The Position

We built De Novo to make whey protein without a cow. A year in, we walked away from it. The science was working. The spreadsheets were the problem.

By then, we’d committed about $2 million to whey and spent $300,000 of it. Twelve months of fermentation runs and customer calls and spreadsheets, all aimed at a protein we were about to abandon.

Walking away meant admitting the protein we’d built the company around was the wrong one, and betting instead on one we couldn’t reliably make yet.

The Model

What saved us is what Shane Parrish calls second-order thinking.

First-order thinking asks one question. What happens if I do this? It’s fast, and it’s usually right about the immediate effect. Everyone else asks the same question and lands on the same answer, so by the time you act on a first-order conclusion, you’re already standing in a crowd.

Second-order thinking asks the next one. And then what? What does the world look like after the obvious move plays out, once everyone who saw what you saw has moved on to it, and the thing you’re excited about today is something a dozen funded teams are all shipping at once?

It’s uncomfortable on purpose because it slows you down when momentum wants you to commit. We chose whey without asking. The question only arrived a year later, once the numbers we’d been building started pointing the other way.

Field Tested

Whey made sense the way obvious bets do. The market was enormous and already proven, and the science sat well within what our team could do. Competitors were popping up everywhere for the same reasons.

The proven part is what eventually undid it.

Whey is a waste product of cheesemaking. Dairies already make it by the ton as a side effect, and it trades at $10 to $13 per kilogram. Our one real edge over them was that ours would be animal-free and sustainable, so we took that to customers and asked what it was worth. The answer was nothing extra. They cared about price, and they weren’t going to pay a premium for a greener version of something already cheap.

Then scale. Whey is measured in grams, and the world goes through about 1.5 million tons a year. Producing that much protein would take bioreactors of 300,000 litres and up, the giant tanks the fermentation runs in, and there aren't many that big.

The model didn’t need much more analysis after that. When a path is that easy and that obvious, the margin is usually gone before you reach it.

So we asked the same question about the protein nobody was racing towards. Lactoferrin isn’t a by-product of anything. It’s is a rare bioactive protein found in tiny amounts in milk that binds and transports iron in the body; the world produces about 800 tons a year, and demand sits well ahead of that.

It’s dosed by the milligram, so the volumes fit the 100,000-litre bioreactors that already exist everywhere. It sold for $650 a kilo, and our customers wanted more of it at a price and consistency they could plan around, which was exactly what precision fermentation was built to deliver.

The catch was the science. Lactoferrin is a larger, more complex protein than whey, with functional properties that host organisms find hard to replicate. It was a much harder problem than the one we’d planned for.

One of our technical founders understood this better than I did, and she was hesitant. She told me, “We are going to need significantly more experienced talent to pull this off.” She was right. Choosing it meant admitting that the people we had, me included, weren’t yet enough to build it.

That difficulty was the whole reason it was worth doing. A protein this hard to make is one nobody can flood the market with the moment they notice the price, and the science alone keeps most competitors at bay for years.

Supply eventually increases, and the price comes down with it. That was always going to happen. The bet was on the cost beneath the price: precision fermentation keeps improving, so as we lift our yield, our cost per kilo keeps falling. We ride the price down and keep the margin. The thing that scared us was the thing that would protect us.

Staying on whey was the safe call. It was working, the team could do it, and a year of money and relationships already pointed that way. The case for holding course was strong, and for a while it prevailed.

Eventually, we made the call. A year into being a whey company, we changed what business we were in. We caught it after $300,000, with most of the two million still unspent.

What I still sit with is why it took a year to run that question on the one bet the whole company stood on.

The Action

Ask the same question this week on the one bet you’ve stopped letting yourself question. It doesn’t have to be a product. Second-order thinking works on anything you’ve committed to and quietly stopped examining: a channel, a market, a hire, the way you’ve decided to grow. Pick the one with real money and months in it, the one you’d be embarrassed to reverse because you’ve told people it’s the plan.

Then ask the next question. When this works the way you’re betting it will, and everyone who can see what you see has moved on it, what are you actually competing on? Can someone else reach the same place more easily than you, because they’re built for it and you’re not? Write down the honest answer, not the comfortable one.

If it holds, commit harder than before. If it doesn’t, you’ve found out now, while it’s cost you a few months instead of the company.

The Bearing

One edition to go. Next week, I’ll recap the season in one place, because scattered frameworks won’t drive action.

Second-order thinking supports all the others; it’s what ensures we use the right framework at the right time. I’ll share how they fit together.

One question. One week. Then we go again.

— Jean

The Position

We once spent months at De Novo running experiments that gave us almost nothing to act on. The team had worked hard, the data was all there, but none of it forced a decision.

The strange part is how much we do measure.

Ask me how the science is doing at De Novo, and I can answer in seconds. Production yield, protein purity, and fermentation cycle time. I can give you those numbers without looking them up.

It is the same for you with your product. Build status, the spec, what shipped, the bug list, and the burn. You could recite all of it right now without opening anything.

Ask either of us what happened last week that decides next quarter, and the answer slows down.

The product numbers feel like progress because they move where you can see them. But most of what you are watching is a read-out of decisions you already made. By the time a number like revenue or yield confirms something worked, the work that produced it is weeks behind you, and you cannot reach back and change it.

The numbers worth having are the ones that tell you early enough to still change the week.

The Model

Andy Grove, who built and ran Intel, framed this well: the difference between a lagging indicator and a leading one.

A lagging indicator reports a result. Revenue, profit, units out the door, protein yield at the end of a run. Each one is true, and each one is already history by the time you read it. The inputs that produced it were set in motion weeks or months earlier. You can study the number, but you cannot change it.

A leading indicator measures the activity upstream of the result, while that activity is still in your hands. Qualified buyer conversations move before pipeline, and pipeline moves before revenue. An experiment built around a clear hypothesis moves before months of lab work. Release cycles move before product quality.

