The kingdom of heaven is like a king who prepared a wedding banquet for his son . . . but when the king came in to see the guests, he noticed a man there who was not wearing wedding clothes. He asked, “How did you get in here without wedding clothes, friend?” The man was speechless. Then the king told the attendants, “Tie him hand and foot, and throw him outside, into the darkness, where there will be weeping and gnashing of teeth.” For many are invited, but few are chosen.
—Matthew 22:2, 11–14
TUESDAY, APRIL 5, 2011
Ten million dollars.
Twitter was officially back in the game. They had finally come back with a real offer. While we hadn’t seen a formal term sheet—and the devil was absolutely in those details—it was clear Twitter was now in the realm of 2011 tech-bubble (in)sanity. Our stonewalling had paid off. Even Sacca and I couldn’t claim this wasn’t a tempting offer.
I was riding that particular high when the phone rang.
“Hello, Antonio!”
It was six thirty p.m., and this was the much-awaited phone call from Amin, reporting on our proctological day of interviews at Facebook.
“So, I talked to our engineers and we have the final feedback.”
In the Aaron Sorkin cinematic adaptation of this story, this is where the violins will start scraping their tension-building wail.
“I’m sorry to say that we won’t be moving forward with a deal for AdGrok. The feedback on Argyris and Matt was mixed, and I don’t think it’s a go right now.”
Fuck!
Take another kick in your scarred mug, startup guy.
“In the interests of AdGrok, can you tell me what some of the feedback was?” I sputtered.
Amin changed to that slightly hushed and tense conspiratorial tone that people use, as if hiding in the bushes, when in fact I assumed he was in a closed-door conference room. He proceeded to do some pro bono dragoman-ing. Argyris would have been a possible hire, but Matt was a definite no. Clearly MRM was a gifted engineer, but Facebook had very specific conceptions of engineering greatness. Also, there was a bit of that nebulous “cultural fit” blocking as well.
After my day of interviews, I had imagined that MRM and Facebook would get along about as well as a Berkeley hippie and a Marine scout sniper, and I was right.
“Sorry to hear that, Amin. Thanks for your time and that of everyone at Facebook.”
“Hold on. While that was the feedback for the engineers, the feedback for you was different. We want you to come and join the Facebook Ads team. Your feedback was excellent, and everyone really felt you were an extremely strong candidate.”
My mind stuttered a bit on that one. When in doubt, act coy.
“Well. . . . Amin, as you can understand, I’m somewhat committed to both AdGrok and this other deal we have. I’ll have to think about this.”
“Think about it. But again, we really want you to come to Facebook.”
I looked up suddenly toward our windowed office. I had isolated myself on our balcony in an obvious bid for privacy. Argyris was inside, looking at me with a worried frown. He wanted an answer as badly as I did. I raised two fingers and mouthed “two minutes” to indicate I’d need a bit more time. He nodded, and went back to his screen.
What the fuck should I do? I couldn’t tell the boys this, at least not yet.
Stealthily, making it look like I was readjusting my phone, I hung up and dialed British Trader. While I had moved out a few months ago and we were officially apart, we were still regularly in touch. You don’t just cut contact with the mother of your children, and besides, she still wanted to hear about the AdGrok saga.
“Hey, what’s up?”
“So, get this. Facebook doesn’t want the boys, but they want me. Argyris is right here. I don’t know what to do.”
As an oil woman, British Trader was completely outside the tech scene, and knew little about the intricacies. But she had a savvy read on human nature in a professional setting. Also, given that I was wholly devoid of most human boundaries or morality, she provided a mainstream sanity check on my actions.
“Don’t tell the boys. It will just destroy their confidence. You’ve got to figure out some way to manage it.”
We went back and forth, with me sketching out more details, and her sharing her take.
I looked at my watch. Almost seven p.m. In a few minutes, Argyris would be off like a shot to spend quality time with Simla, the girlfriend turned wife. If I held out for a few more minutes, he’d be out of the office, and I could ignore the boys on email and have a night to think about it.
