The whole scheme
Alteryx starts playing a new game. Plus, a lie! On Twitter!
A brief programming note: This blog will be off next week. Happy holidays; bah humbug.
It's not that startups are Ponzi schemes, but they are pretty Ponzish.
Ten years ago, a friend of mine was offered a job at Optimizely, which was, at that time, a leading A/B testing platform for ecommerce websites. The company was doing well: It had been a standout in its Y Combinator class, and recently raised $28 million from Benchmark, a top-tier VC firm. It was on pace to bring in more than ten million dollars in annual revenue, and was growing at 400 percent a year. And it had all the glittery intangibles—a new office in downtown San Francisco with space to grow from 70 people to 450; an upcoming conference with talks from tech celebrities;1 a book.
Like most startup compensation packages, my friend’s offer included stock options. Those options gave her a claim on equity in Optimizely; that equity gave her a claim on Optimizely’s future profits. In theory, the value of that equity is dependent on Optimizely making money, and distributing that money to its shareholders.
But no, obviously not. This is not how anyone thinks about startup equity. Nobody—neither employees nor the early-stage VCs who invest in startups—value a startup’s equity by estimating the future dividends that the startup will pay its shareholders and plugging a bunch of numbers into a Black-Scholes model; they value the equity by guessing how much someone else will pay for it in the future. Their return isn’t funded by the operations of the business; it’s funded by a future investor.
In Silicon Valley startups’ employees and investors, this is the whole scheme: Not to build a business that steadily makes a little bit of money for a long time, but to build a business that someone—another company, a private equity firm, public investors on the stock market—buys for a lot of money all at once.
All of Silicon Valley is built around this scheme. Venture capitalists exist solely to fund the scheme: They invest in a company “until it reaches a sufficient size and credibility so that it can be sold to a corporation or so that the institutional public-equity markets can step in and provide liquidity.” Most founders start companies to get rich from the scheme; companies like 37signals are notable precisely because they don’t do the scheme. People say things like “from idea to IPO” because an IPO is when the scheme cashes out; they say “a startup is at least a seven-year commitment” because the scheme takes about seven years to run.
And startups hire their employees by promising to let them in on the scheme. Ten years ago, Optimizely was very explicit about this. When my friend got her offer, they didn’t tell her how much her options were worth, like you might when you join a big public company that offers stock grants (usually as RSUs). Instead, they told her how much those options might be worth, depending on how well the scheme went. If Optimizely got acquired as a unicorn—the word had just entered the lexicon back then, and we all wanted to use the cool new slang—her shares would be worth $100,000.2 If Optimizely became the next Workday, which was recently valued at over $7 billion after its “breathtaking” IPO, her shares would be worth $750,000. The comparisons kept going up, all the way to Salesforce, then the biggest SaaS company in the world and valued at $50 billion. My friend’s offer, she was told, would make her more than $5 million.
In 2020, Optimizely sold to Episerver, a Swedish technology company, for “less than $600 million.” According to at least one alleged insider, the deal returned nothing to the early employees who joined Optimizely at the time my friend got her offer. The scheme is the whole reason most people are here, but the scheme doesn’t always work.
There are other schemes. People open small businesses—gyms, coffee shops, HVAC repair services—that try to make more money than they spend, and the owners keep what’s left over. Law firms and dental offices create partnerships where senior employees divvy up the company’s annual profits among themselves. Some publicly traded companies attract investors by paying dividends.
And, as described in an excellent essay by Cedric Chin, there is the capital allocator scheme:
Let’s say that you’re a startup founder. Let’s say that you raised a small seed round and built a B2B SaaS (software as a service, selling to other enterprises) and it turns out — five years into it — that your startup will never hit venture scale. It will at most be worth $50 million in valuation; your top-line hits a ceiling of $8M annual recurring revenue (ARR). You think you might be able to get it to $10M ARR, but it’s been two years of stagnant growth now, so you’re not sure.
This is a common outcome in startupland. You’ve rolled your dice and failed to hit the jackpot. The common outcome is to sell the business, pocket whatever money you make (which may or may not be a life changing amount, to be fair), and then reroll your dice. This presumes that you may not keep the results of your current roll and roll a new, second die. It presumes you cannot buy the results of other people's rolls. And it presumes that you cannot add the results of new rolls to your current roll.
It has always puzzled me that so many people are aware of Buffett’s story but do not then reason about the implications of his actions. Buffett bought Berkshire when it was a failing textile business in a dying industry. It had worse than no moat: its death was guaranteed.