Grove’s instinct from the factory floor was plain: you do not wait for the finished batch to fail quality control to learn something went wrong. You measure the inputs and the steps, because a problem caught at step two costs you an afternoon, and the same problem caught at the end costs you the batch.

The lesson here is: if the feedback loop is long, the control point has to move upstream.

Field Tested

This goes back to the early days of De Novo, in a small and fairly depressing office in Cape Town.

The other founders and I sat down to go through the experiments the lab had run and the data it had produced. One of our technical co-founders took us through them, one at a time, results and all.

Something to keep in mind about a lab experiment in precision fermentation: it is slow. From running it to knowing whether what you did worked is three to four months, and that is just the lab and pilot stage, before any scaling into the larger fermenters. Every experiment on that table was a season of someone’s work.

We went down the list. Experiment after experiment, the answer landed in the same place. We had plenty of data. None of it added up to an answer. Science is messy, and results are rarely clean. Anyone who has worked near a lab knows that. But this was past messy. The data was not pointing anywhere. We had done the work and could not say what the work had told us. The line that kept coming up that day, more than once, was “it’s difficult to say.”

By the fourth or fifth experiment, the room had gone quiet. We had walked in to read the data. We were now stuck on a worse question: whether these experiments had been designed to tell us anything in the first place.

The cause was not the science team. It was how the experiments had been set up. Each one had been run as a standalone shot, with no real thought for the one before it or the one after. Nobody had sat down first and asked how this set of experiments connects, what each one has to prove so the next is worth running. We had run the project one experiment at a time instead of as a project.

The cost was not really the reagents, the equipment time or the salaries. All of that is real, and all of it is small next to the thing that actually hurt. Time.

Every inconclusive run pushed the date we could put samples in front of customers and start closing deals out by weeks, sometimes months. In a biotech, science and commercial are bolted together. A lost season in the lab is a lost season on the commercial timeline. And every month De Novo runs costs tens of thousands of dollars, whether the experiments told us anything or not.

Once we named it, the fix was not complicated. Every experiment now opened with a written hypothesis and one explicit question: what does this run need to prove so the next decision is worth making? Each one mapped against the whole project sequence rather than fired off on its own. SOPs, done properly. Nobody pushed back. It was obvious to everyone the moment we named it, and the newer and younger scientists needed that scaffolding most.

One thing I want to be clear about, because the wrong lesson is easy to take here. This is not “science must get everything right and stop taking risks.” Experiments fail, and messy data is part of the job. Because missteps in the lab are a certainty, and because they land straight on your commercial timeline, you build them into the plan and the strategy on purpose. You schedule for the misses, because the misses are coming.

I realized later that the same mistake was sitting on the commercial side of the company. We were talking about hiring someone to run business development before we had built the machine that person would run. No defined process, and no weekly signal to tell us whether a given week had moved us closer to revenue or only produced activity. Same mistake, different room.

The Action

This week, before you build anything, sort the numbers you already have into two piles.

Take everything you currently track and go down the list. Is this a result, something already decided by work that is behind me? Or is it an activity, something still in my hands over the next seven days? Most early-stage dashboards, where they exist at all, turn out to be almost entirely the first pile. Doing that sort honestly is worth the hour on its own.

Then pick the two outcomes that actually decide whether the company is still standing in a year. At this stage, it is usually some version of pipeline and product, or pipeline and cash.

For each one, name the activity that sits upstream of it. The work that, done well and in enough volume, makes the outcome more likely in eight weeks. If the outcome is pipeline, the upstream activity is probably qualified conversations and next steps booked. If it is product, it is probably something like experiments closed out cleanly, or release cycles finished. You know your business better than I do. Name the real one, not the convenient one.

Put a number on each. Five qualified conversations a week. Three experiments documented end-to-end. Whatever the honest target is.

Then look at those numbers every week, on the same day, and treat a miss as this week’s problem rather than next quarter’s footnote. The entire value of a leading indicator is the window it gives you to act before the result is set. Review it monthly, and you have thrown that window away and turned it back into a lagging indicator with extra steps.

Run it for a month. If your leading numbers look healthy and the outcomes still are not moving, you picked the wrong indicators. That is useful. Go find better ones. If your leading numbers are weak, the market has not failed you yet. You are looking at the reason the outcomes are quiet, early enough to still do something about it.

The Bearing

There is one trap in all of this.

A leading indicator is only useful if it is actually connected to the outcome. Track the wrong upstream activity with enough discipline, and you can hit every weekly target while the business quietly goes nowhere.

Choosing what to measure is a prediction. You are saying: I believe this, done now, causes that, later. Get the prediction wrong, and the dashboard is lying to you.

That is next week’s problem. Learning to think a move or two ahead, so the number you commit to chasing is one genuinely wired to the outcome you want.

— Jean

The Position

Year three of building De Novo. Two B2B conversations a week.

Five operational meetings every week on top of that, for the consumer product we’d added that year, to show traction faster than the B2B sales cycle was going to deliver it.

That was our weekly calendar for four months. Essentially, two companies’ worth of completely different meetings inside one early-stage team.

We told ourselves we were covering both bases, but in reality, we were running B2B at half pace because half of every week was being spent operating something else.

That realisation came from looking at my calendar, the work that was actually the company (B2B ingredients) was getting the leftover hours.

The B2C product was getting the meetings because it felt more urgent and tangible.

The thing we’d built De Novo to do was getting managed in the gaps between the things we’d convinced ourselves to do as well.

The Model

Matt Lerner’s term for what we hadn’t yet found is channel-market fit. He spent years running growth at PayPal before founding SYSTM, his advisory practice for early-stage startups. It is one of the most underused ideas in early-stage growth.

For every business, there is a specific channel where the audience, the format, and your natural strengths line up cleanly enough that the channel starts compounding without being forced. The job is to identify that channel, commit to it before you have a signal, and stay long enough to find out whether the signal exists.

Three indicators tell you whether you’re in the right one.

First, the audience is concentrated there in the right shape. Specifically, your customer, in numbers worth working for, in the role that does the buying or the influencing.