British Trader and I kept on going, and sure as shit, Argyris got a phone call from the better half, and took off with a wave and a concerned look.
With considerable relief, I hung up with British Trader, gathered my startup kit of messenger bag and laptop, and cleared out in case Argyris came back.
Here’s some capital-H History for you:
Right around 1961, when the Cuban government was televising political executions like they were the Super Bowl, with death warrants signed by that Argentine mama’s boy Che, whose face graces more than one misguided hippie’s T-shirt, my parents fled Cuba. Like many, they left as minors, alone, rushed onto the last flights by panicked parents who foresaw (correctly) that the Iron Curtain would take a Caribbean detour around Cuba before long.
Forty-four pounds of luggage is what they could take. That was expected to encapsulate a life.
For my grandmother, who had considerably greater difficulties than my parents in getting out, five of those pounds were consumed by one essential thing: her heavy, hard-as-diamonds domino set.* Double-nine rather than double-six (Cuba being the only country where that’s the standard game), the tile backing was green, and the whole set encased in a robust but simple wooden box. The only connection with a world that had been riven by revolution and then throttled in the titanic struggles of the Cold War, that domino case was the vessel of memory. It recalled the evening sunsets on the veranda, the warm conversation with friends, the inky black coffee drunk late into the night, with the click-clack of tiles as soundtrack.
And now where was that domino set?
Sitting unused in a closet. Thanks to Zynga, Facebook, and other companies, old Cubans like my mother were too busy playing social games like FarmVille to gather around the table and spin the tiles that had been so painstakingly smuggled. Too busy clicking to buy $0.99 pink tractors and $1.99 spotted digital cows.
Facebook got Cuban old ladies to play computer games! And pay for it!
Think of that miracle for a moment.
And it wasn’t just Cuban old ladies.
In December 2010, Zynga launched a FarmVille clone called CityVille. That game, a moronic rip-off of the far cleverer game The Sims, had accumulated one hundred million users in a month.
One hundred million users!
If humanity had waited until 2010 to invent masturbation, it would not have caught on as fast as CityVille. That’s how fast Facebook could make something happen.
Here’s another data point for you: As part of our push to woo Facebook, I had been getting Google Alerts on the company for months. One in particular had caught my attention. In October 2010, a mother in Florida had shaken her baby to death, as the baby would interrupt her FarmVille games with crying. A mother destroyed with her own hands what she’d been programmed over aeons to love, just to keep on responding to Facebook notifications triggered by some idiot game. Products that cause mothers to murder their infants in order to use them more, assuming they’re legal, simply cannot fail in the world. Facebook was legalized crack, and at Internet scale. Such a company could certainly figure out a way to sell shoes. Twitter was cute and all, but it didn’t have a casualty rate yet, no matter how much this Lady Gaga person was tweeting.
Facebook it was.
But Twitter had come up with the solid offer for AdGrok, while Facebook hadn’t come up with a solid offer for anything yet.
The shambolic hipsters with the expensively decorated offices, thousand-dollar fixies in their bike stand, and the Fail Whale?* Or the hoodie-wearing frat boys with an imperial mandate who coded while they shat? Which was it going to be? Could it possibly be both?
Here’s another truth about tech life: anyone who claims the Valley is meritocratic is someone who has profited vastly from it via nonmeritocratic means like happenstance, membership in a privileged cohort, or some concealed act of absolute skulduggery. Since fortune had never been on my side, and I had no privileged cohort to fall back on, skulduggery it would have to be.
Managing a combined deal between Facebook and Twitter was like trying to engineer simultaneous orgasm between a premature ejaculator and a frigid woman: nigh impossible, fraught with danger, and requiring a very steady hand.
We’ve mentioned Mick Johnson before in our narrative. His company had been in my YC batch and disappeared under mysterious circumstances a few months prior, with Mick magically reappearing inside Facebook. He had made the initial introduction to Facebook Ads that had kickstarted this soap opera.
