Ian Sigalow

The through-line

Sigalow describes venture as an empirical, numbers-driven trade in which the hardest work is not finding companies but choosing among them. "The hardest part I think is actually the picking" [1], he says, and it is a skill that resists shortcuts: it takes reps, it cannot be learned quickly, and hindsight is misleading because the information available at the time was rarely as clear as it looks later [1][6]. Alongside that sits a refusal to theorise from the armchair. "I can't sit in a dark room and pull the shades and dream up the future" [1][6]; you have to go and test it, which is why he calls venture "a hand-to-hand combat business" and why Greycroft sits in New York, LA and San Francisco, the ecosystems where the top technical talent understands "the state of the possible" [1][6].

The other constant is candour about luck. Greycroft closed its first fund in the autumn of 2006, Amazon launched AWS roughly thirty days later, and Apple launched the iPhone a few months after that, none of which was in the firm's PPM [1][6]. "You can't get better than lucky" [1][6]. What has shifted over eighteen-plus years is the scale and the structure: from a $75 million friends-and-family pool to a platform of more than two billion dollars across six early stage funds and three growth funds [4][5], from a digital media thesis built for trade sales to a discipline of underwriting companies that can reach a billion dollars of revenue and go public [2][5], and from meeting founders after they have a prototype to tracking people years before they start anything at all [2].

On what makes a founder fundable

Asked for the one thing he looks for, Sigalow answers with three. The first is a "master of two domains", a CEO who both understands the product deeply and can sell; plenty of people sit in one bucket or the other, but true expertise in the middle of that Venn diagram is rare, and the founders who have it have historically built the biggest venture-backed outcomes [2]. Selling, in his definition, is far broader than booking enterprise revenue: it is every round of funding and every hire persuaded to leave Meta or Google and take a ninety percent pay cut for startup equity. "If you can't sell your own product, you can't expect anybody to sell it for you" [2]. The second is a market big enough for a company to go from zero to a billion in revenue, because a billion is the number that lets you take the company public at the end of a ten-year fund that is honestly a fifteen-year fund [2]. The third, and his favourite, is an "earn secret": something about a product, market or structure that leads in success to a winner-take-all outcome, and that stays secret because if someone had already built it there would already be a winner in the category [2]. The three tend to collapse together. A genuinely exceptional master of two domains has usually also found the secret, and if they are very good they only want to operate in a market that can support a public company [2].

He is quick on the negative screen and slow on the positive one: "I know if they don't right away", inside five minutes, but confirming the rest can take a long time, because fundraising ability masks other things, market size is a risk you take, and the earned secret may never have been tested by anyone [2]. Venmo is his cautionary case. Two founders out of Penn came in pre-iPhone with a crazy idea, the social feed drove engagement and virality, and Greycroft still wrote the investment off and sold into Braintree because the company was burning cash, before it flourished under PayPal [2]. The earned secret in hindsight was that a feed made payments social and that incumbents were never going to ship a good mobile app, buying enough time to reach critical mass, and other investors backed similar companies that could not execute against the same secret [2]. His broader conclusion is temperamental: young investors carry overconfidence, experience brings humility, and "I've learned to think in more of a probability state" rather than a certainty state, which is what makes it rational to take a one-basis-point swing if the outcome could be enormous and the price is right [2]. He has also spoken about the most visionary founder he has worked with and what made that person special [9], and about deals he chose to pass on and how they played out [11].

On dreaming bigger

The corollary of needing billion-dollar outcomes is that modest ambition is disqualifying: "if it isn't big it isn't interesting" [2]. He finds this is where founders outside the Bay Area systematically undersell themselves. "I tell people to dream bigger all the time. I don't tell people in San Francisco as frequently to dream bigger" [2], where people sometimes pitch four companies in one meeting. He calls it a possibly unfair criticism, and locates the cause in exposure rather than talent: founders elsewhere are simply not around people pitching grandiose ideas often enough [2]. He endorses the board advice one of his CEOs received, to stop running the same play repeatedly and try some flea flickers and Hail Marys, and acknowledges that this is hard for operators to internalise when the base case outcome is already good [2].

