Discover more from benn.substack
Does venture capital ruin great products?
It's at least a little strange how we assume that spending more money building something makes it worse.
The modern data stack, we're now regularly told, is dead. There were many murderers. It was killed by overcrowding. The Fed did it with low interest rates. The Fed did it with high interest rates. Vladimir Putin did it. Sam Bankman-Fried did it. It died in two parts. Reverse ETL killed it. Usage fees and compute charges billed it to death. It was murdered from inside the house, by the naive hubris of inexperienced founders. It died of natural causes and old age—animal spirits and hype cycles have short and searing lives, and aren’t made to last.
Of course, nobody really cares how it died, though. Who actually killed it—or if it is even dead, or if philosophical epochs can die at all—isn’t being investigated. Nobody is on the case; the New York Times won’t send us an alert in thirty years to tell us it’s been solved.1 Because we don’t talk about the death of the modern data stack to bring it to justice; we talk about it because we have something to sell—a counterculture product that is oh-so-different from the status quo, a personal brand to monetize, a Top Voice badge on LinkedIn—and nothing gets clicks like headlines about something being dead. If it bleeds, it generates leads.
Still, whether they’re happening in cynical SEO bait on a corporate blog or in the comment section of a tabloid,2 these conversations matter, because narratives matter. Every post and Twitter thread adds a dot to our collective memory of what happened over the last few years; eventually, those dots will blur into a generally accepted pointillism. If the dots blaming founders dominate, that’s the story we’ll remember and tell; that’s who will be guilty in the court of common knowledge. Beware of young founders starting enterprise companies. Beware of companies built by data practitioners. The history we paint now—regardless of whether it’s right or wrong—is the history that we’ll remind and warn everyone else about in ten years.
Though our painting of the murder is still incomplete, at least one theme is emerging: The VCs did it. The rough story here is well understood. Prior to 2019 or so, the modern data stack was a loose community of tools that aspired to move data products to the cloud. It was a pleasingly modern alternative to the clanky machinery sold to us by Oracle and others—an ergonomic upgrade; a cheaper way to work; a more accessible set of open technologies and languages. It was a set of tools born from the firsthand pain of the people who were creating them, by practitioners for practitioners. It had a short short creed and, over time, a lot of fervent converts.
Then, the story goes, after the pandemic exploded and interest rates hit the floor, venture capitalists caught wind of the club. And not the pioneering VCs, but the kind who chase trends, who ask who else is on the cap table, and who are still looking for the suckers who are inflating the bubble. For eighteen months, diligence went to zero, valuations went to infinity, and a bunch of crossover funds started cannoning money at every data startup they could find.3 The money, the narrative continues, cost companies their soul. User delight became an urgently monetizable asset. Community features were prioritized behind fiduciary duties. Founders were no longer humble builders; they were pawns in a pump and dump.
This used to be fun, we say, and our tools used to be good. But no longer, because VCs ruin everything: “The sudden infusion of billions of dollars in capital without due diligence that brought in so many people looking to make a quick buck on whatever the flavour of the day is.”
It’s a compelling story. Noble protagonists tried to do a noble thing; smooth-talking villains convinced them to play a game that they had rigged in their favor. Everyone sold out, everything got corrupted, and the platforms we loved got destroyed.
Everyone needs money
It also…doesn’t really make sense? It’s not that VCs are Good—far from it. It’s an industry full of entitled, performative, and insecure men4 who are desperately trying to prove how smart they are to entitled, performative, and not-yet-insecure 25-year olds.5 The problem with the narrative that VC money ruins products is that it has to ruin them through some mechanism. When a wholesome and well-meaning startup deposits $25 million into their bank account, their once-loved product doesn’t immediately become terrible through some sort of divine transubstantiation. The money has to work its way to the product, through the choices the company makes. And not all that clear why a company with a bunch of money would make worse choices than a company without it.
A brief programming note: There are two ways in which VC money can affect a startup. One is in how it affects its founders and its people. The second is in how it affects the product it brings to market. This post focuses on the latter question. But the first one is at least as interesting, and will probably be what next week’s post is about.
