On Starting Another Company

Jason A. Hoffman, PhD | July 19, 2026

I have started and sold several companies. I am still a technologist and an operator, but I now have more ways to help than I did when I was younger: investor, advisor, first customer, partner, recruiter of the person who should run it. Capital and time are both things I allocate now. Starting a company is one choice available to me, not the default expression of believing in a technology.

So when I ask myself whether to start another one, the question is not whether I have an idea. I always have ideas. That is the bias, not the qualification. The question is whether a specific company deserves to exist, whether it deserves years of anyone’s time, which role I should play, and whether we can make the choices that shape its outcome instead of letting those choices be made for us.

The person I most want to hand this to is a younger version of myself. I was a PhD scientist and engineer when I left academia. I knew how to formulate a difficult technical question, pursue it rigorously, and recognize technically serious work. I did not know how much capital a company could responsibly take, how an investor’s fund economics would shape its choices, how to recruit without an institutional brand, how to turn a first customer into evidence, or how to show an eventual buyer why acquiring the company made more sense than building the technology itself. I learned nearly everything nontechnical about building a company on the job, in sequence, with the consequences already attached.

My technical training made the exposure worse in a particular way. The technology was real enough that I assumed the company around it would form. Then a fund’s return requirement, a customer’s contract, a hiring sequence, a missed market window, or a valuation that seemed flattering could remove the outcome I thought I was building toward. I came to understand that the company can fail while the technology succeeds. When I looked only at the technology, that failure felt mysterious. It usually was not. An important variable was omitted from the model.

I have run the experiment both ways. My first company was technically superior to everything around it, and it was done the usual way: build the thing, raise what was available, hire who seemed necessary, take the revenue that appeared, respond to the board. It was a good outcome. It made millions. It was also twelve years old and a mess when it was acquired. My fourth company was designed end to end before it existed: three specific hires (a combined general counsel and COO, a head of HR, a CTO), every job description, every contract, the exact amounts and timing of the Series A and Series B, payroll, all of it constructed against a defined objective of an acquisition in a defined price range within four years. We built real capability, capacity, and software. We did no revenue. We were acquired on schedule, everyone who participated beat the S&P 500, and I made materially more than I would have made staying employed at a large company.

The difference between those two companies was not the technology. Both technologies were real. The difference was that in the fourth company nothing I could affect was left to chance. For everything I could not affect, I decided what to watch and how we would respond instead of assuming the world would cooperate.

I wrote this in the first person because it is, first, a guide for myself. It is what I would use before giving another part of my life to another company. It is also something I can hand to a technically serious first-time founder and discuss with them. It is not a universal startup template, and it is not a test they pass for me. The useful part is answering the questions together and finding where our assumptions differ. There is no virtue in making a founder learn through ten years of scar tissue what can be made explicit before incorporation. The point is not to remove the genuinely hard experience. It is to keep the known parts of building a company from presenting themselves later as unknowable fate.

This is mostly about companies that build a repeatable product or a strategic capability whose value can grow faster than billable headcount. A consulting or advisory firm is a different kind of business. Its customers usually fund the work through current revenue, its capacity grows by adding people, and its value depends on utilization, margins, repeat business, reputation, and whether the expertise survives the founder. Most of the questions below still matter, but the venture-capital and exit math should not be carried over unchanged. The dangerous middle is a company that calls itself a product company while every new customer requires new senior work. Consulting revenue can fund or teach a product company, but until delivery stops scaling with expert hours, it is still consulting.

The Bias I Correct For First

A technologist’s default error is to anchor in a technology and treat everything else as detail to be arranged later. But a technically anchored company is still a bet on macro events, market windows, and an understanding of time, timing, and timeliness, and none of those are technical facts. Social and economic forces come to dominate the path of every company: who sponsors it, who joins it, whose business models fund it, what the world happens to be doing during its window. I hold this as a rule: technological, social, and economic factors must all be accounted for, on purpose, and none by default. A company that has answered only the technical question has answered one question of three.

