That Will Never Work

by Marc Randolph

  “Well,” I said, pantomiming a toast. “Shit.”

  I paused, taking in the absurd particulars of the scene: the Lear’s leather interior, Barry’s billowing Hawaiian shirt, the tray of fruit big enough for a family of five. I smiled, feeling resolve flood my chest.

  “Blockbuster doesn’t want us,” I said. “So it’s obvious what we have to do now.”

  I smiled. Couldn’t help it.

  “It looks like now we’re going to have to kick their ass.”

  17.

  The Belt Tightens

  (2000–2001)

  AMONG CLIMBERS, IT’S A matter of dogma that if you haven’t reached the summit by early afternoon, it’s time to think seriously about turning around. As noted Everest climbing guide Ed Viesturs tells his clients, “Getting to the top is optional. Getting down is mandatory.” When you’re thousands of feet in the air and miles from camp, you have to leave yourself plenty of daylight—otherwise, you can get stuck just where you don’t want to be.

  That’s what happened to Netflix, metaphorically speaking, in the fall of 2000. After Blockbuster refused to buy us, Netflix was in a no-man’s-land: we were no longer in grave danger, but neither were we entirely out of the woods.

  Unlike many companies our size (and larger), we’d survived the dot-com crash. Our business model was good: no-due-dates, no-late-fees was working. People loved Cinematch. We were on track to acquire 500,000 users by the end of 2001.

  But our subscription model was fundamentally expensive. We were still hemorrhaging money, and the landscape we found ourselves in was a far different one than we’d encountered a year before. Burning cash at the rates that had been normal the year before now looked irresponsible. We needed to speed things up. We didn’t have to become profitable, necessarily. But if we ever wanted to go public, banks (and the investors they recommended our stock to) would have to see a path to profitability. If we were still in the cycle of raising $40 million every year, and then posting $45 million in losses, we wouldn’t look like a very good bet.

  We knew that if we wanted any chance of surviving in a post-bubble world, we had to be ruthless in our dedication to the Canada Principle. Over the course of late 2000 and throughout 2001, we streamlined our processes. The “.com” suffix that had been a ticket to free money in 1999 was now an albatross—so we cut it from our name. The “portal” ethos that had ruled the world (and the boardroom) a year before had crashed and burned alongside Dr. Koop—so we shelved it.

  We called this ruthless streamlining scraping barnacles off the hull.

  Companies are like boats: sometimes you have to put them in dry dock to remove the barnacles that have accreted on the hull, slowing down forward progress. In the wake of the Blockbuster fiasco and the crash of the dot-com market, we took time to self-assess, then mercilessly pruned back all programs, tests, additions, and enhancements that weren’t contributing anymore.

  We’d always done this. And it hadn’t always been easy. Sometimes, what looks like a barnacle to you is someone else’s favorite feature. For instance—when we were determining the price point for Marquee, we had tested dozens of prices and DVD quantities. We’d had some customers getting four discs at a time for $9.95, others $19.99, and still others $24.95. We’d allowed people to keep two discs at a time, eight, and every number in between. And while the standard plan was to swap as often as you liked, some customers were limited to just a handful of swaps. We were eventually sued over one of our more interesting experiments, in which we expedited service to some customers to see if that encouraged the light renters (“birds,” in Netflix parlance) and artificially slowed down service to discourage heavier users (internally—and secretly—called “pigs”).

  All of this testing was undeniably useful. Because of it, we didn’t have to argue about whether higher prices slowed down sign-up rates, increased churn, or encouraged higher usage. We had proved whether it did each of these things and knew exactly by how much. But once we’d learned what we needed to learn, the usefulness of the test fell to zero for us.

  Unfortunately, the cost stayed the same. And each new feature we added on to the service had to then work flawlessly with the old ones, so as to accommodate all our customers, no matter which plan they were on. Doing so meant that design got complicated. Testing got harder. And everything slowed down.

  An outdated feature was a barnacle. The drag it imposed might be tiny—but multiplied by a thousand? It slowed us down and cost us money.

  So we scraped. Every meeting, before we could launch into any discussion of what we planned to do going forward, we had to start by looking backward, coming up with a list of what we could decide to stop doing. It wasn’t easy. Most times, deciding what not to do is harder than deciding what to do. Yes, the customers who were on the $24.95 plan were delighted to have their test terminated and get switched onto the $19.95 group. Not so much the lucky few paying $9.95 for the same service. Or getting eight discs at a time.

  After a while, we got pretty callous about it. Let ’em scream, we would rationalize. We’re okay with upsetting a thousand if it means we get it right for ten thousand.

  As 2000 slid into 2001, with the Blockbuster opportunity steadily receding in the rearview mirror and the idea of an IPO shelved for the foreseeable future, Barry was furiously scraping barnacles from every aspect of the business, desperate to make our boat go faster.

  At first, the items on the cut list were easy to tackle. If we weren’t going to go forward as a portal, we didn’t need to develop technology to serve up advertising on our pages. Christina and her team could drop their efforts to develop a showtimes feature, and the content team no longer needed to assemble data for every movie on the planet—we could just focus on the DVD catalogue.