Consider dbt Labs, which has become the poster child for this dynamic. It built a beloved product that was supported by legions of adoring fans, of both the software and the company. Its growth started going vertical just as the funding market did, and dbt Labs quickly went from a bootstrapped consultancy to a venture-backed behemoth, with hundreds of millions of dollars in the bank. It was around then that people began to ask questions. Within a year, it was open season on dbt Labs, and all the whispers were the same: dbt had lost its original magic. It was no longer an open-source project trying to make people’s lives better; it was a desperate commercial enterprise, trying to convert community and customer love into a sales pipeline. As its sales team grew, its product experience eroded. People were angry, and righteously so: They sold out, people said, and justice is served through product criticism and customer churn.
But there’s something very counterintuitive about those complaints. Prior to raising money, dbt was a side project, built in whatever spare time and with whatever leftover money Fishtown Analytics had when they were done paying the bills keeping their consulting business running. The project could only move as fast as that schedule allowed.
Now, it’s a product funded by $400 million dollars of investment, built by a company that can blow a huge hole in its balance sheet to develop it. Regardless of the sustainability of that model, it’s hard to see how something built by a small team on nights and weekends would be less useful, less complete, and less delightful than one that has dozens of designers and engineers figuring out exactly what people want and building that.
Yes, there are arguments for how this could go wrong. dbt Labs—or any well-funded startup; I don’t specifically care about dbt Labs here; they’re just a useful and well-understood example—could stop building features that people want and start building things that people will pay for, like integrations with Okta. They could introduce dark patterns that frustrate users into buying something they don’t want. They could overbuild, building features that make the product marginally easier to sell, but considerably more awkward to use. They could stop talking to the “community,” and prioritize their product roadmap by how deep the pockets are of the people who are asking for stuff. As a venture-backed business, they’re motivated to make money first, and make people happy second.6
But these incentives apply to companies that are bootstrapped too, and maybe even more so. Companies that haven’t raised venture capital have to make money. One bad month and it’s game over. If anyone is motivated to put their own financial needs ahead of their “responsibility” as community stewards, it’s the leaders of a business whose survival depends, pay period to pay period, on those financial needs being met.
So why do people turn on startups like dbt Labs? My guess is that it’s actually psychological. Prior to raising money, startups seem pure and uncorrupted. People accept their flaws because of that purity—they’re scrappy small businesses trying to make enough money so that they can keep pursuing their passion; they’re underdogs defying the odds to bring people something that might make their lives better.
Venture capital doesn’t ruin products—it ruins this myth. Venture-backed startups are transparently capitalist enterprises. We're already in an existential post-capitalist war with the rich, and one of us is going to end up with all the money. It’ll be them, we know it, but people will be damned if they hand over their money so easily now. Well-funded startups have money; why do they need more?
In other words, product NPS is a feeling that’s defined by what we think about who we’re paying as much as what we’re paying for7—and “the product got worse” is a useful rationalization.
Growing versus making money
Ed Hammond wrote in June that Musk should just take Twitter public again, and I am increasingly coming around to that view, mainly because it would be the funniest possible outcome. The problem for Musk is:
He is very very good at attracting fans who trust him implicitly and will invest in his companies at high valuations, but
Twitter is not doing so hot right now, cash-flow-wise.
If Twitter is a private company and you own almost all of the equity, you are mainly exposed to its cash flows, not the meme potential of its stock. If you take it public and sell a bunch of stock to your ardent fans, the business results matter less, and your ability to attract attention matters more. Surely, given the facts here, that increases the valuation?
Levine’s argument is that shareholders can value their equity in a company in two ways: Based on the profits the company makes and eventually owes them, or based on how much they can sell that equity for to somebody else.
Most small businesses think in terms of the former approach—so much so, in fact, that equity is often an afterthought. People don’t start yoga studios or family law firms or coffee shops for the equity in those businesses; gyms aren’t using Carta to manage their cap table. People start them because they think they can make money by selling their services for more than it costs to provide them. For these businesses, profitability is essential, but growth isn’t—so long as they can steadily cover their costs, they can operate indefinitely.
Startups that raise small amounts of money8 can function the same way. If they can pay their expenses with their revenue, they don’t need to expand their product or business beyond what’s necessary to keep customers happy and retained.
