| Episodes | 106 |
| Mentions | 181 |
| Cited here | 39 |
| First — last | #2 — #729 |
| Top guests | Jeri Ellsworth, Avidan Ross, Eric Klein |
| Related | kickstarter · open source hardware · arduino · hardware startup · startup funding |
Venture capital is a form of financing in which specialised firms invest other people’s money in young companies in exchange for equity, with the expectation that a small number of very large outcomes will pay for a large number of total losses.[394][491] A venture fund is organised as a partnership between limited partners, who supply nearly all of the money and hold very little control, and general partners, who supply perhaps one to five percent of the fund and hold most of the control; in exchange for that asymmetry, most of the financial upside returns to the limited partners.[402] Because the model’s filter is a company reaching roughly a hundred million dollars of revenue within about five to seven years, the structure of venture capital shapes which companies can be funded, on what terms, and how they must subsequently be run.[402]
Fund structure
A venture fund is a partnership in which the limited partners put in the money and hold little control, while the general partners direct the investments.[402] The general partners’ own contribution is typically only one to five percent of the fund.[402] This structure is comparatively recent in origin: venture investing began as wealthy individuals deploying their own money, an arrangement that carried no management fee because there was no outside capital to charge one against, and the modern structure in which a firm invests other people’s money is now the overwhelming majority of the industry.[402]
Raising a first fund can be easier from individuals than from institutions: a partner at an established firm may write a personal cheque where the firm itself would not, and the mechanics of management fees make institutional money less attractive to a new manager in any case.[402]
Raising against the fund’s plan is slow even for an operating business. When Microchip was losing money at a time when venture capital was difficult to obtain, chief executive Steve Sanghi spent about nine months raising against a written business plan, during which he did not know from one fortnight to the next whether payroll would clear.[632]
Return expectations and portfolio arithmetic
A venture fund’s own business model sets the filter it applies to companies. The industry target — termed “venture returns” — is a business reaching on the order of a hundred million dollars of revenue within about five to seven years, and the commonest reason for declining an otherwise sound company with a good product is that it will not reach that scale in that time.[402] The defining trait of the model is not temperament but arithmetic: investors place money in risky ventures and require a large return in proportion to that risk, and this governs their behaviour regardless of the individuals involved.[147]
The arithmetic is dominated by losses. A fund’s results rest on a dozen or so large winners sitting alongside thousands of investments that returned nothing at all, on the arithmetic that one hit in five or ten returns the fund.[491][394] A consequence is that a business which is somewhat profitable but never explosive — termed a “lifestyle company” — is unwanted in a portfolio, and companies are pushed toward high-risk moves that either succeed or fail quickly and clear the books.[394]
Investors therefore evaluate a pitch as a plan to dominate a market, knowingly discounting a large share of what they are told and taking the chance anyway; engineers who pitch too practically fare badly at this, which is part of why implausible companies get funded.[394]
Terms imposed on founders
The capital arrives with structural conditions. Investors hold preferred shares while founders and employees hold common stock, so in a wind-up or a sale the investors are repaid first, and a founder’s nominal percentage of the company says little about what they receive.[301] The preference operates in practice: a company sold for one and a half million dollars against at least twenty million invested returns the entire proceeds to the investors before anyone else sees them.[188]
Accepting venture funding transfers control of direction: afterward the founder either goes along with the direction the board sets or leaves the company.[208] The unwritten terms attached to funding matter as much as the money: taking it implies raising a larger round the following year and selling or listing within about five years. On Jason Huggins’s account of his own companies, an instrument that can be bought out, such as a convertible note, removes that obligation and lets a company be run for profit instead.[369] Even a light investor relationship imposes discipline: Huggins’s investors request financials quarterly, which obliges the company to keep them, bringing a bookkeeper, an accountant, and real profit and cash-flow statements.[369]
The expectation reaches working arrangements quickly. A research group folded into a venture-funded consulting firm on the understanding that it would keep operating as it had was questioned within a month or two about what its people were doing and what the exit strategy was.[60] Once a company has taken venture money it cannot simply be operated as a profitable business at its existing size, because the investors’ return expectation is a multiple of what they put in.[447] A company built on a technology from a doctoral thesis and funded by investors is under pressure toward an early sale rather than toward becoming a durable vendor, because the return has to be realised.[2]
Failure modes
Raising more capital than the business needs is itself a failure mode: a company with a hundred million dollars of revenue can still collapse, because the obligation created by the money is to reach the investors’ target rather than to be profitable.[441] A recognised pattern is funding a company just short of what it requires, so that the final tranche is negotiated when the founders are desperate and costs a large additional share.[147] Financing decisions are also made under uncertainty about timing: Larry Sears’s company took a further venture investment that in hindsight it did not need, because of the lag between finishing the product and the sales arriving.[109]
The size of the investment itself is bounded by what building a business costs. Seed-stage cheques from a hardware-focused fund run from around one hundred to one hundred and fifty thousand dollars at the low end up to five hundred or seven hundred and fifty thousand, scaled to the company’s progress and the partners’ confidence in the founders.[402] Avidan Ross’s fund writes five-hundred-thousand-dollar seed cheques, which he characterises as large to write and small relative to what getting a business built costs; companies that raised comparable sums through crowdfunding have failed for exactly that reason.[327]