This inverts the usual founding sequence. The usual sequence is technology first, then find a market, then find capital. I start with the window: which technological, social, and economic conditions are coming together, and when. Then the buyers: who will pay for the capability, the equity, the jobs, and eventually the company itself, and what has to be true for each of them to say yes. Only then do I ask what to build, how to build it, and on what schedule. The technology is something I choose for the opportunity, not the reason I assume the opportunity exists. After that, I can ask whether I am the right person to build it and whether I want to.

Starting a Company Is Not the Only Choice

A technical accomplishment does not imply a startup. The right home for it might be a new company, somebody else’s company, an internal program, a license, a joint development effort, a lighthouse-customer relationship, or a partner that already has the distribution, permissions, capital, or ability to operate it. It may need more work until one important uncertainty is resolved. It may be a paper. It may be advice. It may be nothing yet.

When I was younger, the main thing I knew how to do was found the company and operate it. I can now help build a company or a market in more ways. I can provide capital without being the CEO. I can help decide how the company should run without taking an operating role. I can be the first credible customer whose use makes the next customer’s decision easier. I can structure a partnership, connect the technology to an existing market, or help an engineer find the person who is better suited to build the company around it. This does not mean I should sit in every seat. It means I should be clear about what help is actually needed and not take a larger role simply because I can.

When I was in academia, I helped my PhD advisor spin a few companies out of the lab. His advice was absolute: a founder should never work at a company they started. In the biotech and pharmaceutical world he was talking about in the 1990s, there was a logic to it. The scientist could remain infatuated with the technology and with discovery, while the company’s job was much narrower: take a single compound through an expensive development and regulatory process. The lab kept creating possibilities. The company had to choose one and advance it. The person best suited to the first job was not automatically the person best suited to the second.

In the world I knew then, lab spinouts were a more familiar pattern in biotech and pharma than in technology or computer science, unless you happened to be near a place like Stanford. Software often requires continued invention, and the technical founder may be exactly the right person to build and lead the company. I do not treat my advisor’s advice as a universal rule. I keep the distinction underneath it: founding a company, working at it, and running it are three separate decisions. Equity is not a job description.

Technologists sometimes hear corporate development as the deal work that begins after the real work is finished. I think that is a category error. Good corporate development asks whether to build, buy, partner, or invest; who has the pieces that are missing; what each party needs from the relationship; and where the technology should ultimately live. This is systems work across company boundaries. A contract defines how two companies work together. A capitalization table shows who owns what and which outcomes work for whom. A partnership has to keep working even though neither side controls the other. An acquisition has to move technology and people into a new home without destroying what made them valuable. It is the same way of thinking about composition that I described in Zen of Unix Tools, applied to institutions instead of programs.

Corporate development is not a dirty word for a nerd. It is how the nontechnical parts needed for a technical outcome get worked out. Refusing the vocabulary does not make those parts disappear. It only leaves the decisions to whoever arrives later.

The Company Is Mostly Other People

Whatever role I choose, I still have to be honest about what a company is. It is mostly other people: customers, employees, investors, acquirers, competitors, and regulators, all making their own decisions for their own reasons, and all adjusting as conditions change. I cannot change what other people are doing or why. There is stuff going on in the world, and then I start a company inside it.

What I control is narrower: which investors I take money from, which customers I sign, who I hire, the terms I agree to, what behavior the company rewards, and how it responds when something changes. Left alone, a company follows its strongest incentives. You can end up with the business your customers want, not the one you started. You can end up with the business your investors’ fund economics require, not the one you started. Neither is malice. It is what happens when reasonable people follow the incentives in front of them. My job is to choose those incentives deliberately and set limits before they pull the company away from what we meant to build.

This is also where “explicit, predictable, forecastable” needs its correct qualification. A company full of people who keep adjusting to each other and to the world cannot be made predictable. What I can make explicit is what I believe, what I have promised, and what would make me change course. Then when something unexpected happens, we can compare it with what we actually agreed instead of arguing about what we meant. What I can forecast is a range, not a point: the paths the company might take, the signs that it is on one rather than another, and the decisions each path would require. I wrote in On Strategic Decision-Making that a commitment under certainty is a plan and under uncertainty is a decision, and that the output of real analysis is a distribution. Founding a company is the largest instance of that. Finding structure in the chaos does not mean eliminating the chaos. It means knowing where the company is, where it can still go, and which of those paths I would accept.