Second, the format plays to how you actually communicate. A founder who thinks in technical frameworks writes well on LinkedIn, and a founder with a strong personal story works on video (I’m terrible on video). Picking the wrong format means trying to be someone you’re not, for weeks on end, at scale.

Third, your content or outreach performs disproportionately well there compared to anywhere else. Noticeably better, often before you can fully explain why.

When all three align, go deeper, because adding a second channel in month two slows both down. You spread the same hours across more surfaces, and the first channel that was about to compound just stops getting the time it needs.

A channel finding its compounding curve at month three needs you to keep showing up. The sibling channel you’d start in month one is a way to look like you’re doing something while you wait.

I worked at AP Moller Maersk for five years before founding anything, and I was there when a similar decision was made and implemented at a corporate scale. Maersk was the world's largest container shipping line by volume and chose to stop competing for that title.

The shift was to serve fewer, higher-margin customers more deeply rather than chase every container shipper at the same level of attention. From the inside, you could watch the trade-off play out in real time.

We moved fewer containers but made more money on each one, with better service for the customers we kept. Depth over breadth was a deliberate corporate choice.

Maersk had every option available and picked the narrower one.

Doubling down on what works applies the same way at thirty billion dollars in revenue as it does at a three-person team.

Field Tested

Edition 4 was about whether to kill the B2C product at De Novo. This edition is about what killing it unlocked.

The kill is Edition 4’s story. What I want to walk you through this week is the period after, and the specific channel decision we made next that I now think mattered more than the kill itself.

Before the kill, our cadence was two B2B conversations and four or five B2C operational meetings per week. After the kill, the B2B conversations went from two a week to five. Same team, no new salespeople, just attention that had been somewhere else.

Five B2B conversations a week was better than two. Even at five, we still didn’t have a distribution channel, but the next decision changed our trajectory.

We sat down and chose where to go deep.

Cold email was the obvious starting point. We had the lists, and we could have built outbound sequences inside a week.

But as a first touch, cold email had the worst job in the funnel.

The buyer had never heard of us. They had no reason to trust the science. They were reading us against every other vendor email that landed in their inbox that morning. Even if they opened, we had thirty seconds to move from stranger to credible.

LinkedIn gave us a different first touch.

The buyer encountered us repeatedly through their own feed before any direct contact.

The technical credibility of the science accumulated over weeks of posts in a way an email subject line never would. By the time anyone reached out, or by the time a direct message landed, they had already done part of their own due diligence on us. The first message from a new buyer started with “We like what you are doing, let’s work together” instead of “What exactly is it that you do?”

There was also a more practical reason. De Novo’s company page had already been showing pull on lactoferrin content. There was already proof on LinkedIn before we made the channel commitment. Cold email had no equivalent proof yet. As a first move, it would have been a guess. LinkedIn was a known platform we had tested but hadn’t leaned into enough yet.

So I committed. We were going to be on LinkedIn until the channel either compounded or showed us it wouldn’t. Either answer required staying long enough to know.

I systemised the posting so it ran as a public story rather than as ad hoc content. The technical credibility of the science came through on every piece. Lactoferrin specifically got its own consistent thread. The audience was narrow on purpose. Formulators, ingredient buyers and operators within large food and beverage companies who needed to know what we were building before they needed to buy it.

After the JV with Earth First Food Ventures was formed, the channel gained two engines.

I kept posting from my profile on the technical and operator side.

EFFV ran cold DMs on LinkedIn for the commercial side. One channel, two motions, both targeting the same narrow audience.

The first proof that the channel was working came in the DMs. We went from barely any messages a week to three or four a day from operators inside the industry, well before any deal closed. Distributors asking about carrying lactoferrin and formulators asking for samples. The volume was low compared to a consumer brand running ads; however, every conversation was qualified. The room we were in only had the people we wanted to talk to.

Interestingly (though not that interesting), EFFV found us on LinkedIn.

They had been speaking to several lactoferrin producers when they reached out. What came back later, after the JV was signed, was that they had chosen De Novo for the quality of the science and the broad expertise of the founding team. Our science hadn’t changed from the year before, but the room we were posting in had.

After the JV formed, the cadence improved even further. Ten B2B conversations a week, sustained for six months. Ten of those moved into ongoing discussions with documentation, samples sent, NDAs signed, and commercial conversations getting concrete. None of those conversations were inbound when we were running half a week of B2C meetings. All of them came after the segment commitment plus the channel commitment.

What the channel actually was: B2C was killed; B2B was doubled down on, and LinkedIn, chosen deliberately because it was where large corporates already congregated. That, mixed with a founder voice and technical credibility, created a format that played to our strengths instead of forcing us to become a consumer brand.

Each layer reinforced the one below it. None of them would have done the work alone.

The Action

Look at your calendar this week. Pull up the last four weeks of it if you can.

Count the meetings that were directly about the one business you’re trying to win at.

Count the meetings that were about something adjacent to what you’ve also been running. If the ratio is anything less than four to one in favour of the one you’re trying to win at, you have your answer.

Then go a layer deeper. For the business you’re trying to win at, write down the specific people inside specific companies whose attention has to move for you to win this quarter. Be specific. Two named companies, three named roles. If you can’t write that list, the segment isn’t tight enough yet, and the channel question is premature.

Once you have the list, ask one question. Where are these specific people, in these specific roles, already reading and watching while doing the part of their job that involves keeping up with their space?

If the answer is LinkedIn, go to LinkedIn and stay there until the right people start reaching out to you. If the answer is a specific industry publication or conference, write the case study or the talk that lets you show up there. If the answer is a trade Slack or a specialist newsletter, find a way in.

Pick one. Commit to it for ninety days. Do not start a second channel inside those ninety days. If at the end of it the channel has produced nothing, you have a clean answer and you can move to the next candidate.

Ninety days, one channel, no parallel experiments. Whatever the channel tells you at the end of it is the answer worth having.

The Bearing

The first sign LinkedIn was working for us came well before any deal closed. The DMs went from once or twice a week to several a day.

