David Tisch

The through-line

Across everything David Tisch says, the same distinction keeps surfacing: at the late stage venture is finance, and at the early stage it is something else entirely. "When you're later stage in venture, you're doing finance. When you're early stage venture, you're back in dreams" [1]. That framing shapes his whole practice. He judges companies before there is much to judge, on "an idea on a deck on a demo on you know minuscule traction," which means "all we're able to judge is you and what you're telling us how you're telling it to us and and all of these intangibles that are not math or science" [2]. He is blunt that this makes the job less scientific than the industry pretends: "i believe my job is mostly our minimal science" [2].

The second, harder-edged half of the through-line is his insistence that investors are not the interesting part of the story and should stop behaving as though they are. He describes an ecosystem where "investors and founders are viewed as so much uh in the same bucket in the same conversation," and argues that anyone who has actually started something knows "how separate those worlds are" [2]. His own long feedback loops, his regrets, his fund timelines: "that's my shit and the more that i put that on you as a founder the less you should work with us because that's my problem it's not yours" [2]. What remains constant from the first $25,000 angel cheque to a firm of six making forty to fifty investments a year is that the relationship with the person, not the transaction, is the asset [1][2].

On what he actually looks for in founders

He gives the framework and then dismantles it. The textbook answer is "team it's market it's product it's idea," pressure-tested by "why is this team the best team to work on this product in this market right now," plus a why-now: "is there something that's happened or is there some secret or unlock that you have to make it interesting right now" [2]. Then comes the honest filter: "do we give a shit," which he stresses "doesn't mean that your company's good or bad it just might not be for us" [2]. He is candid about the sectors that leave him cold, ad tech, monetization tech and digital media among them, and about the ones where he lacks the competence to help, battery technology being his example [2]. The escape hatch is the founder's ability to make him care anyway. BoxGroup funded a rocket company "not because we can help with rockets or space or anything but like they were so compelling that we got there," and got to eight synthetic protein companies only after learning enough about go-to-market, commercialisation and the consumer behaviour that would create demand to make the sector legible [2]. What he wants to know from a technical founder is whether they can abstract the work commercially enough that he can then bring in someone who understands the depth [2].

Underneath the sector question is a character question. He wants "ambitious determined hungry like killers," people who will see it through a trough: "you're going to bleed along the way in so many different ways mainly emotionally" [2]. He does not ask founders to enjoy the compromises, only to have understood what they were signing up for [2]. Pandora's two hundred rejections is his shorthand for the trait: "how determined you have to be to go to that 200 first meeting like literally they got shut down by everybody they're like we're gonna keep going we got we're right we're gonna keep going that's grit" [2]. And when he tries to explain what tips a decision, the vocabulary collapses into something deliberately unquantified: "how do i describe what i like i don't know that's pretty that's cool that's that's interesting that's it" [2].

On what founders should want from investors

His advice on fundraising is mostly advice to stop taking investor advice seriously. "if i tell you why i don't want to invest it doesn't probably matter find somebody who wants to invest," he says, adding that unless "a super obvious pattern" runs through all the feedback, founders should "be who you are" and "just find people who believe in you and what you're working on and don't try to go fix the nose" [2]. He also thinks grouping investors as a class is "a silly exercise," since every one of them is different [2]. His ranking of what matters is deliberately deflationary: "the most important thing is money the second most important thing is like good money and then the worst thing is bad money," bad money being the investor who "just going to tell you what to do how to do it" [2]. The question of whether founders genuinely want direct feedback from investors, and whether it is good for them, is one he has returned to elsewhere [7], as is the argument that the best founders do not need their investors at all [4].

He is equally sceptical of the company-builder pose. "there are very few investors in venture that are company builders," and even those who did build something built it in a different era of technology, tactics, culture and go-to-market, so their playbook is an n of one and a dated one [2]. He extends this to founder mythology generally: "make your own movie don't try to replicate someone else's," and do not read a book and conclude "oh that's how you do it" [2]. The structural reason he gives for the misalignment is money. A fund manager has capital committed for eight to ten years and a salary that comes with it; the founder is "putting every single thing on the line," and "you're literally your n is one my n is 150" [2].

On portfolio construction and the seed market

BoxGroup's shape is unusual and he defends it. The firm has been "as small as two and as big as six," makes "about 40 to 50 new investments a year" at that pace for over a decade, and has backed more than 450 early-stage internet companies as total generalists across consumer, enterprise SaaS, B2B vertical software, healthcare, fintech and what he calls quasi life sciences [2]. He rejects the spray-and-pray label with numbers rather than argument: the first fund put money into 152 companies, of which nineteen were worth over a billion dollars at the time of speaking, on track for roughly twenty-three, against a bar where "we are good if we fund three to five of 150 that hit a billion" [2]. He is careful not to overclaim from it: "ten years ago we were good i don't know how we did yesterday" [2]. The related mechanical questions have been a recurring subject for him, including whether concentrated investing works at seed [7], whether ownership matters and how multi-stage funds have changed seed rounds [3], and a set of positions he has argued publicly: that uncapped SAFEs are good, reserves are bad, signalling risk is overstated, and price does not matter [3][4]. He has also laid out what he considers the three variables that actually matter to founders raising a round [3].