Four Products, Four Sets of Buyers

I used to describe a founding CEO as having three products: the technical product that customers buy, the financial product (equity and debt) that investors and lenders buy, and culture as the product that employees buy with their working years. The correct count is four. The company itself is a product, sold once to a company for which buying it makes sense, to the public market, or to nobody. If it remains independent, its own cash flow and later investors eventually have to let earlier shareholders turn some of their ownership into cash.

The technical product. Customers buy it. They evaluate utility, reliability, price, and switching cost.

The financial product. Investors and lenders buy it. They evaluate return, liquidity date, control, and information.

The culture product. Employees buy it with their working years. They evaluate cash, equity, authority, meaning, and whether the risk was disclosed honestly.

The company itself. Another company, public investors, or the company’s own cash flow eventually buys out its shareholders. They evaluate what the company can do, what unresolved problems come with it, how hard it will be to integrate, and whether its value can be verified.

Every one of those buyers has a business model and expectations, whether or not they state them. The four products have to reinforce each other. The most common failure is allowing the next financing round to become the real product: the company starts producing whatever the next group of investors wants to see, and everyone else gets what is left over.

Five Questions Before I Start

Before anything exists, five questions. All five must pass, and they are ordered so that each answer sets up the next.

One: exactly what am I accomplishing technically? Say it plainly enough that it could be proven wrong. What evidence would convince me that it works? Who could verify that independently? What will it cost in people, dollars, and time, expressed as honest ranges rather than a single optimistic estimate? And underneath all of that is the harder question: how difficult is this really? Not how difficult it is to describe or demonstrate, but to make work at the reliability the eventual buyer requires.

Two: what does it become? A capability can stay a capability, in which case the likely outcome is that another company buys it and I should keep it easy to evaluate and acquire from the first day. Or it can become a product, which requires sales, a broader team, more capital, and a larger outcome. Or it can become part of a portfolio, which is a different company again. I want to decide in advance when we will choose among those paths, what evidence will decide it, what each path costs, and who has to agree. Drifting from one into another is how a small, focused company quietly becomes an expensive company nobody intended to build.

Three: why now? Which technological, social, and economic conditions have to hold or arrive, and when? What evidence says I am neither early nor late? Which events does the plan depend on that I cannot schedule? A company built to create a strategic capability before it has revenue is especially sensitive to timing. Too early, and no buyer is ready to act. Too late, and the continued absence of adoption becomes evidence against it. For a timed company, the window is not a planning detail; it is what the plan is built around. If that window moves by two years, I want to know now what the company would be worth in that world.

Four: does the ownership and capital math work? At any credible outcome there is only one hundred percent of the company to divide. Before issuing the first share, I want to know what the founder needs to make, what early and later employees should make, and what return the investors will require after dilution and contractual preferences. By terminal value I mean the credible value of the company at the outcome I am actually building toward, not the largest outcome I can imagine. The arithmetic is short. If outside capital is allocated terminal ownership b and requires multiple M, then at terminal value X the company can raise at most A = bX/M. If the amount supported by that equation cannot pay for the technical work in question one while giving investors the return they require, the company does not yet work. There is a technology and there is hope. The shortfall will eventually come out of somebody’s share of the outcome; the only question is whether that is understood at the beginning or discovered during a sale. The postscript works through a concrete example.

Five: is it mine to do? Does this deserve years of my time when I consider the full range of outcomes, not only the most exciting one? What am I actually willing to do, and for how long? Founding and operating is only one choice, and for me it is the most expensive. What would count as an unacceptable success: an outcome that is financially fine and personally wrong? One test I use on myself is whether I would still start the company if I had to hire someone else to run it. That is the part of my advisor’s rule I kept: the fact that I founded the company does not answer what job, if any, I should have in it. If it only works with me running it, I want to know whether I am genuinely scarce or merely in love with the technology.

A no to any question does not mean the idea dies. It may be a project, a paper, an investment, more technical work, a customer or partner relationship, or advice to somebody else’s company. Deciding not to start is a legitimate answer, and one of the most underused.