Comments on posts shifted from generic to specific. Operators in the industry started responding to technical posts with questions and frames I hadn’t provided.

LinkedIn was telling us it was working before the JV closed and well before any revenue followed.

That kind of feedback shows up earlier than the P&L. It tells you whether you’re going to win before the win arrives.

Next week: the numbers that tell you the truth about your business before your instincts do.

— Jean

The Position

A Sunday evening at my desk. The contract was in front of me. I was on the fourth read of the same royalty clause that week.

The work was real and exactly that was the trap.

There is a moment in any work that matters where the most damaging thing you can do is keep doing what you have been doing. The work looks indistinguishable from the work you actually need to do. That is what makes the trap invisible.

You are probably in some version of it right now. I am, on a different week, with a different clause to read for the fourth time.

The Model

The framework I keep coming back to for this trap is Tom Bilyeu’s. He built Quest Nutrition into a billion dollar company before founding Impact Theory, and his thesis on why capable people fail to execute is a single sentence.

You will never consistently execute at a level that conflicts with how you see yourself.

If you see yourself as someone who prepares before they act, you will find ways to keep preparing. The preparation will look like what someone preparing looks like. It is the system working as designed.

The trap fires hardest in one specific situation. When you imagine your work being judged later by people you respect more than the people currently around you. The bigger the imagined retrospective audience, the harder your identity defends itself. You reread the clauses you imagine being asked about two years from now. The clauses you cannot defend in a future room, in front of people you have not met yet.

The trap does not announce itself because everything it produces is real. The contract gets better. Sentences in the pitch deck get crisper. Whatever you are working on, you can usually point to where it is improving. The question the trap will not let you ask is whether better is what the decision needed, or whether better is what you needed.

Field Tested

A Sunday evening at our office in Raleigh, North Carolina. The contract was in front of me. I was rereading the royalty structure clause for what was probably the fourth time that week.

We had been negotiating a licensing agreement with a large dairy company in the US for six months. Royalty structure, who carried which responsibilities: upscaling, strain engineering, commercialisation, payments tied to every part of the project. Material things. Things that would have shaped what De Novo became.

The work was real. That was what made it hard to see.

I had not closed a major partnership deal before. The counterparty was an incumbent in a category we were trying to enter. I had not put the fear into words, but it was specific. In two years our shareholders and investors would read this contract and ask me how I had signed something that boxed us into the wrong shape. I was reading the same clause for the fourth time on a Sunday evening because I wanted to be incapable of being the founder who got the deal wrong. Identity protection wearing diligence’s clothes.

Legal costs ran to about $15K over the six months. Cheap by the standards of a deal that size, because we were trying to keep it lean. The real cost was attention. Six months of it, on a deal that had a problem I could not see from inside it.

Halfway through the negotiation, the EFFV opportunity surfaced. EFFV is bigger and more experienced than us, though nothing like the scale of the dairy incumbent. The deal we discussed with them was a joint venture rather than a licensing structure, and it lined up more closely with what we actually wanted out of a partnership.

We closed it in under two months.

The dairy deal did not close.

The realization arrived later, on a Friday afternoon in Cape Town. I was visiting our office there, sitting with Richard, our COO, going through where things stood. The two deals sat on the table next to each other. Two months on EFFV. Six months on the dairy contract and then a walk-away. Richard said something close to “the pace on the dairy deal was different to how we usually move.” That sentence made it visible.

I cannot tell you in hindsight exactly how many of those six months were necessary diligence and how many were me trying to feel ready to sign in front of future investors who weren’t even in the room. I cannot separate the two lines any more cleanly now than I could then. That is the cost of operating with identity protection running in the background. You don’t pay it in cash. You pay it in months you can’t fully account for.

What I can tell you is the diagnostic. Identity protection scales with the size of the imagined retrospective audience. A big incumbent contract gets studied later. A founder-to-founder contract gets signed and filed. The fear running the calendar is about how the contract will read in two years to people you haven’t met yet.

I caught it weeks after the fact, on a Friday afternoon, in another country, with someone else in the room. That timing is the trap’s defining feature. You see it only when something else moves at the pace you should have been moving at, and the gap between the two paces tells you what was running the calendar.

The Action

Pick the one piece of work you are currently inside that has been moving slower than the rest of your work. You can usually feel which one it is. It is the work you find yourself doing on Sundays when you should not be.

Once you have it, ask the diagnostic question.

Do I genuinely need more information before I move this forward, or am I protecting myself from finding out the answer?

This is the question the trap will fight you on. It will tell you that of course you need more information. The clause needs another read. The pitch needs another rehearsal. The list of things that need another pass is infinite. Some of that need is real. The other half of it is the trap. The diagnostic question forces you to say which half is bigger. It will not separate them perfectly. The bigger half becomes visible. That is enough.

If the protection half is bigger, act inside 72 hours.

The smallest version of the action that produces a data point. If you have been polishing an email, send the third version this week instead of the seventh. If you have been preparing for a conversation, have it on a call by Friday. If you have been refining a deal, do the thing I should have done six months earlier: write the high-level fundamentals on one page before you bring lawyers in. Lock the structure on one screen. Then go deeper.

72 hours is the right window because it is short enough to be uncomfortable and long enough to not be reckless. The trap lives in weeks. Three days starves it.

You will not always be right. Sometimes the trap was actually diligence and you needed the seventh version. That mistake is recoverable. The other mistake, the six months you cannot fully account for later, is not.

The Bearing

The diagnostic question doesn’t make the trap go away. I’ve caught myself inside smaller versions of it more times than I can count this year. The dairy contract was the one I can name. The ones I am inside this week I won’t catch until next month.

What the diagnostic does is shorten the half-life. The trap that used to cost you three months might cost you three days. The cost stays. It becomes a cost you can survive.

Once you are moving, the next question is whether you are moving toward something or away from something. That depends on what you are measuring.

Next week: the numbers that tell you the truth before your instincts do.

One thing to do. One week to do it. Then we go again.