On the funnel and the twenty slots

Sigalow reasons about venture through the shape of the funnel. When Greycroft started in 2006 the goal was a company at a hundred million of revenue, which could go public at roughly a five hundred million market cap, and one or two of those in a fifty or hundred million dollar fund made the business work [2]. Twenty years later the number of companies going public each year is still about twenty, one of the industry's constants, but the scale required has moved to a billion of revenue and market caps far in excess of inflation or economic growth, partly because retail investing through index funds requires companies large enough to enter an index [2]. In 2021 a hundred companies made the journey, but eighty of them have since gone out of business [2]. What has changed at the top of the funnel is volume: seed investors now look at something like a hundred to two hundred thousand companies a year, which has to be sorted down by a factor of a hundred to a workable one or two thousand, then sorted again to the ones you actually fund, and if you are very good, one of your twenty a year makes it all the way [2].

He rejects the idea that seed has no signal, and this is why he likes it as the hardest problem in venture. Working backwards from the companies that go public, the vast majority still have a founder CEO at the time of listing, so the talent you are underwriting at seed is the same talent at the IPO [2]. And having read the business plans of many companies that went public, he finds almost all of them are doing at the time of listing the exact same thing they pitched at the seed round, expanded but not changed [2]. The signal exists, provided you are focused on the right problems in big enough markets and the right calibre of founder [2].

On sourcing before there is a company

The industry, he says, has shifted earlier: it used to be that a founder built something, raised friends-and-family money, then came in for a one or two million dollar Series A with a prototype and sometimes customers; "today the industry is now about formation money" [2]. Technical founders coming out of places like Facebook AI research have no LinkedIn profile announcing a company, and by the time they do they are already funded, so early sourcing is heavy inbound networking, knowing who the key researchers are, and putting yourself at the centre of the right ecosystem, after which you do "a little bit of dynamite fishing" and build a brand so they come to you [1][6]. As companies mature the motion flips from inbound to outbound, and partners carry the burden of getting quality meetings with busy, high-quality people [1][6].

Doing that at scale requires technology and people data, which he says have become quite good. The approach is close to Minority Report: identify the roughly ten thousand people in his universe of high-potential software CEOs, stay in touch with them, and be the first phone call when they eventually start something, since many of the twenty annual winners are likely to come from that pool, sometimes after a wait because they have the energy before they have the idea [2]. The signals come from exhaust data on the internet: papers published to arXiv and whether the research is interesting, rate of career advancement inside a hyperscaler, network centricity, running all of engineering at Palantir, launching the v0 product at Vercel, building products that grew virally from a dorm room, or rising to prominence in the right division of Israel's 8200 [2]. The people who both pioneered research and built and shipped at scale stand out, and there are very few of them; he cites funding Noam, who invented the transformer and went on to start Character AI [2]. He is untroubled by how easy company formation has become, noting the Thiel scholars going after fourteen-year-olds and that some of them go on to build Ethereum [2].

On research as the foundation of a firm

One of his own earned secrets after twenty years is structural: "I believe that all great investment firms are built on research" [3]. The type of research follows the prediction horizon. Renaissance is trying to predict whether a stock moves in the next second; Bridgewater is trying to predict whether the dollar remains the world's reserve currency; venture is trying to predict what happens to technology trends underpinning the global economy a decade out, such as what social media looks like in ten years or whether phones are still how we communicate in five [3]. The obstacle is volume: five or six thousand software papers are published to arXiv every week, "way beyond the scope of what people can read by themselves" [3]. So Greycroft has been building internal technology to machine read all of them and extract trends, which he argues makes the firm's predictions better and more tangible [3]. That sits alongside a public commitment to the empirical: the picking itself is done in public inside the firm, with the five software partners and the wider investment team [6].