We both loved hoppy beer, so over pints of Lagunitas at the Creamery he shared the scoop on what had happened with his company.†
He and his Aussie cofounder, James, had a long work history together. They’d been hacking mobile for years and trying to find something that stuck. After two years of making $2K a month (or less), James was done with it. He was getting serious with his girlfriend and sick of the startup gig. They agreed the company had to be sold. So Mick mustered some courage and waded into the talent-acquisition market. They pitched themselves to one and all, going through the M&A process with Twitter, Zynga, Google, Facebook, and smaller companies.
They had gotten furthest with Zynga and Twitter, with Twitter making an outright bid on the entire company. Mick was unimpressed with Twitter and didn’t want to go there. Shaking a few trees, Mick had gotten an intro at Facebook. They had run Mick and James through the M&A wringer and came back with an offer for Mick and Mick alone. Zynga had also come back with an offer, but for the entire company, bringing the number of serious suitors to three. Sound familiar?
What followed was a convoluted imbroglio of haggling that would have made a Somali pirate-ransom negotiation look orderly. The net conclusion of all this was that Mick would go to Facebook, while Zynga would get James and the company. The major problem here was that both Zynga and Facebook had to make concessions to get the deal done, but neither wanted to subsidize the other’s acquisition by offering more value for the hybrid sale. They perceived themselves to be locked in a zero-sum game with a company they didn’t particularly like. The final terms, which I never got out of Mick, were some weird combo of cash up front, equity on separate vesting schedules in both companies, and a corollary deal that got the investors paid.
As I would come to learn, my situation wasn’t unusual, though not generally talked about. Companies with acquisition wherewithal and the nerve to use it bid for what they wanted in deals. You came in with your team and your product; they gave it the once-over, and said, “We want person A and B, but not C, and we don’t care about the tech.” They then offered you a lump sum for what they wanted, and you were left to double-deal, buy out, or otherwise fuck over whoever in order to get the deal done. The company—and places like Facebook and Google did this commonly—cared only about net price per engineer (or product person), not the absolute cost. And they certainly didn’t care what investors got. Many an early-stage acquisition unfolded in this vulturelike way.
I took Mick’s example to heart for two reasons.
The first: he had actually done it. He had rounded up a circle of the Valley’s leading companies and played them against one another until he got the deal he and his cofounder wanted. He stuck it out while Zynga was getting its act together, even at risk to himself and his Facebook deal. He had done something even Paul Graham had never seen and had advised him could not be done. Furthermore, he had played his hand with virtuosity from a place of total weakness. His company had little traction and was running out of the small money it had managed to raise. If the deal had fallen through, he would have been fucked. His only real power was the ability to get Zynga and Facebook into the same room and fight.
The second reason I took his advice-by-example is that I liked Mick, and in a city full of cranky, asocial, self-absorbed, narcissistic startup founders, he came off as a real guy, a mate one could trust. He stood to gain nothing from my deal, and was helping me only as a way to pay it forward to a fellow startup founder in a bind.
By the time we had drained our glasses, I was convinced that if Mick could hack such a deal, then so could I.
But was there a cost?
Long after this AdGrok drama, I’d hear an East Coast tech guy perfectly sum up the reigning attitude in Left Coast tech: “It’s like they have no memory there. It’s the Land of the Stateless Machines.”
A bit of context: “State” is a technical term referring to data kept in memory that a program or function needs to operate. A “state machine” is an abstract model of computation whereby a computational process alternates among a series of states, each defined by a certain set of instructions or data, and transitioning among the various states as triggered by external stimuli. Hence, a stateless machine is a device that simply processes according to some set of instructions, without any knowledge of prior history, like an amnesiac. Our East Coaster’s snide point was that Californians are incapable of rancor or grudges, no matter how outrageous the effrontery. Conversely, they’re not particularly rewarding of generous behavior either.
As every new arrival in California comes to learn, that superficially sunny “Hi!” they get from everybody is really, “Fuck you, I don’t care.” It cuts both ways, though. They won’t hold it against you if you’re a no-show at their wedding, and they’ll step right over a homeless person on their way to a mindfulness yoga class. It’s a society in which all men and women live in their own self-contained bubble, unattached to traditional anchors like family or religion, and largely unperturbed by outside social forces like income inequality or the Syrian Civil War. “Take it light, man” elevated to life philosophy. Ultimately, the Valley attitude is an empowered anomie turbocharged by selfishness, respecting some nominal “feel-good” principals of progress or collective technological striving, but in truth pursuing a continual self-development refracted through the capitalist prism: hippies with a capitalization table and a vesting schedule.