— Jean

The Position

I sat at my desk on a Saturday morning, trying to decide whether I was about to kill a project for the right reason or out of fear.

For four months, we had been building a B2C product at De Novo, my biotech startup. I had pushed for it. The science team had been pulled in, two days out of every five. We had spent close to $150,000 on formulation, packaging, regulations, and contractors.

Three days earlier, our advisors had walked us through the launch plan in detail. By the end of that meeting, the list of things we would need to either hire for or contract out had grown so long it looked like the assembly instructions for a second company.

That morning, I wrote down the sentence I had been avoiding. We were building an entirely new company on top of the one we already had. And the company we already had did not have the shape to carry it.

The Model

There is a strategy framework I keep coming back to from Roger Martin and A.G. Lafley, the former CEO of P&G. They wrote a book called Playing to Win. It lays out five questions any strategy has to answer.

What is our winning aspiration. Where will we play. How will we win. What capabilities must be in place. What management systems are required.

I could answer the first three without much effort. The fourth is the one I had not asked.

What capabilities must be in place. The capabilities you cannot do without to win in the arena you’ve chosen.

Every arena demands its own. A B2B ingredient business demands technical sales, formulator relationships, regulatory expertise, and patience for long contract cycles. A consumer brand demands performance marketing, retail operations, brand, customer service at scale, and the cash to fund customer acquisition for years before payback. The arenas look adjacent from the outside. The capabilities required to win in each one do not overlap.

The question Martin’s framework forced on me is whether the arena I had chosen required capabilities we already had, capabilities we could build without becoming a different company, or capabilities that quietly meant we would have to be a different company entirely. The third answer is the one that kills companies. The arena is right for a different company than yours.

Field Tested

I want to be honest about why I pushed for the B2C product, because the lesson is in the why.

De Novo’s commercial path is long. The science alone takes years. After that, B2B sales cycles for food and beverage formulators routinely run 12 to 24 months before a contract closes. Four or five years of technology development, plus another two on the sales side, before the first meaningful revenue lands. For an early-stage company, that timeline is hard to live inside, and it is hard to keep investors enthusiastic inside it.

The B2C move was, in my head at the time, a way to compress that. We could put a consumer product on shelves. We could show traction earlier. The end consumer of our ingredient was the same kind of person who would buy a finished product. The story lined up.

I was actually looking for a way to make the timeline feel shorter, to investors and to myself, and I built a story about traction to justify it. The product alignment was real. The company alignment was not, and I did not test the second one before I committed the team to building.

Our advisors did the work I had not done myself. They walked through marketing capability, distribution channels, retail buying, social media operations, customer support, performance ad spend forecasts, and brand work. By the end of the meeting, they had described a company we had not built and could not staff. I had no honest answer about who in our existing company would run any of it.

I sat with it for three days. I had been carrying the doubt longer than that, weeks, possibly months, but those three days were the first time I let myself test the kill on first principles. I needed to know whether I was killing it because the math did not work or because the work had become hard. The math did not work.

I told the founders and executives the following week. Same room I had pushed for the project in. I laid out what had changed in my thinking. There was less pushback than I expected. Some of the team was relieved. Others had spent serious time on it and were disappointed. Everybody understood. I told the board separately. They were quietly disappointed and supportive. They could hear that the call was not a panic.

The aftermath surprised me. I had expected to feel some version of failure. What I felt was a mountain off my shoulders. The relief told me what I should have known weeks earlier. The doubt had been a load I had been carrying privately, and the cost of carrying it was higher than the cost of naming it.

The arena was real. We were not the company that could win in it.

The Action

If I were messaging you back tonight about how to think through this for your own company, I would start with this exercise.

Take the arena you’ve chosen, or are about to choose. Be specific. The arena is something like “mid-market HR teams buying compliance software,” or “specialty coffee shops sourcing dairy alternatives,” or “biotech procurement leaders sourcing functional ingredients.” Vague labels like “the wellness market” will not work for this exercise.

Then write down the capabilities required to win in that specific arena. What is structurally required to compete. Sales motion. Channel infrastructure. Customer support model. Pricing model. Marketing function. Whatever the arena demands of anyone competing in it.

Then mark each capability against your current company in one of three ways.

1. Have it.

2. Can build it without becoming a different company.

3. Would have to become a different company to build it.

The exercise is a fit check. Does the arena you’ve chosen require a company you can become from where you are today, or a company you would have to start over to be?

Run this once before you commit your next quarter to the chosen arena. If most capabilities mark “have it” or “can build it,” you have your right to play in that arena. If most mark “would have to be a different company,” you have either the wrong arena or the wrong company. Either answer is worth having before you spend another quarter trying to win a game your shape cannot win.

The Bearing

Distribution is the area where I still have most of my lessons ahead of me, not behind me.

The B2C kill taught me to test the arena against the company before committing. The question I am working through now at De Novo is which of the arenas we can win in is going to compound the fastest. I do not have that answered yet. I will tell you when I do.

The other thing the kill taught me is the gap between knowing what to do and doing it. I carried the doubt privately for weeks before I named it. Knowing was the easy part.

Next week: why knowing what to do is not the same as doing it. The trap I have walked into twice.

One thing to do. One week to do it. Then we go again.

— Jean

The Position

I’d had a coffee at home before sunrise, two more on the way to the customer’s office, and by the time Leah and I sat down across from the CSO of one of the bigger ingredients companies in the US, my hands were shaking, and I was trying not to show it.

We’d been chasing this customer for six months. If the deal landed, the first year was worth somewhere between $150k and $200k to us, and the relationship moved past $500k by year three. Numbers like that were a rounding error for a company buying ingredients on the scale they did. We were pre-revenue with two new hires already on the books against this ingredient and a year of pilot runs already paid for. The deal was our next milestone.

Leah and I went over the pitch in the car. It was the version we’d been working through for weeks. We were going in to talk about the science of the ingredient and why we believed we’d built something better than the alternatives already on the market. We thought that was the part of the story that mattered to the customer.