  But we didn’t need Barry’s spreadsheets to tell us what was becoming increasingly obvious to everyone who sat around our E-Staff table every Monday: we had more people than we needed.

  Under ordinary circumstances, being a little heavy would have been okay. At the rate we were growing, we could have easily absorbed extra employees by waiting a quarter or two, until the added volume and complexity of the business both warranted and economically supported having more staff. But things were different now. As Barry worked through the numbers and confronted the new post-crash landscape, it was clear that we had to be not just a lighter company—we needed to be a different one.

  In a post-bubble landscape, we couldn’t look like a money pit. There had to be a point where not only were we making money from each customer each month, but there were enough customers to cover the fixed costs of running the business. In the past, we had mostly focused on only one side of the equation: getting more customers. Now, it was becoming increasingly obvious that we had to focus on the other side of the equation: spending less money running the business.

  There were only so many costs we could cut by eliminating big things. And we were running out of barnacles to scrape. We had a clean hull and the destination was clearly defined, but the boat was still too heavy. If we were going to make it to shore, we’d need to lighten the ship.

  Our usual Tuesday E-Staff meetings always began with the same agenda item: Who fucked up? That wasn’t the official name for it, of course—but that’s what I called it. In the interest of transparency and radical honesty, each of us would go around the room and say something that wasn’t working. We didn’t need to know what was going well—things that were working didn’t need any attention from the rest of us. Instead, we wanted to know what wasn’t working. Who, in my elevated parlance, was fucking up.

  It’s a maxim of startup life: You’re going to get things wrong. You just don’t want to get the same things wrong twice.

  At one meeting in the summer of 2001, just after the morning’s ritual flogging was complete, Reed gestured to Barry that it was time for new business. Barry stood up from his seat and walked to the whiteboard, grabbed a green marker, and wrote in large letters: 2,000,000.

  “That
’s the number,” Barry announced, turning toward us. “At our current overhead levels, that’s how many subscribers we’ll need to get to profitability.”

  He leaned down and squinted at his laptop.

  “But that’s seventy-three weeks away. And every month between now and then, we’re burning money. We’ll run out long before we’ll get there. And I don’t need to remind anyone that people aren’t exactly lining up to throw money at our feet.”

  Barry paused, then squinted down at his laptop again. “We have to cut our expenses. And deep. We have to be lean enough to make it to profitability on the money we already have. And the only way we do that is by lowering our costs so that we can be profitable on a smaller subscriber base.”

  He picked up the green marker and turned back to the board. With the edge of his palm he erased the 2 at the head of the number and replaced it with a 1.

  “We’ll only survive if we can be profitable on one million subscribers. And this?” he said, opening a manila folder and sliding a stapled set of papers around to each of us. “This is how we’re going to do it.”

  Layoffs. That was Barry’s plan.

  After his dramatic reveal, Reed, Patty, Barry, and I met for lunch every day, chewing through dozens of scenarios. Which departments should we cut to the bone? Which should survive unscathed? Should we cut highly paid (but more valuable) employees or make it up in volume, by drastically reducing the ranks of customer service agents?

  These were complicated questions. We needed to cut expenses dramatically, but we had to do it in a way that wouldn’t damage our ability to grow our business.

  After a particularly grueling session, I swung by Joel Mier’s desk and tapped him on the shoulder. Joel was my director of research and analytics. In his role he had to straddle art and science, data and intuition, and his appearance supported that. At 6′4″, he was a powerful presence, but he held himself in a gentle way that made him approachable. In an office full of shorts and unwashed T-shirts, Joel dressed like a university professor: button-down shirts, cardigans, corduroy pants, and black oxfords. He was extremely measured in his speech and chose his words carefully. He was an even more thoughtful listener, nodding slowly and considerately at even the stupidest utterances, as if the failure to find some genius in them must lie with him.

  But his professorial manner hid a sharp wit and a hooligan-in-the-back-row sense of humor. Joel loved a good practical joke—he was a consummate player of Coins in the Fountain and loved nothing more than to stock the Netflix kitchen with half-edible delicacies. He’d once put out a bowl of freeze-dried garbanzo beans that had prompted an enraged Mitch Lowe to scream: “You almost broke my tooth, motherfucker!” I still laugh when I think about it.

  Joel’s combination of data-driven intelligence and sophomoric humor had made us fast friends from day one. It didn’t happen often, but whenever we could, Joel and I liked nothing better than to sneak away at lunch to a booth at the Black Watch, the sole dive bar in Los Gatos, to carefully dissect our colleagues, our heads collapsing helplessly in laughter onto the beer-stained tables.

  “How’s it hangin’, boss?” Joel murmured now, barely looking up from his screen.

  “Up and to the right,” I said, giving the ritual response to our usual greeting. I flipped my head toward the stairs. “Let’s go for a walk.”