But once a company raises real money, this is no longer true. Because VCs don’t just invest in startups to make money; they invest in startups to make money by selling their equity for more than they paid for it. They don’t want the rights to future cash flows; they want to cash out, and make all their money all at once.
Because of that, it’s not true that bootstrapped startups and startups that have raised meaningful venture capital both want to make money. The former does. The latter wants to grow. A successful business that people love and produces a reasonable profit isn’t good enough. It has to keep finding new customers, new markets, new features, and new upsell plans. There is no cruising altitude for a venture-backed startup. The options are to fly higher, or stall trying.9
As Randy Au pointed out earlier this week, that’s what ruins products. It encourages companies to push them beyond what they are naturally suited to do. They bolt on unnecessary new features; they build things that only a few people want; they try to colonize adjacent use-cases because that’s what’s required to grow. It makes the product worse, but no board of venture capitalists would approve the alternative—holding steady, leveling off, growing slowly and turning a modest profit.10
I keep asking myself whether it is possible to avoid the fate of wanting to become an all-in-one product. I had a huge amount of trouble doing so because it just felt so weird and unnatural.
I think this is right—it is unnatural, at least for startups that are under a constant pressure to grow. But could they grow in a different way? Our default assumption is that products grow by gradually expanding like urban sprawl: Dead features rise like mountains beyond mountains, and there's no end in sight.
It seems, however, like there might be a solution to this: Build more suites. Suppose that a company builds a useful product that does something small well. It’s found its wedge, and has customers happily using it to solve a narrow problem. Instead of expanding that product, could they just build another, adjacent product?
Most obviously, this would protect the initial product from bloat. The business could grow without forcibly growing the surface area of something that’s already reached its comfortable size. Thinking in suites—or at least, being open to the possibility of building new products rather than always defaulting to expanding the existing one—could also help people draw more natural lines around what a product, or set of products, is meant to do. Features wouldn’t need to be shoehorned into existing workflows.
This would also keep existing products from getting sanded down. The pressure to grow can make specialized products that serve niche customers worse because it encourages vendors to chase bigger, more generic audiences. An all-in-one leviathan has to have wide appeal. A suite of tools can include a simple product for everyday buyers, and a breathtakingly beautiful one for
suckers power users.
There are real obstacles to this—most notably, people don’t necessarily want suites. For example, I’m not sure where, exactly, a tool that’s desperately trying not to be a BI tool draws the line around its first product and starts building its second. But today, we give in to pressure to always be growing, as a product and as a business that VCs will happily back. We'd all be better off—and maybe there'd be fewer reports of our death—if we thought about ways to grow the business without always growing the product.
I was on a plane when The Alert went out this week. Do Androids get the warning? I’ll find out when the aliens show up.
VCs in early 2020: If you thought that I was ratchet with my ass hanging out, just wait until the summer when they let me out the house.
Ok, technically, eleven percent of them are women, which is twenty years behind that notably trailblazing body, the United States Senate. (Also, wait. Forbes has been putting this list together since 2011—so there have been 1,300 total potential honorees—and Adeyemi Ajao, number 96 on the 2023 list, is the first black person on it?)
Not exclusively full of them! Some of them are good. You’re one of the good ones.
To me, a small amount of money is any amount that is low enough that investors never follow up to ask what happened to it. If you raise $1 million from a big angel investor, they’ll never email you to check in or tell you what to do; they just want you to give them more than $1 million dollars at some point in the future. If you raise $100 million from a VC firm, they will want to attend board meetings and get updates and maybe fire you. There’s no exact line for VCs where the former type of involvement becomes the latter, but I’d guess that startups could raise somewhere around three million dollars from professional investors before they start caring about where it went.
Raising venture money also affects how companies are allowed to fail. A struggling small business can just shut down. Though they could try to save themselves with a bunch of unsavory growth hacks, they could also decide that they’d rather close with dignity. Struggling venture-backed startups are encouraged to pull out every stop to save their businesses, including ugly product anti-patterns and marketing maneuvers (e.g., everything Twitter has done over the last year).
A lot of people have said that data companies’ valuations are what are ruining their products. I don’t think that actually matters all that much. Had dbt Labs raised their last round at a $1 billion instead of a $4.2 billion valuation, growth would still be the imperative. There is no option to level off for a $1 billion company either.