On AI as the next fifteen-year cycle

Sigalow places the current moment in the same class as the events that made Greycroft's first two decades. The firm was born within six months of the iPhone and at the dawn of the cloud, and the industry got an unbelievable fifteen-year run on the back of those two [3]. When GPT-3.5 appeared in November 2022, he and his partners concluded "it's the same moment as the iPhone or the internet or the cloud and it's going to power venture returns for the next 15 years" [3]. Since then the focus has been largely the application layer with some model infrastructure, funding application companies growing very fast [3]. He describes the capability itself as magical: "we're able to get humanlike intelligence from software for the first time", with expression limited mainly by individual creativity, and huge services industries such as marketing analytics, legal tech and healthcare set to be supercharged [3]. He is unbothered by public market volatility, because the whole software and cloud services market is perhaps a trillion dollars against a global GDP above a hundred trillion; CEOs and CTOs deciding to invest two dollars where they used to invest one produces a massive windfall for AI companies precisely because the sector is still a small speck of the total [3]. He led Greycroft's investing in AI companies at pre-seed and seed globally [2], and Greycroft raised a billion dollars in one of the toughest fundraising environments [8].

On the four big trends of the next decade

Offered with the caveat that he is terrible at prognosticating, his ten-year map has four pillars [1][6]. Software is the first, and he treats crypto and AI as subsets of it, alongside quantum, spatial compute and infrastructure; he would discourage anyone from shorting broad software growth after four decades of it, while noting that where within software Greycroft plays changes every couple of years [1][6]. Life sciences is the second, an area Greycroft does not play in today, but where CRISPR, gene editing and the application of AI to drug discovery suggest an evolution driven by the same forces as software [1][6]. Consumer products is the third, on the logic that consumer is seventy percent of US GDP, that novel formulations and material science keep driving innovation, and that while demand is fickle, persistence in great consumer products is extraordinary: he does not know what phone he will use in a hundred years, but expects people to still be drinking Coca-Cola and using Heinz ketchup, and points to the stability of the cash flows and to Warren Buffett's holdings [1][6]. The fourth is impact, with climate or sustainability as subheadings, which he sees as the one that poses an existential threat: you cannot express a strong climate thesis in public markets today beyond Tesla and a handful of others, but it is coming, and if the planet is not livable in a hundred years there is little point building flying cars [1][6].

On the Ohio test

His screen for whether something is real traces back to Akron. At Summit Mall, kids were pulled aside to watch commercials before broadcast, because Middle America was treated as representative: if it works there it works everywhere [1][6]. He still applies it. "Would it pass the Ohio test" is his way of asking whether the rest of America will use and understand a thing, or whether it is fringe behaviour confined to the one percent that will not extend [1][6]. He attributes the broader analytic habit to a household split between a child psychologist mother and a trial lawyer father at Jones Day, stats on one side and stories on the other, with dinner table conversation spent defending a position, which combined with a science background he calls a good crucible for early stage venture, where you have to understand people as well as technology and markets [1][6]. He also notes how many people in venture come out of the Midwest, and that as the second generation in his family to go to college he grew up with some grit [1][6].

On how the firm was built, and its successive acts

Greycroft in 2006 was not a response to other venture firms; it came from where he and his partners saw opportunity [1][6]. The thesis was that consumers were spending increasing time online, the advertising world was lagging, and there was a chance to buy beachfront property in fast-growing consumer sites with capital-efficient business models [1][6]. That was contrarian at the time, when Khosla and Lux and others were doing clean tech, nanotech and hard science to work out what came after the internet bubble, and Greycroft went straight back into the internet [1][6]. The first act was digital media: roughly seventy percent of the portfolio was ad-supported, with heroes in Buddy Media, WideOrbit and Extreme Reach, and Huffington Post bought with the explicit expectation of a trade sale rather than an IPO, entered around sixteen to twenty and sold for around three hundred and sixty million [5]. He is admiring of what made Huffington Post work: a flywheel where scale attracts writers who want a megaphone, more writers means more content, more content means better indexing and more traffic, plus the election-cycle hack of creating a billion landing pages for individual donor names that nobody else was SEOing, converting passers-by into addicts [5]. That, he says, is the best part of the job, when a founder comes in with an insight about how the world works that sounds out there until they say it [5].