  Earlier that week, Reed, Barry, and I had brought all the Netflix directors into the cone of silence and shared our intentions with them. Since the majority of the employees in the company directly reported to this group, they had the best idea of who was actually getting things done, who was irreplaceable, and who could be easily done without. I had gotten in the habit of taking daily walks with most of them, making lazy circles around the building as we talked out loud about each person in the department. I needed their help, because deciding whom to let go was complicated. Talent and indispensability turned out to be the easiest dimensions to figure out. It was the other stuff that got us tangled. How much weight should be put on personal situations? What if an employee were the sole breadwinner in her family, with a new baby? Should she stay on rather than someone young and single (but more talented)? And what about the handful of married couples working at Netflix? Would it be just too cruel to fire both of them at the same time?

  Joel was pulling on his jacket as he pushed open the building’s front door with his hip and joined me on the sidewalk. Without saying anything we turned and began the loop, circling clockwise around the building.

  “I’ve been thinking, boss,” Joel began, as soon as we were safely out of earshot of an employee getting out of his car.

  “That’s always dangerous,” I replied.

  Joel smiled and continued.

  “I know we have talked about doing this LIFO, but I’m not feeling it.”

  LIFO stood for last in, first out, meaning that the last people hired, having the least tenure, would be the first ones to be let go. We’d borrowed the term, which usually applied to inventory. Although LIFO, for our purposes, didn’t necessarily correlate to talent, it did correspond to experience on the job, and to some degree it made people feel like there was a rational basis to a process that could often feel pretty random.

  “It’s Kyle that I’m worried about,” Joel finally blurted out. “He’s definitely not next in line to go when you use seniority. But his attitude…” Joel left the thought hanging.

  I knew exactly what he meant. In one of the analytics meetings that we periodically held, where we brought together people from different disciplines in the company to wrestle with particularly challenging analytical problems, Kyle was consistently proving himself to be…how shall I put this…difficult.

  At Netflix, there was nothing wrong with disagreement. In fact, disagreement was a critical component of our culture of radical honesty. We expected disagreement, because we encouraged vigorous debate. In Netflix meetings, there was no seniority, and no one’s opinion was more valuable because of their title, age, or salary. Everyone was expected to fight for their point of view until a consensus had been reached.

  Still, no matter how passionate the argument, there was a shared expectation at Netflix that, once the self-evidently correct conclusion had been reached, it was time to fall in and implement it. Disagreement was collaborative, not ego-driven. It didn’t matter who was right—all that mattered was that we got it right.

  And that was where Kyle was falling short. He just couldn’t get over it when things didn’t go his way, and his bad attitude was infecting everyone.

  “I get it. He’s out of here. We’ll keep Markowitz,” I said to Joel, referring to one of Joel’s direct reports.

  “Good,” he said flatly, without looking up at me.

  I’d known Joel long enough to understand that there was something on his mind. As we turned the corner and passed the picnic tables that sat in a small courtyard to the side of the building, I finally figured it out.

  “By the way,” I said quietly, “if I haven’t already made it clear…you’re safe.”

  The look of relief on Joel’s face was immediate. He nodded and with a big smile said, “Up and to the right, boss. Up and to the right.”

  Barely a week later, we were once more gathered around the table in the conference room, but the crowd had swelled to include the entire management team. It was a Monday night, just a few minutes before 8:00, and people were still shuffling in from dinner, looking for places to sit or stand. Outside the room, the rest of the office was deserted, the clusters of empty chairs and cubicles a somber premonition of what would be happening in less than twenty-four hours.

  I caught Joel’s eye from across the room and he gave me a curt nod. Patty sat at the front of the room, two white binders open on the table in front of her. Reed hovered over her left shoulder, looking down intently and gesturing at something. Patty whispered something back, and with her pen drew a line through a row of type halfway down the page.

  One major player was already gone: Eric Mey
er. He’d been nudged out the day before, his talents—although prodigious—no longer right for the challenges ahead of us.

  As for everyone else? After two final marathon sessions over the weekend, we had come to our final list. Now it was time to act.

  Patty looked up from her notebooks, pushed up her sleeve to check her watch, and loudly pushed back her chair. “Okay, boys and girls. Here’s how this is gonna work.”

  The email came out right on schedule, at 10:45 on Tuesday morning. It was short and to the point: there would be an important announcement at 11:00 in front of the building.

  The company had long since swelled past the point where we could fit everyone in a single room. Not even the lobby could hold us. Now, when we needed to assemble the entire company, we had to rent the 100-year-old Los Gatos Theater on Santa Cruz Avenue or just gather like a mob outside, in the picnic area by our front entrance. Today, it was the courtyard. There was no sense renting a theater to tell 40 percent of your company that they were losing their jobs. That would just be cruel.

  As I walked through the rows of cubicles on my way to the stairwell, I instinctively stepped into the conference room at the head of the stairs. This would be where my departments would be doing their layoffs, and I thought I should do one last check to ensure that, at the very least, there were two chairs in there. Getting fired was bad enough, I thought—you should at least get a chair to sit in.

  But it was okay. Two chairs. An empty table, a clean whiteboard. Nothing to indicate that there wouldn’t be a regular meeting there.

  Outside, there was already a big crowd gathered in the courtyard. People stood in small groups, nervously talking. I found Joel and stood stoically next to him. A few minutes after 11:00, Reed climbed up onto one of the picnic tables. The crowd quieted.