Buffett installed a new manager, Ken Chace, atop Berkshire, educated him on ‘return on invested capital’, and told him to harvest (and, if possible, optimise) the cash flows even as the business was in its death throes. Buffett would then use that sputtering source of cash to reinvest into other companies.
For people (like me) who’ve spent most of their careers in Silicon Valley, the idea that employees and investors make money by increasing the value of the company they work for is deeply ingrained in us that it can be hard to recognize that our scheme is a scheme. Chin, again:
It is in the venture industry’s interests that you believe that the only way to a billion dollar outcome is to swing for the fences. It is in their interests to have founders believe that anything less than a venture scale outcome is failure; that they must reset, reroll their dice, and try again for glory. … Consequently, there are many articles about the venture route.
Though people don’t talk about it as much, this capital allocator scheme is also everywhere in Silicon Valley—it’s just not being deployed by startups; it’s being deployed against startups. Stagnating early-stage startups that can’t find another VC willing to pay higher price than their last VC get bought for parts and talent; stagnating late-stage startups that people no longer believe will become Workday or Salesforce get bought by private equity firms to be mined for revenue.
As Chin says, that’s not a bad thing; it’s just not an exciting thing: “The compounding takes a long time, the decisions are decidedly unsexy, and there is only so much surplus that gets created as a result of your businesses.” But if you can’t make a bunch of money all at once, making a little bit of money every day is the only other option.
Earlier this year, Alteryx, a publicly-traded data company, approved their executive team’s compensation packages. Mark Anderson, Alteryx’s CEO, was offered a total package of $57.1 million.
That offer, however, was kind of like my friend’s “$5 million offer” from Optimizely—it was built on the promise of the Silicon Valley scheme. Anderson was being paid a base salary of $650,000; he could make another $1.3 million bonus if Alteryx’s revenue growth well exceeded the annual target. The rest of the offer was tied up in performance-based restricted stock units, or PSUs, that would only be awarded to Anderson if Alteryx’s stock price exceeded certain targets. At the time of the offer, Alteryx was trading at about $50 a share; Anderson would be awarded the first tranche of PSUs if the share price rose to $90. To receive all of the available tranches—i.e., for his offer to be worth the $57 million headline3—Alteryx would have to trade above $240.4
Last week, Alteryx agreed to sell itself to Insight Partners and Clearlake Capital Group, two private equity firms, for $48.25 a share. The scheme is the whole reason most people are here, but the scheme doesn’t always work.5
There are several ways to interpret a deal like this, but here’s one: It’s the end of Alteryx’s run at the Silicon Valley scheme. Prior to the acquisition, Alteryx’s executive team—who all have packages structured like Anderson’s, where relatively moderate bonuses were paid out based on company performance, and relatively huge bonuses were paid out based on stock price performance—was heavily incentivized to focus on how much Alteryx was worth over how much money it made. Even the income-statement-based bonus used “ARR as the sole performance metric,” whereas companies like Snowflake gated their bonuses on revenue, margins, and other metrics.
If the acquisition closes, those incentives will change.6 According to Jamin Ball’s latest weekly summary, most of Alteryx’s key financial metrics—revenue, growth rate, gross margins—are all roughly the same as the median public SaaS company, as are its R&D and G&A expenses. In sales and marketing, however, Alteryx spends roughly fifty percent more than the median SaaS company. In part because of this expense, Alteryx loses money every year instead of making it. And in part because of that loss, Alteryx was valued—and in a sense, for sale—at 4.4 times its revenue, compared to a median multiple of 7.2.
If I had to guess, that’s what this deal is all about: Insight and Clearlake saw a $900 million annual revenue stream that they can buy at a relative discount. Slash sales and marketing spend,7 flip the annual loss into a profit, and bank the cash for as long as Alteryx’s sticky and slow-moving enterprise customer base holds up. If the business takes off, great; Alteryx could be spun out or sold at a higher price than they paid for it. And if the business stagnates, that’s fine, so long as it’s still making money. For capital allocators, that’s their whole scheme.
Here’s another, more general way to interpret the Alteryx deal: The Silicon Valley scheme is getting harder to run. Alteryx has 8,300 customers, makes $900 million a year, and is growing at twenty percent a year. In this market—of higher interest rates, and governed by more aggressive anti-trust regulation—a company with those performance metrics is valuable because of the cash it can make, and not because of what it might eventually be worth.
That should have reverberations for other tech companies—and for data companies in particular, many of which were built exclusively for the Silicon Valley scheme. Like Alteryx, a lot of these companies will likely find themselves as part of a capital allocator’s scheme instead. The only question is whether they’ll be the master of their own game, or a pawn in someone else’s.