What would the Valley make of my betrayal then? How much was I sacrificing by making it?
The capitalist hippies would take me back, I reckoned, I just had to be minimally successful. The Land of the Stateless Machines would continue crunching away, ingesting people and money, and outputting products, and they’d still be happy to grind me into the mix as well.
WEDNESDAY, APRIL 6, 2011
I had decided to deceive my cofounders for the first time in our harried time together.
As with many such lies, the rationalization was that it served the greater good. The boys were already stressed to the point of hyperventilation with all the shit we’d been through, and now we were betting it all on this very flaky acquisition process that could collapse in an instant. If they realized that this Twitter process was truly do-or-die for them, they’d choke. So what you do as a CEO is internalize that stress for the company and let it consume you instead of the rest.
How’s that for a masterful rationalization?
What’s more, it wasn’t even clear the Facebook offer was for real. In God we trust, everyone else show me an offer letter. I had called Gokul that morning and mentioned I found Facebook’s interest flattering, but I needed to see an offer before I could even begin to manipulate the other side of the deal to spring me.
Today, though, was a day for Twitter.
Part of any acquisition process is what’s loosely called “due diligence.” Taking both technical and legal forms, it’s the snooping around an acquiring company does to make sure it’s actually getting what it thinks it is. On the technical side, it means understanding the company’s “stack”; that is, the pile of interrelated user interface and back-end server technologies that power the product. It might even be as detailed as line-by-line code reviews with the startup’s engineers. You can fake a lot in a startup these days, what with Amazon Web Services and all sorts of off-the-shelf back-end components that let any even minimally competent duffer set up a Web app that does something. Intelligent planning for growth is rare among early startups, but it’s the name of the game at a large, rapidly scaling tech company. Waiting for a team to grow from technical adolescence to mature talent was too long even for a larger company.
As a first step, Twitter had invited us in as a group to talk technical turkey with a pack of engineers that reported to Kevin Weil. We spent a tense and wonky hour locked in a room with the senior engineers on the Twitter Ads team, walking them through our back-end stack that made AdGrok possible. I’m using the corporate “we” here, as it was completely the boys’ show. It had been so long since I’d even touched AdGrok code, there was little I could have said about it. While the meeting seemed to have gone well, the fact that we were going deeper in with Twitter underscored the fact that we were approaching a point of no return in terms of AdGrok investment.
“Look, we’ve got to figure out if we’re selling or what,” I said once we were out of earshot of the Twitter offices.
We were sitting at the picnic tables in South Park, the boys across from me. This was where Twitter itself was conceived in 2006, during a brainstorming session held on one of the park’s slides. The irony was striking.
After some awkward dithering, and lots of downcast study of the green tabletop, we finally got to talking. For probably the first time, I confronted the boys with the fact that we hadn’t shipped anything since launch almost a month before, and that the commitment from the technical side of the team seemed to be waning. Given the occasional wall between the technical side of AdGrok and everything else (i.e., between them and me), I wanted to confirm that they also had the same vibe.
They didn’t disagree with me.
MRM himself seemed checked out and hadn’t delivered anything on the new code front in weeks. Argyris and I had chatted about it, but so far all we’d done was to call him the mornings he was late to tell him to get his ass to AdGrok. Argyris was holding up his end, but the two had lost that wonderful mind-meld synchrony that had powered AdGrok’s development from the first days in our ratty Mountain View apartment. The dev team is the engine of a tech company. If they were done, then we were dead in the water. If that engine couldn’t be fixed back into productivity, then it was time to sell the company while we even could.
I looked from one to the other: they seemed tired and worried and done with the startup game. They agreed we should pursue the acquisition process to its conclusion. We had to sell AdGrok to Twitter, or else.