Exit pressure
Funding history bounds an eventual sale price. Roughly fifty million dollars of venture capital into the board vendor Arduino implies it will not be sold for much under two hundred million, while its actual business — about two million boards a year with roughly eighty percent of income from hardware — does not obviously support that figure.[707]
A funded company is less precarious than it appears from outside, however: once a company has its first round, its board is filled with people who have a vested interest in its success, and a team that delivers, or delivers on enough and pivots sensibly, can generally raise the following round.[407]
Hardware and venture capital
Venture capital began in the 1960s funding hardware — Fairchild and Intel are the founding examples — but modern practice runs against the sector: given a software and a hardware investment with comparable expected returns, an investor takes the software one, because hardware carries capital requirements and demands eight to ten distinct disciplines among the first fifteen employees.[495] Consumer hardware in particular is avoided because the ratio of failures to successes is skewed so heavily toward failure that the occasional large acquisition does not compensate.[495] The standard rule for consumer hardware is that the selling price must be about four times the bill of materials for the business arithmetic to work, a multiple founders routinely find implausible until they run the company.[495]
Hardware sold at twenty to thirty points of gross margin is not by itself an attractive business, which is why hardware companies were relatively underfunded for a period; what changed the assessment was the redefinition of hardware as the means of delivering software into a user’s life away from a screen.[402] The premium investors place on growth speed disadvantages hardware structurally: scaling a physical product from one unit to a million is a machine that has to be built and cannot be built quickly, so capital flows to businesses whose marginal cost is near zero.[491] Entering the medium-volume hardware market takes roughly two to ten million dollars, which in practice comes from venture investors, and those investors expect the intellectual property to be protected — a direct tension with releasing the design openly.[113]
Capital attention, rather than technical merit, determines which manufacturing technologies are publicly prominent: 3D printing is better marketed than precision casting largely because venture money is going into one and not the other.[405]
Sector structure also conditions outcomes. A market with two entrenched vendors is harder to enter than one with a single dominant player: in programmable logic the two leaders fight each other and a new entrant is crushed incidentally, and although around twenty companies have declared an intention to enter, none has succeeded, even as capital continues to go in.[103] Capital requirements vary enormously by sector: an EDA startup can begin as two people with laptops writing code, and one argument put to that industry is that a ten-million-dollar business run at twenty percent profit — about the scale of a dry-cleaning franchise — is a perfectly good outcome.[99]
Deferring outside money is possible even in silicon. Øyvind Janbu’s fabless semiconductor company was founded with about eight or nine employees on the founders’ own money, much of it from a previous acquisition, and operated for its first few years before taking venture funding from two firms.[95] The reason to delay a raise is price rather than independence: the same sum of money buys a smaller share of a company that has already been built up — ten percent rather than fifty.[95]
A vendor’s funding structure is also a purchasing consideration for its customers, because no exit is good for them: a listing pushes the vendor toward growth and revenue extraction, and an acquisition usually ends in the product being shut down or absorbed.[542]
Investors and selection
Angels and firms
The formal difference between an angel investor and a venture firm is the size of the cheque; the practical difference is that angels tend to be more useful to a company while venture firms write larger cheques.[163] Angels invest for reasons beyond return and expect the portfolio rather than any one company to pay; venture investors are doing it only to make money. The distinction is between someone whose job is investing and someone whose hobby it is.[163]
Entrepreneurs in residence and accelerators
An entrepreneur in residence is paid and given an office and a firm’s full network of contacts, on the understanding that they will start a company the firm finds interesting; the firm may or may not fund it in the end.[129] The arrangement carries a hazard: if the sponsoring firm then declines to fund the company, other investors read the decision as a mark against it regardless of the actual reason, and the founders still have to run the ordinary fundraising circuit.[129]
Firms use more than one selection model: some choose the people and trust the eventual idea to follow — the accelerator model — while others pick an area of interest and fund several companies attacking it.[129] An accelerator’s purpose is to bring a team to the point of being fundable, and what investors will fund is high volume or a path to a hundred-million-dollar business; that requirement is what makes a domestic-only hardware accelerator difficult, since the manufacturing engine to replicate the product at those volumes has to exist somewhere.[113] A hardware accelerator supplies more than money — in some cases a machine shop and dedicated staff working alongside the founders — because hardware lacks the network of experienced people that software accelerators can draw on.[113] For an investor, the format replaces a decision made on a pitch in a week or two with three to six months of observation of how the team functions under stress, overcomes obstacles, and pivots, so that by demo day the investor knows both the team and the product’s evolution.[113] Accelerator operators do not necessarily hold the capital themselves: they raise it from investors on the strength of their expertise, and those investors take a share and get first access to the companies coming through the programme.[209]
Fund theses
Some funds treat the investment as the start rather than the end of the work, advising and coaching companies afterwards instead of acting only as “check writers”, and staff themselves with people who have built the kind of product they fund.[437] A technically-run seed fund selects on the founders, and the founders who resonate arrive with half-disassembled prototypes wanting to discuss the algorithms running on the microcontroller rather than a polished pitch deck.[437] One seed thesis in the hardware sector splits between the products themselves — low-cost robotics and connected devices — and the “picks and shovels”: the engineering, developer, manufacturing, and prototyping tools used by everyone making products, with the fund entering as the first institutional money after the founders’ own.[327]