The boardroom was the kind that signals a customer is taking you seriously. There was a long table with fancy chairs around it and a tray of snacks and drinks already laid out for us. The CSO sat at the head of the table with his head of R&D and his lead formulator on either side of him. We delivered the pitch the way we’d rehearsed it. I was nervous, and my mouth was dry, and I drank a lot of water, but the pitch came out the way we’d planned.

Then the room got polite.

The CSO went quiet. The questions came from his subordinates. What does it cost? Where are you manufacturing? They were the kind of questions you ask to fill the silence in a room you’ve already decided to leave. I didn’t read it that way at the time.

We left, got in the car, and on the way back, I was the one who said it first. “I think that went well.” Leah agreed. We spent the drive talking about what closing this customer would do for us, the press release we’d write, and the next round of funding it might unlock. We were already living inside the version of the story where the deal had landed.

The next morning, I was at my computer at 6 am with another coffee, going back over the pitch. Somewhere between the second and third reread, my stomach dropped. We had walked into that room with an ingredient we were proud of and spent forty minutes telling the customer about ourselves. The cost question and the manufacturing question had been the room politely closing the meeting.

I texted Leah. “I don’t think we landed it.” She wrote back that her excitement had thinned overnight, too.

The Model

At 8 am, with the cost-and-manufacturing questions still rattling in my head, I worked out what those polite questions had really been about. They were the room’s quiet way of asking why I’d spent forty minutes on the ingredient when the conversation needed to be about what they could do once they had it.

I’d read Alex Hormozi long before that meeting. He runs a framework called the Value Equation. The numerator is the customer’s dream outcome and the likelihood you can deliver it. The denominator is the time delay and effort it takes for them to get there. His point, the one I’d read and nodded at long before, is that the denominator is where most of the wins live. The pitch with the shorter path tends to beat the pitch with the bigger promise.

I’d known the framework going in. I’d walked into the boardroom and pitched the wrong half anyway.

The pitch we’d delivered was all numerator. The science of how we made it, the proof of mechanism, and the reasons we believed our chemistry was better than what was already on the market. What the formulator had been waiting for was on the other side of the equation. He wanted to know what he could fit in the sachet now that the dose was smaller, what he could put on the shelf alongside our ingredient that he hadn’t been able to before, and what kind of product he could now build that the constraints of the ingredient had stopped him from building.

Hormozi’s vocabulary didn’t tell me anything I didn’t already know. It told me which half I had skipped.

Field Tested

We were a technical team, and we’d spent the previous year in the lab building this ingredient. We were proud of it. When the moment came to tell the customer about it, we defaulted to talking about the part we’d built ourselves. Forty minutes later, the room went polite.

We had known the problems we should have led with from day one. The dose volumes formulators were stuck with on with current ingredients on the market. The flowability problems that came with five grams of an ingredient going into a sachet. The way the dose ate up so much of the package that brands ran out of room for the other ingredients, which made the product worth selling. We’d built the ingredient to fix exactly those problems. Two grams of ours hits the same effective dose as five grams as what’s currently on the market, with better flow and less volume in the sachet. That story had been in our heads from day one, and we just hadn’t put it in the pitch we delivered.

Once Leah and I got on the phone after the morning text, the first instinct was to write the polite thank-you email and wait to see what came back. That is what most people do after a meeting goes quiet, and it lets you keep believing the meeting might still convert without having to admit, out loud, that it didn’t.

What we did instead was write the CSO directly. We told him we didn’t think the meeting had gone the way we’d intended, that we’d framed the offer the wrong way, and asked for 15 minutes to take another run at it. He wrote back the same day and gave us the time.

The second pitch led with the formulator’s problem. Within ten minutes, the head of R&D said, “This is exactly the problems our customers have been encountering when formulating their products.” Within a week, we had an NDA on the table, samples being shipped, technical evaluation booked, and a commercial discussion on the calendar.

The Action

Pick a meeting from the last three months that went politely and went nowhere.

Send that person an email. Tell them you’ve been thinking about how you framed the conversation, that you don’t think you led with the part that mattered to them, and ask them for fifteen minutes to take another run at it.

The reason you’ll hesitate to send it is that it requires saying out loud, to the customer, that the last meeting didn’t work. That is the act most founders are wired to avoid.

There are only two answers. They say no, which means you weren’t going to close that deal regardless, or they say yes, which means you’ve earned a second crack at a room the rest of the market would never ask for. Most never get either answer because they never ask the question.

Before you sit down to write the new pitch, read the last one out loud and pay attention to where the sentences are landing. If most of them are about what you’ve built and how it works, with not much about what changes for your customer once they have it, you have a denominator pitch to write.

That’s the rewrite to send the same week.

The Bearing

The instinct to lead with what you’re proud of doesn’t leave. I still catch myself doing it. The discipline is to hear the room go polite and correct earlier than the next meeting.

What polite sounds like, in case you’re listening for it, is quiet from the senior person, while the subordinates ask the kind of questions you ask when you’re checking boxes rather than buying. The smiles are the giveaway. They look like engagement, and they aren’t.

An offer that lands is one of the harder skills in the job. It also won’t get you very far if you’ve taken it to the wrong customer. Next week is about where your customers actually live, and how to find them when they aren’t where you first went looking.

Next week: where your customers actually live. The right people are almost never where you first look for them.

One thing to do. One week to do it. Then we go again.

— Jean

The Position

I drove home from that call, sure I wasn’t going to build the thing.

A few hours earlier, I had been certain I was. I had been researching the KYC and FICA compliance space for months. South Africa had been placed on a money-laundering watchlist, and new rules were landing hard on companies not built to handle the administrative load. The offering was structured around the outcome rather than the feature list. I went into the conversation expecting confirmation.

The consultant ran dealer networks for most major automotive OEMs. He told me exactly what I was hoping to hear. The problem was top of mind, and in his words, it was a question of when, not if.

Then he said, “It’s kind of like someone trying to tell me to pay tax.”