Correction - I lied
I’m not a lawyer, but I assume if I put out a press release that said “I raised a billion dollars from Jeff Bezos to launch Frisco’s Bar and Grill,” but I had actually raised a thousand dollars from my college roommate, I would get sued for fraud. I also assume that if I put out a press release that said “I raised
a billion dollars from Jeff Bezos a thousand dollars from my college roommate to launch Frisco’s Bar and Grill,” I wouldn’t get sued for fraud? I mean, I don’t know, but the first one seems like a clear attempt to make money by getting people to come to my restaurant through “fraudulent representations,” while the second seems like a weird joke? The SEC and Department of Justice probably wouldn’t like the joke, but they’re reluctant to sue people for jokes. Plus, if it was clearly a joke, people would never see my press release anyway. I’m a relative nobody with something like 5,000 followers on Twitter; people—and at some point, the SEC and the Department of Justice—would only read my press release if it said Jeff Bezos invested a billion dollars. Nobody would care if I made a dumb joke.
But what if I put out a press release that said “I raised a billion dollars from Jeff Bezos to launch Frisco’s Bar and Grill,” and then put a footnote at the end of that sentence that says “jk, it was a $1000 and it was from my old college roommate Mike.” Would that be fraud, even if the press release itself says it’s not true? What if that footnote was an endnote, and it wasn’t directly attached to the sentence about Jeff Bezos? What if the endnote wasn’t an endnote, but a second press release, issued a day after my first press release went viral, in which I said, “correction, Jeff Bezos was not involved, the investment was led by Mike”?
I guess I could still argue that these are jokes, but I suspect the SEC and the Department of Justice would eventually disagree. The line between a dumb joke and fraud probably exists somewhere between “
Jeff Bezos invested no Mike did” and “now that everyone is talking about Frisco’s, I should admit that I lied about the Jeff Bezos part.”
I’m excited to say that we might get to find out where that line is. Earlier this week, Andrew Young launched a new “AI romance” app—Her, as a PG-13 Tamagotchi, basically—on Twitter. In the second tweet of his announcement thread, he said that “we're grateful to be funded by @naval” (@navel is the handle of Naval Ravikant, one of Silicon Valley’s most prominent investors). Then, after fifteen tweets about all the ways you can customize your Pixar Kelly Kapowski,8 Young issued a “correction,” saying, ah, oops, “naval is not involved.”
First, what? And second, is this fraud?
On one hand, definitely? It’s hard to come up with an innocent explanation here. Naval is one of the most famous dozen or so investors in Silicon Valley; he’s not some two-bit toady who people forget about being on their cap table. On the other hand, if Young had posted his thread as a blog post, though it’d be pretty weird to say Naval was an investor at the beginning of the post and to say he wasn’t at the end, it doesn’t seem like fraud. And Twitter threads, you could argue, are kind of like blog posts. Even the SEC itself uses them that way.
In other words, functionally, Twitter threads are clearly a collection of independent press releases that get shared and read based on their content and their content alone. But legally, are they all one document? If they are, incredible. What a loophole. Young—a relative nobody with something like 5,000 followers on Twitter—posted a video; lied about it; people talked about the lie; that helped the thread go viral; Elon Musk tweeted about it; Naval tweeted about it several times; the app got a bunch of downloads; it shot up the App Store rankings; the video got 22 million views, and the correction only got 300,000. Like, sure, it looks bad to lie, but if you're building some nauseating app so that misogynists can date Lara Croft, how worried are you about having an honest brand anyway?
I have no idea what the actual number was; $100,000 is made up.
Even this number is too low, actually. This is the accounting cost of the package at the time it was awarded, when Alteryx’s equity was valued at about $50 a share. If Alteryx’s stock rose to $240, Anderson would be awarded 1.5 million shares, which would be worth $360 million.
These sorts of compensation packages aren’t uncommon—it’s how Elon Musk became the richest man in the world—but having such a large percentage of the potential payout gated on the stock price alone is unusual.
I mean, the scheme didn’t work perfectly for Alteryx’s leadership team, because Alteryx sold for well short of $240 a share. But it definitely worked. Just from the shares they already own, Anderson will make $7 million for the acquisition, Alteryx’s series B investor will make $50 million, and Alteryx’s founder will make $350 million.
From an internal FAQ about the acquisition: “At the closing, any RSU that is subject to performance-based vesting and that has not become eligible for vesting based on the actual or deemed achievement of the performance metrics of the RSU (including any such achievement in connection with the closing) will be cancelled and will not receive a payment.”
The internal FAQ also includes a cryptic, one-sentence response to a question about potential workforce reductions: “We will continue to adapt the organization as needed to align with industry and business trends.”