On regret, timelines and the broken feedback loop

He describes living "in a state of regret," and defines the real cost precisely: "the opportunity costs in our business are not making bad decisions and funding bad companies it's it's not funding a good company" [2]. The founding regret is Zynga, the second deal he was ever pitched, when Mark Pincus described poker on Facebook and Tisch "told him it was the stupidest thing i've ever heard" [2]. Watching it become a defining company of the following six years set his internal bar: "i can't fuck up this bad again" [2]. Later misses include Scale AI and Casper, the latter passed on for a reason he now finds absurd, picking one of five brands that week and choosing a direct-to-consumer running clothing company instead [2].

He is unusually clear-eyed about why judgment is hard to improve here. "the feedback loop is somewhere in the neighborhood of five to ten years," with failure arriving faster than success, "it's two years on a negative sense" [2]. His illustration is that in 2016 he would have named Blue Apron and Warby Parker as the best companies from the first fund, when the answers were Plaid and Airtable, two 2011 investments he still considered early years later [2]. Blue Apron he describes as an object lesson in the gap between a good decision and a good outcome: backing a company that reached a billion dollars of top-line revenue was right even though the IPO was not [2]. ID.me, his third investment, makes the same point over a longer arc, a 2009 cheque that "feels like overnight success 16 years later and still early in their journey," with a founder in Blake Hall who is "still as hungry and ambitious as he's ever been" [1].

On timing as the hidden variable

Timing is his organising concept, both for markets and for lives. He belongs to what he calls the Xennials or the Oregon Trail Generation, the 1978 to 1983 micro-cohort that "experienced the development of technology in real time," from dial-up at 2400 baud through the web, analog-to-digital cell phones and Wi-Fi, and he thinks living through those inventions rather than inheriting them "sets a foundation for life that's so different" [1]. The point is not nostalgia but sensitivity to how fast context moves: his brother, seven years younger, had a fundamentally different childhood because the internet already existed, and "seven years is a dramatic difference in what the world is like" [1]. He traces the analytic habit to studying history, where his obsession was "studying today in a historical context," and he applies it directly to the work: "thinking about the future in a historical context is really important because you're trying to say where are we going? And it's very hard to say where we're going if you don't know where we've been" [1]. Timing also explains his own path. Graduating in 2003, after the internet "had went up and all the way down," with no tech scene in New York, meant technology "just wasn't a career path or a viable option" [1]. He has separately argued that technology has shifted from being a vertical to a horizontal, and that New York's rise as an innovation hub came from what the city chose to double down on [8].

On careers that are not linear, and being a bad employee

He resists the tidy retrospective narrative. "when you look back on people's stories, a lot of the time, they're not linear. They're random. There's jumps. There's risk. There's timing" [1]. Law school was "a nice punt of making a decision for three years" that he expected to teach decision-making and turned out to be vocational training for a job he did not want [1]. A year in real estate finance followed, then TechCrunch and Ruby on Rails convinced him to jump, then a first company with a college friend where "he wasn't ready. I wasn't ready. We had no idea what we were doing" [1]. Then two years running the internet division at KGB, the B2B provider behind 411, which he calls "a big dumb company" of 13,000 people, and describes as the source of many of his core lessons: "when you work at a big dumb corporate american company you report to the ceo and you're charged with figuring out how to fix it you can learn an egregious amount in two years both good and bad" [2]. The specific lesson was about innovating from inside: "working inside a big companies and trying to innovate is always friction filled," and without top-down buy-in "you were building here and fighting friction there" [1]. The personal lesson was blunter: "I learned very quickly I'm not a good employee. I'm a bad employee" [1]. He counts himself lucky that the 2007 to 2009 window let him hire when nobody else was, which put him in contact with people who became foundational to the modern New York tech scene [1]. The entry point to investing was similarly unplanned: the Angel Boot Camp conference in Boston, where "i don't know that i learned" but the room contained future luminaries, including the founder of Techstars, a conversation about why Techstars was not yet in New York, and the chance to launch and run the New York program [1].

On building BoxGroup as an institution

The firm's origin is almost accidental. His first angel cheque was $25,000 into Boxie, a Roku competitor built on the thesis that "cable was evil" and people should escape it, a company that went on to raise from Fred Wilson and USV and to sell to Samsung [1]. Not wanting his own name on the cap table, he formed an LLC named after Boxie and a Manhattan nightclub, choosing "group" because it "felt bigger than me," never expecting the name to last [1]. The capital structure matured slowly: the first nine years ran on personal capital from himself, his father and his two brothers, deliberately deployed through a fund structure so the early vintages could be counted as funds one, two and three, with outside limited partners arriving only in fund four [1].