The second act was vertical commerce, hunting deliberately for areas where Amazon could not compete or was not competing well, which is why the firm made intentional plays in fashion, including Trunk Club and its stylist-on-demand model [5]. He gives Stitch Fix credit for the better version, charging separately for the stylist to create membership commerce that retains better than a services model, and thinks Trunk Club should have gone into women's [5]. He is generally cool on asset utilisation businesses of the rent-and-return kind: they are capital intensive, fixed asset utilisation is tricky unless manufacturers absorb the capital cost as you scale, and you inherit the problem of disposing of end-of-life product [5]. The third act is multi-disciplinary, with fintech, global insurance contracts, healthcare IT and software, enterprise software now roughly half of all investments, and broad consumer internet, driven both by scale and by a deliberate search for spaces where companies can be taken public [5].

On firm design, ownership and staying un-straitjacketed

Greycroft is run as a platform of distinct businesses rather than one pool. The software team owns sourcing, picking, winning and scaling for software; a similar team owns sustainability, launched with Coca-Cola systems; each fund has its own LP base, its own mandate and its own strategy, including whether seed or growth is the better entry point in that category [1][6]. His argument is that merging everything produces a hodgepodge and makes outlier performance hard, so specialization and focus should produce better long-term performance [1][6]. On decision-making he contrasts two extremes, the pod shop model from public equities where picking is pushed out to a single person at the edge, and the hive mind in the Bridgewater sense where information is pooled and a small group decides. Greycroft is the hive mind: five partners on the software fund, of whom he is one, deciding together [1][6].

He is blunt that specialisation must not calcify. "You can't be the same person tomorrow that you were yesterday", and a sector-focused fund is fine for now but "becomes a straitjacket", which you do not want to wear to work every day; his advice to anyone raising a fund is to be good at something with a track record and credibility, but for that something to be broad enough to extend into the next new thing [4]. Sector rotation, in his phrase, "is everything in venture capital"; get the markets wrong and you lose all your money [4]. He is similarly unromantic about the barrier to entry, noting that "anybody with a checkbook can be a venture capitalist" and telling aspirants to go fund things and see how it works [4]. He also refuses to claim genius for the first fund, calling it roughly a 3x that he was not happy with given what was missed, and crediting two deliberate choices: a theory of portfolio construction that he now thinks was probably wrong, and a theory of how to be a good co-investor that he thinks was right [4].

That co-investor theory was a reaction against the standard 2006 posture, where every firm wanted the same board seat, the same committee seats and a minimum of twenty to twenty-five percent ownership immediately, so "20 became a very expensive percentage to get" and it turned into a death match [4]. He describes it from the founder's side: a company will raise money for years, so being asked to sell a fifth of it to someone met on a street corner, with customary rights and privileges that mean nothing can be done without approval, reads as acquiring a boss rather than a long-term partner [4]. Greycroft flipped the script, and he notes it is one of two female co-founded venture firms managing anywhere near its scale, with co-founder Dana Settle in Los Angeles making the firm bicoastal from day one [4][5].