My heart sank into my stomach for a few seconds. Hair was on fire, but everybody had already bought a fire extinguisher. The big firms had compliance teams who did nothing but this all day, and nobody was going to outsource regulatory liability to a third party, because if the fine came, it still landed on the company, not the vendor.

By the time I was on the road home, I had no idea whether this was still something I should do. It wasn’t.

The relief came later that evening. I had done the research and structured the offering properly. The only thing I had not done was build anything yet. The walk-away was clean because there was nothing to walk away from.

The Model

There is one concept I come back to more than any other when I am trying to figure out if a market is actually worth building for. Matt Lerner calls it the hair on fire problem. He runs a growth advisory called SYSTM, and before that he ran growth at PayPal.

The idea is clean. Not all problems are equal. Some are mild annoyances people tolerate forever. Others are actively disrupting someone’s work right now, and those are the ones you can build a business around. The tell isn’t them saying “yes, that’s a problem.” The tell is them telling you everything they’ve already tried to fix it.

If they’ve tried nothing, the problem isn’t urgent. If they’ve tried three things and none of them worked, you’ve found it.

That’s the framework. On its own, it is a sharper question than I was asking my own market the first time I tried to run it. But hair on fire is necessary. It is not sufficient. You can find a market that is genuinely on fire and still not have a business. Urgency tells you the problem is real. It does not tell you anything else.

Urgency is the first of three tests. The other two are whether the market has already called a truce with the problem, and whether the economics of your solution would hold if the buyer switched. Miss either of the other two, and you’ve built a business on one leg.

Field Tested

What I had not understood at the time was that I had run one test and called the work done. Urgency was there. Confirmed. That emotional charge of thinking you’ve found an opportunity is exactly what makes you stop asking questions. It took the consultant spelling it out for me to see that the other two tests had been answered too, just not in my favour.

I did explore one more angle. Independent used-car dealers. Smaller operators who might not have internal compliance teams, and who might be easier to sell to. For a minute it looked like a way through. Then I walked the unit economics. Smaller dealers meant smaller budgets and thinner procurement processes for an outside vendor to integrate into. The cost of acquisition would have eaten any margin I could have charged.

That was the third test. Viability. Once I ran it, there was nothing left to build.

Here is what that experience did for me the next time around. At De Novo Foodlabs, my precision fermentation company, we went into customer discovery with a different discipline. We were not looking to confirm a hypothesis. We were listening for what was already keeping our buyers up at night.

An R&D director at a multi-billion-dollar dairy ingredients company, whose job was to figure out what went into the next generation of their consumer products, put it plainly. “It’s almost impossible to get your hands on unless you know someone in the industry.”

That one sentence cleared all three tests at once. The buyer could not fulfil their own commitments without the ingredient, so the urgency was real. The cause of the scarcity was a supply chain tied to dairy production volumes that nobody could control, and the market had lived with it for years without a better option. The price point on lactoferrin was high enough that a better supply solution would have real margin to work with.

We did not build De Novo’s fermentation platform because the technology was interesting. We built it because customers had already told us the way the ingredient was being made was broken. Nobody had fixed it, and the economics of a better solution would actually hold.

The product came after the problem. That order is the difference between building a business and building an exhibit.

The Action

Today, message three buyers and book calls. Live ones, not Typeform. One founder I mentor keeps trying to do this over a form and wonders why the signal is weak. The answers are always going to be weak, because you are not in the room when someone decides what to say.

Use the three tests.

The first is urgency. What is the problem in your business right now that you have tried to fix and could not? If the answer is vague, end the call politely. You do not have urgency.

The second is the truce. If this is still a problem, how are you living with it, and what is that costing you? If there is no truce, the urgency isn’t real. If there is, you are up against something harder to displace than a named competitor. You are up against their muscle memory.

The third is viability. If this were completely solved, what would you pay for it, and how would you justify the spend internally? Price is not the only thing in the way. If the buyer cannot get the budget line approved, your product does not exist in their company, regardless of how good it is.

Three conversations. Look for overlap across all three tests. If there's urgency but the market has already solved it, walk away with a clear head. If the solution holds but the economics do not, walk away. That is how you spend a year on a problem that the other two tests would have killed in a week.

The Bearing

Urgency alone almost cost me a year building something the market was never going to pay for. I was lucky to find it out in research, before I had committed to anything I couldn’t walk away from.

Even now at De Novo, when we’re deciding which ingredient to commercialise next, urgency is the test I still have to stop myself from treating as the whole picture. It is the only one the market tells you about without being asked.

Next week: once you have a problem that clears all three tests, how you turn what you’ve built into an offer the market cannot ignore. A good solution and a compelling offer are not the same thing.

— Jean

The Position

I walked into coffee shops with a product I was genuinely proud of.

We’d built an oat milk at Gourmet Grubb that performed beautifully in coffee. Clean taste, right texture, worked exactly the way a barista needed it to. And it was fresh. We thought that mattered. We thought “fresh” was our edge.

The coffee shops did not care.

Not even a little. They were going to pour it into coffee. Their customer wasn’t going to taste the difference between our fresh oat milk and the UHT carton that lasts three months unopened. Nobody ordering a flat white at 7am is asking whether the oat milk is fresh. It wasn’t a feature they could advertise or a story they could tell. It was just milk that went into coffee.

What they did care about was that once opened, our product lasted four to five days. Think about what that actually looks like on the floor. A café owner opens the carton Monday, writes the date on it in Sharpie, and now she’s watching the clock. If Tuesday is slow, she’s already nervous. By Thursday she’s either pouring the last of it into lattes she didn’t really need to make, or throwing it out. All of that, for a milk that costs more than the UHT carton sitting next to it.

We’d built a premium feature that solved a problem nobody had, and created a real problem in its place.

The feedback was strong. Some of those coffee shop owners told us directly: the extra cost and the extra headache are not worth the value this creates for our customers. They weren’t being difficult. They were being honest. And they were right.