How I Use This With a Founder

I would not send this to a founder as homework and wait for a score. I would sit with them and answer the five questions in order, over hours rather than weeks. The questions divide naturally between us. The first belongs to the founder: they should be able to teach me what they are accomplishing technically and how difficult it really is, and if they cannot teach it, that is an answer too. The second and third we work together; I usually know more about who buys capabilities and when, and they usually know more about what the work requires. The fourth is arithmetic, and it is fast. The fifth each of us answers privately, and then we compare.

When we find a gap, I want to know what kind it is. If it is a fact we do not know, we research it. If the answer is genuinely uncertain, we describe the range, decide what to watch, and agree on what would make us change course. If we disagree, we write down where and why. None of the three gets repaired with an optimistic sentence in a pitch deck.

The gaps that matter are rarely exotic; the same few miscalibrations come up so often that I now expect them. One founder was about to raise half of what the plan in front of us actually required; the shortfall would have surfaced in year two, mid-build, with the capability half-proved. Just as often I give the opposite advice: do not raise yet. Stay as the two founders, get as much done as possible for another three to six months, and then go out, because a company that can show the proof raises a different round from one that can only describe it. Waiting does not mean going quiet. Take the investor meetings anyway and say plainly: here is what we are building, here is what it will prove, and here is when we will raise. All it makes you seem is responsible, and when the round opens, it opens with investors who watched you do exactly what you said. Others arrive planning to raise too much: the round they describe would sell so much of the company so early that the arithmetic commits them to the no-person’s-land of a $700 million outcome by a date that surprises them when we compute it out loud, or else to raising another quantum of capital before the capability is proven. Revenue plans make the symmetric error, with bookings at the end of year two presented as compatible with running out of cash in the first quarter of year three. Underneath all of them is the same pattern: spending more than the proof requires, waiting longer for revenue than the cash allows, or blowing past a strategic exit with a raise that makes it unreachable. The calibration runs in the opposite direction when the technical work is hard enough to set its own clock. A chip costs what the design, the tape-outs, and the bring-up cost, and it takes what the fabrication cycles take; there is no version of that plan where two founders stay small for six more months and emerge with more proof, because the proof is the expensive part. There the equation answers a blunter question: not whether the plan is calibrated, but whether the outcome can pay for work that cannot be phased down to a cheap experiment.

The conversation does not have to end with me investing or the founder incorporating. It can end with a narrower technical experiment, a commitment from a lighthouse customer, a search for a partner, a different role, a smaller financing plan, a later date, or a clear no. Those are not lesser outcomes. The founder should leave knowing which company they are proposing, what has to be true for it to work, why customers, employees, investors, and buyers would participate, what evidence would change the answer, and where I can actually help. That alone rules out many of the failures that otherwise look mysterious.

If the Answer Is Yes

Incorporation is not the beginning of a company. It puts into effect decisions that should already have been made. I make them in this order.

First, write down what company this is. In one or two pages: why it exists; whether it is building one strategic capability, a venture-scale product, a lasting independent business, or a small company that preserves options without consuming much capital; what we are trying to achieve and by when; which other outcomes are acceptable and which are not; what the founder, employees, and investors can each expect; which terms we will never accept; what evidence would make us reconsider; and when we would stop. These decisions should change only when the evidence changes. Anything left unwritten will be renegotiated continuously by whoever has leverage that week.

Second, write down the ways it could turn out. What might the company look like in years three, five, seven, and ten (and years fifteen and twenty if it is meant to last that long)? It might shut down, recover part of the investment, become a small success, produce a good outcome for the founder but not the investors, produce a good outcome for the investors but not the founder, hit the intended result, or become an outlier. These are possibilities, not predictions. For each one, who gets to turn ownership into cash, who provides that cash, and when? A company meant to last twenty years can take money from a five-year fund only if somebody can buy that fund’s stake before the company itself is sold or goes public.

Third, divide the likely outcome before issuing shares. What would the founder have to make for the years to have been worthwhile? What should early and later employees make for taking the risk? What return does each investor need? And if the outcome is larger than all of those requirements, who benefits from the difference? Most companies never answer that last question, so the answer is eventually imposed by whoever prices the next financing round.