On geography and the valuation arbitrage

Sigalow's sharpest structural observation is that public markets price companies rationally regardless of headquarters while private markets do not: "the public markets are totally rational the private markets are totally irrational", and up to 2020 there was at least a hundred percent premium for being based in San Francisco [4]. He illustrates it with the ramp difference, where a Notion goes from seed to an eight hundred million pre-Series A while an equivalent New York company would have raised four rounds at a hundred, two hundred, five hundred and fifty, taking three times longer and diluting by fifty percent [4]. Early stage investors in San Francisco enjoyed that mechanically without necessarily noticing it, frog-in-the-pot style, because fast markups mean less dilution [4]. His favourite case is Shipt: Bill Smith said his company was worth a hundred million, Sigalow offered a term sheet at forty-five pre and told him to go to the Valley, tell them there are no direct flights between Birmingham and San Francisco, and see how many hands went up. Sigalow had no doubt the company was worth what Smith said [4].

Post-Covid he sees that unlocking, with multiples once reserved for San Francisco applied elsewhere, which is why not one of the unicorn valuations Greycroft printed in that period was in San Francisco and why East Coast and Midwest funds saw portfolios perform in ways that would have been much harder when investors had to fly to a board meeting [4]. He thinks innovation is accelerating rather than slowing, partly from macro conditions and partly from people discovering during remote work that they did not love the job they were in and starting side projects unobserved [4]. He is not certain the geography of software production changes fundamentally, though: only a handful of cities in the world produce massively scalable software companies, which is why so much of the world's wealth concentrates in a few places, and he expects the cities that were popular in 2019 to be popular again, in part because loneliness drives a lot of behaviour [4].

On mentors and the long apprenticeship

Sigalow attributes a great deal to being introduced to Alan Patricof at twenty-six, calling him a mentor you could not improve on, gracious with network, advice and feedback, and still coming into the office at eighty-nine with the energy of someone in their sixties after running the New York City marathon [1][6]. The introduction came through Glenn, whose teaching assistant he had been in graduate school, and before that he had been a research assistant to a professor foundational in cliometrics, applying modern econometrics to history, asking questions such as whether the Great Depression would have ended sooner had the Fed done A instead of B [1][6]. The Greycroft founding was itself an act of substitution: he came to Patricof seeking four million dollars for StrongData, his two-factor authentication company built on the CVV2 code, and left having agreed to put StrongData on ice and launch a fund instead [5]. Patricof raised thirty million by faxing friends a memo and a fill-in-the-blank cover sheet, tapping out at seventy-five million by the end of the summer, which is also a fair indication of the average age of the 2006 investor base [5]. Friends told him he was crazy to join a seventy-one-year-old on a thirty million dollar fund rather than go to KKR; he never thought twice, because you have to have a certain mentality about wanting to be a builder [5].

Takeaways

• The three-part test for an investment: a "master of two domains" founder who understands the product and can sell, a market that supports zero to a billion in revenue, and an "earn secret" that leads in success to a winner-take-all outcome [2].
• Selling means every fundraise and every hire persuaded to take a ninety percent pay cut, not just enterprise deals: "if you can't sell your own product, you can't expect anybody to sell it for you" [2].
• The number of companies reaching a public listing has stayed near twenty a year while the required scale moved from a hundred million to a billion in revenue and the seed funnel widened to a hundred thousand-plus companies a year [2].
• Seed does have signal: most companies that go public still have a founder CEO at listing, and almost all are doing at IPO the same core thing they pitched at seed [2].
• Sourcing now happens before companies exist, using arXiv contributions, hyperscaler career trajectories, network centricity and social graph data to track a universe of roughly ten thousand potential software CEOs [2].
• "All great investment firms are built on research", which for Greycroft means machine reading the five to six thousand software papers published to arXiv weekly to extract trends [3].
• GPT-3.5 marked the same kind of moment as the iPhone, the internet and the cloud, and Greycroft has focused on the application layer with some model infrastructure since [3].
• Four investable pillars for the next decade: software, life sciences, consumer products, and impact or climate, the last of which he considers the existential one [1][6].
• Specialise but do not calcify: a sector-focused fund "becomes a straitjacket", so a manager's edge must be broad enough to extend into the next new thing [4].
• The San Francisco valuation premium was an artefact of irrational private markets, not of company quality, and it compressed after remote dealmaking took hold [4].