Going from “I” to “We”
A CEO’s Search for her CTO
“There’s one for each company.”
Our professor pointed to a stacked wall of still-in-box IKEA filing cabinets. “You have 25 minutes to assemble yours as a team — starting now.”
Eight weeks earlier, I had applied to the d.school’s “Launchpad” Accelerator class with a key question sparked by my consulting and design career:
What if I could build a tool that would enable anyone to communicate with images as easily as they do today with words?
At the GSB, this idea engulfed my identity. I became “the one who takes those visual notes.” Building on a decade of painting and Bain slide-making, I’d developed this verbal-to-visual system — and people had started to notice. To prove it (and to get accepted to Launchpad), I offered my visual notes for sale. Two days, 500 sets, and $2500 later, I had earned a spot in the class to start my company — Stick.
On that Tuesday afternoon at the d.school, though, I could feel the flood of failure even before they called “Time!” As the only solo founder in Launchpad, I was surrounded by other teams who had functional filing cabinets ready to be filled, while I had an askew rhombus ready for a dumpster.
Yes, I’d made progress on my own — making prototypes of my product and outsourcing development work to Russia— but if I wanted to build a tool with a billion users during my lifetime, I needed a technical cofounder. Now.
I didn’t realize that it would take me two years, three months of coding bootcamp, a hundred events, over a thousand emails, 554 candidates, 149 interviews, and enough coffees to kill me.
My first step was to go to Gates, the center of Stanford’s Computer Science Department. I was used to my Stanford email address opening doors, but in Gates, my MBA was a scarlet letter. So I edited the “GSB 2016” out of my signature before emailing every Natural Language Processing PhD. I snuck onto the Human Computer Interaction (HCI) Slack channel to sleuth the latest class projects. I cornered the CS448B professors at office hours. But each time I pitched the idea of an AI-enabled text-to-visual communication tool, I got:
“We’re years away from that working.”
“Researchers have been working on cracking this question for decades. What makes you think you could do it?”
If the best computer scientists in the world thought my idea was an unsolved (unsolvable?) problem, I’d learn to code so I could prove them wrong.
Months later, halfway into my no-boys-allowed coding bootcamp at Hackbright, I still believed in my idea. I also understood why the experts had told me it was impossible. As an (amateur) engineer, I spent 10% of my time creating code, and 90% debugging it. I learned that engineers have to find “bugs,” but entrepreneurs have to ignore them — to keep believing in their vision. Or else… failure. I think of Ben Horowitz’s example: When you drive a race car, he says,
“They tell you to focus on the road when you go around a turn … because if you focus on the wall, then you will drive straight into the wall.”
I needed a technical partner who could focus on the walls and believe in the road — at the same time. No wonder this was difficult.
So, I set out on the next chapter of my search. After 73 unsuccessful Angel List interviews, I turned to another piece of advice I’d heard: “You don’t need a genius to get things started.” Maybe my bar was too high, I thought. So when the 74th (initially unsuccessful) candidate asked if he could just work for free, I said, “Fine.”
To build a prototype of our product, my volunteer engineer and I began working on a (painfully) simple Slackbot that provided visual slide layouts for Product Managers.
A few days later, as I was presenting my latest progress to my Hackbright cohort, I started getting notifications from Slack: “Norrisbot has been invited to #general.” Who is Norrisbot? Our Slackbot’s name was Stick. Oh well, I’ll figure it out later, I thought.
But I didn’t have time. As I was setting up my slides, Slack notifications started engulfing my computer by the dozens:
Google won’t search for Chuck Norris because it knows you don’t find Chuck Norris — he finds you.
And then, by the hundreds:
He who lives by the sword, dies by the sword. He who lives by Chuck Norris, dies by the roundhouse kick.
There is no Esc key on Chuck Norris’ keyboard, because no one escapes Chuck Norris.
… and so on, until an unbridled Chuck Norris Slackbot crashed Slack — and my computer — with its thousands of messages.
It wasn’t my “volunteer’s” fault — he was a new engineer, after all. It was mine. By giving into my desperation and accepting “free” work, I had burned a month of progress with low quality code.
Free work isn’t work. It’s a favor, and favors don’t build companies. It was time to pull the plug and keep on.
Over a year into my search, I was running out of time — and sanity. Potential investors wanted to see a technical cofounder, and with some pitching opportunities coming up, I needed to beef up the “Team” slide in my pitch deck.
And beef up I did. I had a Stanford Physics PhD building our visual language database with the latest neural network style transfer methods, a Stanford undergrad in Human Computer Interaction working on User Experience (UX) research, and a former student of mine working on the Slackbot (yes, it survived Chuck Norris).
Even though I was spread thin, I felt good about pitching Stick at Venture Studio’s Demo Day. Oberndorf was packed with hundreds of investors from Accel, Sand Hill Angels, and the like swarming the dozen demo tables lining the room.
After our pitch, the newest team addition — the Stanford undergrad — was sitting with me at our company table when an angel investor walked up and asked him, “So, you’re part of Stick!” My undergrad laughed a bit, then responded, “I mean … not exactly?”
Progress may be hard to define in a “soil-stage” startup, but I knew this wasn’t it. I thought I was building a company, but this blue-chip-Frankenstein team wasn’t going to bring my billion-user beast of a vision to life. Time to fire the volunteers and keep on.
I kept on — through candidate after candidate, interview after interview— even though each batch seemed more absurd than the last. The real Silicon Valley was proving to be more entertaining than HBO’s Silicon Valley.
Consider Carl, who responded to an interview scheduling email with a one-liner: “Absolutely slammed dealing with my Burning Man camp at the moment. Ping me in a couple of weeks.”
Or Hamilton, who submitted an “intro video” along with his application for the Head of Engineering position on our job site. The URL he submitted? http://will-send-you-a-video-if-you-are-funded-and-planning-on-paying-a-competitive-salary-please-let-me-know.
Or Yu, who replied to one of my email blast pleas for a partner with, “Hi, Alli. This is a one afternoon project.” If only I had met you earlier.
In the beginning, I felt frustrated. Half a dozen later, I started a “Happy” folder in Gmail to store them all, just in case HBO ever needs more material. And I kept on.
After a two year search, I expected some kind of firework display when I finally met our Future CTO, but that wasn’t how it happened. Andrea slipped into my Applicant Tracker Excel just like 553 others had, and sounded just as pleasant on the phone screen as most did. It was not an epiphany, but a slow unfolding as I appreciated the match we’d found.
He had the capability, the character, the vision. He grilled me in the interview. I thought I was interviewing him for a job, but I was being interviewed for a partnership.
Once I knew this was it, I had to earn it by doing my #1 job — selling. I was on the brink of going from “I” to “We” — and We needed contracts. As I started selling this “product currently in development” with screens I’d built in Sketch, it worked. We had our first big customer — and Stick had a CTO.
So, what worked?
I kept on.
I’ve done some work to find a cofounder — from accidentally winning the world’s first Emojicon costume contest as the (hand-painted) Peach Emoji and ending up in the New York Times, to accidentally clicking the button that emails all 3,147 of your LinkedIn contacts at once — including my mom’s law associate who replied asking, “What do you mean by CTO? Certificate To Operate?” I’ve felt physically ill with fatigue, embarrassment, and frustration. But I didn’t stop. That’s the only thing I could control.
If I measure myself on my CTO search efficiency — on how quickly I could build the cabinet — I fail miserably. But if we measure intention and persistence, I think I’ve done pretty well. Instead of a shoddy IKEA filing cabinet that’ll last a few months, we’ve got a sturdy oak chest, ready for the long haul — even if we are still rearranging the drawers.