Fourth, work backward from that outcome to each financing round. Raise money to remove a specific risk, and size the round using A = bX/M. The post-money valuation (the stated value of the company immediately after the investment) should be no greater than Xq/M, where q is the fraction of the investor’s ownership that will remain after later dilution. When a term sheet arrives, turn the calculation around before admiring the valuation: X = (post-money × M) / q. That tells me the outcome the investment now requires. A financing can quietly commit the company to a much larger outcome than the one everyone thought they were building toward. I also want an investor whose fund is the right size, for whom a plausible outcome matters, who asks for evidence the company can reasonably produce at each stage, and whose behavior is understood both when things go well and when they do not. I count on follow-on money only when it is legally committed. Otherwise the amount in the plan is zero.

Fifth, decide what revenue is for. Revenue changes how a company is valued; it is not automatically good. What must the first revenue prove: willingness to pay, repeatability, a minimum value for the company, or the ability to fund itself? When should revenue begin? What quality of revenue counts: recurring, retained, diversified, attached to a product rather than custom work, and actually collected? By what date will the company choose between looking for a buyer and building a full sales operation? Some revenue destroys value. A customer contract can constrain the company as surely as investment terms can. Exclusivity, IP rights, data rights, and change-of-control clauses all affect the eventual outcome even though they never appear on the capitalization table. I keep those terms next to the cap table, because an acquirer will examine both.

Sixth, plan the proof around other people’s objections. Everyone the company needs has a good reason to say no. An investor can put the next dollar somewhere else. A customer can build it, buy it from somebody bigger, or wait. An employee can leave during the hard middle. An acquirer can build rather than buy. The next investor can wait for more proof. A milestone matters only if it changes one of those answers. Once the company is running, I ask not only whether we are on plan but whose answer we changed this quarter and whose we need to change next. Objectives say what we are doing. Objections say what we have to prove, to whom, and by when. I keep no more than three questions alive at once that could kill the company. Everything else is ordinary operating work or can be delegated.

Seventh, decide when to step back and look again. Notice things daily without redesigning the company every time something happens. Step back weekly to see what they add up to. Once a month, ask: what kind of company are our recent actions making us? Once a quarter, revisit the likely outcomes, ownership, capital needs, possible buyers, and whether a reasonable outcome can still be good for founders, employees, and investors at the same time. Some events deserve that review immediately because they can change the company by accident: a term sheet, a major customer, an exclusivity request, an executive hire, a technical delay, or an acquisition inquiry. For each one, write down what we know, the decision we made, and what has to happen next. Keeping those three things separate lets us later see why we made the decision and whether we followed through.

Metamorphosis

Every company I have watched closely, including my own, became a different kind of company at some point. The question is not whether that will happen. It is when, why, and how, and above all whether the people involved choose the change or merely discover it afterward. There is also a normal progression from the fun of creating something with a small team to the very different work of running an organization. That change is not drift. Pretending the work has not changed is.

Choosing the change means saying plainly that the evidence has changed, revising the plan, restating the promises to everyone affected, and telling the people who joined, funded, or govern the company what kind of company it has become. Discovering it afterward looks different: one custom customer, one strategic investment, one aggressive valuation, one large sales hire, and one bridge round, each reasonable on its own, until a focused company built to create one valuable capability has become a heavily funded enterprise sales company that nobody chose. Small decisions accumulate, and options disappear in an order no one examined as a whole. The monthly question, “what kind of company are our recent actions making us,” is how I try to notice while there is still a choice.

The Tests I Keep

Underneath all of it, I keep the tests simple.

Everyone gets paid. At an ordinary successful outcome, the founder, employees, and investors should each receive enough to have justified the risk they took. I check the actual order in which money would be paid, not just the headline value of the company. A company that works for everyone only in the best case does not work.

Everyone wants to do the next one. A serial founder’s greatest advantage is that the same engineers, executives, investors, and buyers want to work together again. In my fourth company, everyone made a return better than the index. That fact is worth more to a fifth company than any one relationship from the fourth. An outcome where I win and the other people do not is a failure even if the transaction closes.