Here’s the part that still bothers me. I studied supply chain at Stellenbosch. I’d spent years treating shelf life as a variable to be optimised, not a cost my customer would have to carry. I had the experience to see this coming, but I wasn’t looking at the right layer. I was thinking about our product when I should have been thinking about how a coffee shop actually operates on a Tuesday morning. I paid for two degrees only to forget where the supply chain actually ends.

We should have had that conversation with our customers before we built the value into the product. Not after.

The Model - Jobs To Be Done

In the 1990s, a Harvard professor named Clayton Christensen was hired to help a fast food chain sell more milkshakes.

The obvious move was to make the milkshake better. Thicker. Sweeter. More flavours. They tried all of it. Sales didn’t move.

Christensen’s team did something different. They watched. And they asked one question: what job is this person hiring the milkshake to do?

Turns out nearly half of all milkshakes were sold before 9 am. To commuters. Alone. No other food. They had a long, boring drive ahead of them and needed something to make it bearable. The milkshake was thick enough to last the whole journey, one-handed, no mess, and fit in the cupholder.

The competition wasn’t other milkshakes. A banana disappeared in two minutes. A bagel made a mess of the steering wheel. Coffee didn’t last the drive. The milkshake won because it could survive the whole commute, one-handed, in the cupholder.

This is Jobs To Be Done. Your customer doesn’t buy your product. They hire it to do a specific job, in a specific moment, with constraints you usually can’t see. When your product does that job cleanly, you win. When it creates friction in the middle of that job, even if the product itself is excellent, you lose.

Every job has three layers: functional, emotional, and social. Miss any one of them, and you’re working against yourself.

Our oat milk was functionally good. But the job we thought we were doing was “put the best possible milk in coffee.” The job the café owner was actually hiring us for was “don’t make my week harder or my margins thinner.” We failed that job. Not because it was a bad product. Because we’d never honestly asked what job it was being hired to do.

Field Tested

I want to be specific about this because the lesson is in the detail, not the summary.

We had the logistics figured out. Refrigerated transport, cold chain intact, distribution mapped. That wasn’t where it broke down. It broke down at the point of use. In the hands of the person who was supposed to buy it repeatedly.

A coffee shop owner doesn’t think about your product the way you do. You’re thinking about quality and what makes you special. She’s thinking about the carton she opened on Monday that now needs to move before Thursday, and the new barista who keeps forgetting to check dates. Meanwhile, the UHT option next to hers lasts three times longer and never once caused her a problem.

We were selling freshness. They were calculating risk.

The gap between what you think you’re offering and what your customer actually experiences is where most products quietly die. Not in a dramatic failure. Just in a polite “we’ll think about it” that never turns into a second order.

The moment I realized we’d built our value proposition around something the customer didn’t value was not some dramatic epiphany. It was quieter than that. I sat with the feedback, and the thought that kept coming back was that I’d been telling people “fresh was the edge” for months before I ever asked a single customer whether that was the edge they wanted.

The Action

Here’s a question that will tell you more about your product in ten minutes than your last three internal meetings combined:

If your customer could get the exact same outcome from something simpler, cheaper, or less hassle, would they?

If the answer is yes, you don’t have a product problem. You have a job problem. You’ve built something that serves a job you defined, not the job your customer actually needs done.

This week, pick one customer. One who’s gone quiet, or who tried your product and didn’t come back. Call them. I know the instinct is to send a Typeform and consider it research. But don’t.

Ask them: “What were you hoping this would do for you, and what actually happened?”

Then shut up and listen. You’re going to want to defend the product, or explain what they misunderstood, or pitch what’s coming in v2. Swallow it. Let them finish. Let the silence sit after they’re done talking. That’s where the real feedback appears.

If they describe a friction you didn’t know about, something your product created rather than solved, write it down. That’s the job you’re failing. That’s where the real work is.

You don’t need to rebuild your product this week. You need to see it clearly. The oat milk was good. The job it was hired to do was not the job we’d designed it for. Knowing that earlier would have changed everything we built after it.

One honest conversation. That’s the starting point.

The Bearing

Understanding the job your customer hires your product to do is the foundation.

Every decision downstream (pricing, positioning, distribution, messaging) is built on whether you got this right.

I still catch myself forgetting it. Twice this quarter at De Novo, my biotech startup. The instinct to pitch what you’re proud of doesn’t leave. You just get faster at spotting it.

Knowing the job isn’t enough, though. You also need to know whether solving it is urgent enough that someone will actually pay you for it.

Next week: the market already knows what it needs.

— Jean

You make decisions without complete information. You move without a clear map. You back yourself on judgment calls that could go either way, and then live with the outcome, good or bad.

That’s not a beginner problem. That’s the job. At every stage, at every size, in every industry.

I’ve been building companies for long enough to know this doesn’t change. At Gourmet Grubb, navigating an industry most people didn’t understand yet. At De Novo Foodlabs, raising capital and building deep-tech in one of the most demanding spaces a founder can operate in. The uncertainty doesn’t disappear with experience. You just get better at moving through it.

The founders who make it aren’t the ones who wait for perfect information. They’re the ones who learn to navigate with what they have. A clear mental model, a process for making decisions, and the discipline to take action before they feel completely ready.

That’s what this newsletter is.

Every week, one note on what I’ve learnt and an action you can take before the week is out. Ideas that have been tested in actual companies, with the honest account of what worked, what didn’t, and why.

By the time you’ve worked through everything here, you won’t just have better ideas about your business. You’ll have a system for thinking about it, one where strategy, communication, and execution aren’t three separate things you’re trying to juggle, but one integrated way of operating.

Dead reckoning is what navigators did before GPS. No perfect information. No guarantee of arrival. Just your last known position, your best judgment, and the discipline to keep moving.

I don’t have all the answers. Some of what you’ll read in these editions are lessons I learned the hard way. Some are frameworks I’ve been using intuitively that writing this newsletter is forcing me to articulate clearly for the first time. All of it is tested in real companies, with real consequences.

This newsletter is as much a learning process for me as I hope it will be for you.

That’s the job.

Let’s get to work.

— Jean