Established industries with expensive problems and substantial revenue but no internal research and development are systematically underserved, because investors concentrated in one place tend to fund the problems they personally encounter and to look for a single outsized consumer success to cover the rest of the portfolio.[327] Funding one’s own problems is a specific selection failure: investors who are already wealthy will back wine services, private-jet programmes, and thousand-dollar kitchen appliances, none of which address a problem an ordinary customer has.[327]
Patents and open source
The first question from investors is often how many patents a company holds, so patents get filed to make the company presentable rather than because the founders wanted them.[173] The signal runs both ways: investors commonly ask an open-source company about its patent position, but a founder who opens with their patents signals litigation rather than a business.[327] In semiconductors a patent portfolio is both defence and currency: a company of any size will eventually be sued whether or not it set out to infringe, so the size of the portfolio determines what it has to trade.[129] A failing company’s portfolio has a second life: it is taken to an investor as a rescue, and the investor monetises it by asserting the patents against companies that are actually making things.[297]
Open licensing creates a direct tension with the model. Investors who arrive after founding have no attachment to the conditions under which a company succeeded; their interest is the return, which puts an openly licensed product under pressure once the target is a company worth hundreds of millions rather than tens.[114] The share-alike condition is the specific deterrent for investors, because building on the licensed work obliges the company to release its own work in turn.[114] One open-source hardware company concluded after several years and its venture funding that it could not continue as a fully open company and grow into a major player in its industry, and closed several aspects of the design.[127] The expectation of defensible intellectual property is a plausible contributor to the semiconductor industry’s shift from open sharing to secrecy, since capital that funds a capital-intensive business wants to own what it funds.[501]
The funding itself is declinable. Stephen Hawes, developer of the open-source LumenPnP pick-and-place machine, turned down a venture offer from one of the project’s own backers rather than close the design.[686]
Alternatives to venture funding
Most products sold in general retail were never venture backed, so the first question to settle is whether one wants a venture-backable business or a good business; the second is well served by a modest crowdfunding campaign, shipping it, and moving to the next product.[327] The model has been summarised as fuel poured on a fire: a small fire can be built up over time with twigs and logs, but pouring jet fuel on a fire that cannot accommodate it does not help.[327] The term lifestyle business is used disparagingly but describes a deliberate choice; the trade being made is a larger possible outcome against the conditions of one’s own working life.[355] The distribution of company formation is a power law whose visible tail is young founders, but the bulk of companies and of the revenue they generate comes from established engineers and business people; a startup does not have to mean a venture-funded attempt at a billion-dollar outcome.[83]
Bootstrapping
Bootstrapping means starting with an amount of one’s own money one can afford to lose, selling a small batch, and reinvesting the profit into a larger one. The failure rate of that path is very low where the failure rate of venture-funded ventures is high, because nothing is staked beyond what the founder can absorb.[24] Retaining ownership retains control: the majority shareholder decides, and a founder diluted to a few percent answers to shareholders on questions the founder used to settle; putting in one’s own money also aligns the effort, since the founder carries all of the risk.[24] The instrument company Saleae was bootstrapped on roughly a thousand dollars to build ten units, selling them and compounding the profit into the next batch, with the internet supplying the distribution that would otherwise have needed a channel.[237] Bootstrapping has become unusual enough that building a thirty-three-million-dollar business without outside funding is treated as remarkable, and graduates commonly believe that a funding round or a crowdfunding campaign is the only way to start a company.[264]
Some businesses cannot be bootstrapped. A design tool built by an average of seven people over three years with effectively no sales represents about a million dollars of payroll before any revenue, which founders cannot self-fund, and working on it part-time means being overtaken.[163] Capital efficiency is achievable even in that model: the design-tool company Upverter raised about four hundred and fifty thousand dollars through an accelerator and a demo day, took roughly six hundred and fifty thousand in total, and ran on it for three years before raising a priced seed round.[163]
Crowdfunding
A crowdfunding campaign changes the negotiating position with investors: terms offered before one campaign demanded a large share of the company, and once the campaign was visibly succeeding the same investors were calling to close, which let the company push back.[173] One assessment at that time was that crowdfunding remained viable for hardware but not for software, except where the project was narrowly defined for a technical audience.[173] An investor behind a crowdfunding campaign changes what the campaign can offer: instead of pricing at a hundred dollars against a fifty-dollar bill of materials at a five-hundred-unit minimum order, the company can commit to ten times the order quantity and price at seventy-five, reaching a wider audience.[327]
Other patrons and roles
Some problems are funded outside the venture model entirely, by a wealthy individual who prefers difficult problems that take years and would not otherwise be backed.[407] For engineers starting a company under any funding model, the standard advice is to be honest about what they are not good at and bring in a business person early to set the company up, handle the legal work, find the market, and raise the money — itself high-value full-time work.[407]