Nothing I can affect is left to chance. For everything else, I decide in advance what to watch, what would make me act, and how to limit the damage if I am wrong.

Explicit beats implicit. Much of what founders experience as stress is not uncertainty; it is goals, promises, and acceptable outcomes that were never stated. Writing them down does not reduce the risk. It removes the avoidable confusion.

Simple survives. The description of the company fits on a page or two. The financial model fits in one spreadsheet. No more than three important unanswered questions are active at once. If it takes a library of documents to keep the company pointed in the same direction, I have made it too complicated. The test is whether we have fewer repeated arguments, not more documents.

So: should I start another company? The honest default is no. Most capabilities that deserve to exist do not need me to found a company around them; an investment, a partner, a customer, or some operating help is usually enough. The answer becomes yes only when all five questions have good answers, the ownership and capital math works at an outcome I actually believe, the timing is supported by evidence rather than hope, no smaller role will accomplish the same thing, and I want the work for the years it will take. None of this guarantees the outcome. The uncertainty stays, and it should. What it removes are preventable failures: two parties who meant different things by success, a contract that quietly eliminated likely buyers, a valuation that made the intended sale unacceptable, or a founder who discovered in year six that the outcome they wanted had become impossible in year two.

For the scientist or engineer I hand this to, the offer is not that I know their technology better. They should teach me that. I can help with the parts I had to learn by collision after leaving academia. I have now been the technologist, founder, operator, investor, customer, partner, and buyer. They do not need to master every one of those jobs before they begin. They do need to understand how decisions in each area can change the company they think they are building. We can work through that together before the lessons become expensive.

My first company succeeded despite its design. My fourth succeeded because of it. One took twelve years and was a mess at the end; the other took four and was clean. If there is a fifth, it starts on paper, before it starts.

Postscript: Making the Choices Concrete

The four kinds of company named earlier in this post are not a ladder, and one is not more ambitious than another. They differ in what the company has to prove, how it gets paid for, and where the eventual value comes from. A company can move from one to another, but that move changes the company and should be made deliberately.

One strategic capability. The company exists to make one difficult technical thing real and valuable to a known group of possible buyers. A single-compound biotech company is the cleanest example from my own training: the lab discovers, while the company takes one compound through an expensive development and regulatory process. Its job is not to keep generating compounds. A buyer might pay for the years saved, the team, the intellectual property, access to a market, or a risk it no longer has to take. I size this company backward from a credible acquisition value and forward from the work required to prove the capability. It should have the smallest team and take the least capital that can do the work within the buyer’s window. Revenue may help, but it is not necessarily the proof. I know this is the right shape when I can name who should buy rather than build, why, and by when.

A venture-scale product. The company intends to sell substantially the same product to many customers, with revenue eventually growing faster than the expert labor required to deliver it. It needs repeatable sales, support, retention, and attractive margins, and it usually needs more capital before it can fund itself. I size it from the money required to reach each proof point, then check that the market and plausible company value are large enough to pay everyone who supplied that money. If every customer still requires new senior technical work, the company has not yet proved that it is a product company.

A lasting independent business. The company is meant to operate for a long time rather than depend on a sale. Customers and, eventually, profits pay for its growth. Investors may receive liquidity through dividends, buybacks, later investors, a sale, or a public offering, but the company must not require a near-term transaction to make its economics work. I size it from the path to positive cash flow and match the capital to the time it will actually take. It may use revenue, debt, or patient equity rather than conventional venture capital.

A small company that preserves options. The company spends a limited amount to keep a technology, team, or market position alive while one or two important uncertainties are resolved. It is not an underfunded venture-scale company. I size it from the most I am willing to lose and the least it will cost to learn what I need to know. It does not hire ahead of the evidence. It should be able to stop, remain small, license the work, find a partner, sell the capability, or become a product company without any of those outcomes being treated as failure.

Consulting and advisory work sit on another axis. A consulting firm can be a lasting independent business, but it sells expertise, judgment, and execution rather than a repeatable product. I size it from signed work, billable capacity, utilization, rates, margins, collection timing, and working capital. It may be a very good business and never need venture capital. It becomes a product business only when substantially the same thing can be sold and delivered repeatedly without adding senior labor in proportion to revenue.