He treats team continuity as the achievement he is most willing to claim. Adam left Techstars with him in 2012, Nimi and Greg joined in 2015 and are still there eleven years later, Greg having left and returned, and Adina and Clare both arrived before the firm raised outside capital, so all six long-tenured people predate the institutional money, which he calls "a really cool narrative" [1]. Asked about his own talent, he redirects to this: "when I look at talent in a work setting, I think a lot of that is around people," meaning both who works there and who the business grows with [1]. His hiring instinct is contrarian by his own account, having taken a Cornell Tech job he had initially refused precisely because he did not want it, a filter he endorsed on the spot [2]. He has also spoken about the firm's model alongside the partner who was its first hire outside the founders [6], and about what he sees as the biggest misconceptions and hardest truths of seed investing today [4][5].

On academia, ecosystems and pedigree

His scepticism about universities and startups is specific rather than general. Sitting on the advisory committee for the New York competition that produced Cornell Tech, he questioned the timing rather than the institution: places like MIT, Stanford, Harvard and Penn have had enormous impact, but that credibility "taken a long time to build up" [2]. When every applicant promised to make New York and its startups better "next week," his response was "no you're you're not there's no chance you are" [2]. Absent that patience, he sees no direct correlation between an academic institution and the early-stage ecosystem, other than the rebel culture of students dropping out or starting something straight after graduating [2]. He has argued in the same spirit that defining yourself involves more than pedigree [8].

On collecting, and why it teaches investing

Collecting is his oldest frame for asset appreciation and his first real connection to the internet. Baseball cards, he argues, were "the first Internet P2P commerce," beginning on AOL message boards with reputation posts and migrating "almost overnight" to eBay, to the point that on the day eBay went public most of the items listed were baseball cards and Ty Beanie Babies [1]. The distinction he draws between the two is the investing lesson. Beanie Babies were an isolated collectible built purely on rarity; cards carry "real life performance," so buying a rookie is a bet on what a career becomes [1]. That produces exactly the patterns he sees in startups: "you have like hype and then it doesn't work. You have hype and it works. You have hype, it works and doesn't work and then goes back up" [1]. He is careful not to overstate the analogy, saying it is "not actually parallel," but insists "there's a lot of lessons to be had" [1].

On the rebel character of the work

He frames both startups and the industry around them as fundamentally oppositional. "there's a rebellious nature to starting a company," which he defines as one of two claims: "I am going to do something that nobody else has ever done, or I'm going to do something that someone else has done way better than they have" [1]. Either way the founder starts "with almost no momentum," which is why he calls starting something from scratch "the hardest thing in the world" [2]. Investors do not do that, and he draws "a clear delineation between VCs and founders," while describing his own role as being fortunate enough to join someone else's rebellious journey [1]. It is also why he does not bother explaining the job socially: the ideas sound dumb until years later, "but what's not dumb are the people. The people are the glue here," and that part does not translate at a party [1]. He locates the joy of the work in the same place, meeting people "at a point when they have a dream" and getting "the opportunity to be part of watching someone's dream come true" [1]. Consistent with all of this, he refuses retrospective self-assessment: "yesterday is just less important than tomorrow. And you're only as good in venture as your next investment," a job whose relevance "is about today and tomorrow" and demands you "continue to stay relevant and almost operate with one foot in front of the other versus this like big, broad step back thesis" [1].

Takeaways

• Early-stage investing is judged on intangibles, not analysis: with only an idea, a deck or minimal traction, "all we're able to judge is you and what you're telling us," and the job is "mostly our minimal science" [2].
• After the standard team, market, product and why-now framework, the real filter is whether the investor cares about the space at all, which is a statement about fit rather than quality [2].
• The dominant cost in venture is the miss, not the loss: "the opportunity costs in our business are not making bad decisions and funding bad companies it's it's not funding a good company," illustrated by passing on Zynga, Scale AI and Casper [2].
• Founders should discount investor feedback unless a pattern repeats across every conversation, and should optimise for people who believe in them rather than trying to "go fix the nos" [2].
• Rank capital as money, good money and bad money, with bad money defined as the investor who tells you what to do and how to do it [2].
• Volume at seed is not spray and pray if the hit rate holds: 152 companies in the first fund produced nineteen billion-dollar outcomes against a bar of three to five [2].
• Feedback loops of five to ten years make interim judgment unreliable, as shown by naming Blue Apron rather than Plaid or Airtable as the standout five years in [2].
• Universities rarely move a startup ecosystem on the timeline they promise, and the honest response to "we'll make startups here better next week" is that there is no chance [2].
• Career paths in tech are usually random rather than linear, and time inside a large, friction-filled corporate can teach an enormous amount both good and bad [1][2].