter account and ask who will take it. A team can say too much has accumulated and they need to hire. Empty lines get surfaced as work nobody owns. Crucially, these are not tasks and cannot be crossed off, because each one is something that has to keep happening or a decision that has to keep being made.

Structure without bureaucracy. People hate bureaucracy and genuinely want to know what they are responsible for, what makes a difference, and where to focus. The document delivers the second without the first.

Making the implicit explicit. Manning's observation is that problems tend to originate in something that was implicit, got quietly built in, and was never written down.

He also makes a sharp point about job descriptions. Companies write good ones, hire against them, and never look at them again. The valuable part of a job description is not the title, it is the bullet points, and every bullet point across the company should be in this document, organized by team, showing how the responsibilities interact.

His own decision-making sits on top of it. He likes assembling a council of experts and equals, taking input widely and hearing real disagreement, while it remains unambiguous whose call it is. Hiring might be his decision, with required input from six named people.

Asked whether he would share the templates, he agreed immediately, describing himself as an open source fanatic and the artifact as an Excel spreadsheet rather than rocket science.

Founder market fit, and the product he walked away from

Manning's advice for finding work that matters starts somewhere unexpected: get quiet.

His view is that our brains are active and will logically talk us into things, so the first step is quiet enough to hear what is actually there. Walks in the woods for him, swimming for his wife, meditation for both. Then whiteboard it, list it, and see what came out of the quiet.

The test he applies afterward came from Ryan Delk, Kettle's first investor and founder of Primer, who spent years going through bad decks and bad ideas with him. Delk's framing was founder market fit, or founder product market fit, and his read was that Manning obviously had it for this problem.

The story that proves the concept is about a company Manning did not start. Before Kettle he wanted to build an insurance product, so he ran hundreds of interviews asking people what they had spent a lot of money on unexpectedly in recent years. He asked across two communities, one international and one made up of American professional colleagues around his own age.

The answer was fertility treatment. IVF. And it checked out as a real gap, because meaningful coverage was mostly available only to people working at a handful of large technology companies.

Then came the check. His assessment was that he was the wrong person to sell it. As he puts it, he is a dad and a guy, and should not be the one taking that product to market, however clearly the data said it should exist. Underneath that was a longer arc: climate work had been driving him since he was twenty, and he did not want to wake up to this other problem five or six years later.

Leading a team through a crisis that becomes public

Michael's standing question produced the most serious answer the show has had.

While Manning was COO at Ushahidi, someone brought a harassment claim against the CEO. Neither he nor the board had faced anything like it. The organization was based in Kenya, and this was before MeToo in the United States. The CEO was, in Manning's words, way out of line, and was let go.

The sequence was brutal. The legal process had to run and the board had to hold a hearing, which took time, during which nothing could be said publicly. Manning stepped in to lead. Then the story leaked to the press while the legal process was still concluding. Throughout, it was hardest on the person who had been harmed.

He became the person talking to press and writing the public statements. But the part he treats as most important is internal. The team was being dragged through the mud without knowing what had actually happened, and his read is that this was an incident and a person rather than a culture.

So he talked to every funder, every customer, and every single person on the team, in what he describes as an intense listening tour. About a third of the team was in Kenya and he flew there. They held team circles designed for people to say what they needed to say, deliberately without back and forth.

What he names as the thing he is proudest of in his career is that nobody quit during that period, and no funders left either. He is careful to add that things could have been handled better, and then he took over the organization and started rebuilding.

His lessons on crisis leadership are two. First, acknowledgement, which he contrasts with the teenage instinct to hide something and hope. He borrows a line from a teammate that everything important is learned by kindergarten: treat each other kindly and with respect. The techniques he reaches for come from couples therapy, listening rather than defending, nonviolent communication.

Second, culture and history as an anchor. Reminding people who they are, and that a terrible incident does not get to define everyone or be carried by everyone.

The 5 things I took away from this conversation

1. Write down who decides, for everything. The Responsibilities Decision-Making Doc is the single most portable idea on this show so far. Not a task list, a list of standing responsibilities and decisions with exactly one name against each. Most organizational dysfunction I have seen is an unwritten line in that document.

2. Review ownership quarterly, on purpose. Evolution by design is what stops the document becoming a fossil. It gives people a sanctioned way to hand something off, surfaces work nobody owns, and turns headcount arguments into an observable fact rather than a complaint.

3. If your model is that good, back it with your own money. Manning's gold price argument cuts through a lot of positioning. Selling predictions as software caps your upside and quietly signals how much you believe them. It is a question worth asking any company whose product is a model.

4. Founder market fit can override the data. He validated a real, underserved insurance need and walked away because he was the wrong person to carry it. That is a discipline most founders lack, and the reasoning, whether you want to wake up to this problem in five years, is the right test.

5. In a crisis, the listening tour is the work. Talking to every funder, every customer and every single employee, and holding sessions designed for people to be heard rather than answered, is what kept an organization together through something that should have destroyed it. Nobody quitting is a remarkable outcome.

FAQ

What is a practical strategy execution framework for a startup? Manning's is a single document listing every standing responsibility and decision in the company, with one owner named against each, organized by team and reviewed quarterly. It sits alongside project tooling rather than replacing it, because its entries are not tasks that get completed but responsibilities that persist.

How is that different from a project management tool? Task tools track work that finishes. This document tracks things that have to keep happening and decisions that have to keep being made, so nothing on it can be crossed off. Manning also treats it as the living version of every job description in the company, since the useful part of a job description is the bullet points, which are otherwise never read again.

What are the COO role responsibilities at an early stage company? Manning describes the COO as the puzzle piece filling the negative space around the rest of the executive team, which means the specific duties change as specialists are hired. What has stayed constant for him is telling the company's story, fundraising, decision-making, and building the systems that let the team scale.

What is reinsurance and why does climate change matter to it? Reinsurance is insurance for insurance companies, covering catastrophic or tail risk. Primary insurers cannot hold enough capital against the possibility of thousands of homes being destroyed at once, so they transfer that risk on. Manning notes billion dollar catastrophes have tripled over fifteen years, which changes both the demand and the pricing problem.

How do you lead a team through a public crisis? Acknowledge it directly rather than managing around it. Manning talked to every funder, customer and employee individually, flew to where a third of his team was based, and ran sessions where people could speak without being argued with. He also used the organization's culture and history as an anchor, so that one incident did not become everyone's identity.

Also mentioned

• Kettle, and the wildfire modeling behind its underwriting guidelines
• Ushahidi, the crisis response platform, and BRCK, the African connectivity company
• The Presidential Innovation Fellows program and open data work at USAID
• The Clinton Foundation climate work that started Manning's career on emissions
• Primer and its founder Ryan Delk, Kettle's first investor and the source of the founder market fit test
• Nonviolent communication, the approach Manning drew on during the Ushahidi crisis
• Nathaniel Manning on Twitter at @natpmanning, and Kettle at @ourkettle

Listen to the full episode

Nathaniel Manning on Between Two COO's

Between Two COO's is hosted by Michael Koenig. Subscribe on Apple Podcasts, Spotify, or wherever you listen.