A $200 Million Year-Three Example

Suppose the intended outcome is $200 million of equity value in year three. To keep the arithmetic visible, assume no debt, transaction expenses, taxes, secondary sales, or special preference payout, and express every ownership percentage on a fully diluted basis at the sale. A real financing has to add those things back.

Let X be the $200 million outcome, b the fraction investors own at the sale, M the gross multiple they require, and A the total capital the company can take. Then A = bX/M.

Because b is measured at the sale, it already includes the effect of every later round and option grant. An investor who buys 20 percent of the company in an early round owns less than 20 percent by the sale, which means the percentages sold round by round have to add up to more than what investors end with. The factor q in the financing step earlier in the post does that conversion: it turns the percentage bought in a round into the percentage still owned at the sale.

Investor ownership at sale	Capital at 3x	Capital at 5x	Capital at 10x
25%	$16.7M	$10.0M	$5.0M
40%	$26.7M	$16.0M	$8.0M
50%	$33.3M	$20.0M	$10.0M

Take the middle case: investors put in as much as $16 million, own 40 percent at the sale, receive $80 million, and make five times their money. The founders and employees share the remaining 60 percent, or $120 million. If employees own 15 percent at the sale, that is $30 million for employees and 45 percent, or $90 million, for the founders. Those percentages are an example, not a recommendation.

Time matters too. Five times the money in exactly three years is about a 71 percent compound annual return if all the money goes in on the first day. If it arrives in several rounds, each check has a different holding period and should be calculated separately.

If the plan needs $24 million rather than $16 million, something has to change:

At a five-times return and a $200 million outcome, investors need 60 percent of the company. With 15 percent for employees, 25 percent remains for the founders.
If investors own 40 percent and put in $24 million, their return at $200 million is 3.33 times, not five times.
If investors put in $24 million, require five times, and own 40 percent, the required outcome is $300 million.
The company can keep the original $200 million outcome and 40 percent investor ownership if it reduces the plan to $16 million of outside capital and supplies the other $8 million through revenue, customer funding, a partner, grants, debt the cash flow can support, or simply spending less.

Founders owning less can support more capital. Requiring more capital can require a larger outcome. Capital with a lower return requirement can support the same outcome with less dilution. These are not separate facts; they are the same equation viewed from different sides.

What Year-Two Bookings Have to Become

A strategic buyer may value a capability without revenue. A product company usually needs operating evidence. If the $200 million value is expected to come from a multiple k applied to year-three recurring revenue R, then R = $200 million/k.

Bookings, ARR, and recognized revenue are not interchangeable. For this example, let B mean the annual recurring contract value signed by the end of year two, not the total value of a multiyear contract. Let c be the fraction of that value that is live and retained as ARR at the valuation date in year three. Ignoring new sales during year three, B = $200 million/(k × c).

If 80 percent of year-two booked recurring value becomes live and retained year-three ARR:

Assumed multiple on year-three ARR	Required year-three ARR	Required year-two booked recurring value
5x	$40.0M	$50.0M
8x	$25.0M	$31.25M
10x	$20.0M	$25.0M

At an assumed eight-times multiple, the company needs $25 million of year-three ARR. If only 80 percent of what was signed by the end of year two is live and retained at that point, it needs $31.25 million of booked recurring annual value in year two. If only 60 percent converts, the requirement rises to about $41.7 million.

Those multiples and conversion rates are scenarios, not market claims. The point is to make the assumptions visible. A buyer may care about recognized revenue, gross profit, growth, retention, concentration, EBITDA, replacement cost, or strategic value instead of ARR. Bookings can include multiyear value, estimated usage, contracts that have not started, or customers that never fully deploy. ARR itself is not standardized, and neither bookings nor ARR is cash. The company has to define each term, model when signed work becomes live revenue and collected cash, and use the measure the eventual buyer will actually value.

This arithmetic does not predict a valuation. It forces the intended outcome, investor return, ownership, required capital, and operating proof to agree with one another. If they do not, the company is not ready to be financed on those terms.