That Will Never Work

by Marc Randolph

  We laughed. And for a moment, the stress of the previous weeks melted away. We had a deadline, sure. We had people counting on us: investors to satisfy, employees to pay, and customers to reach. But when it was all said and done, we were a website that gave people access to DVDs. We weren’t changing the world like Reed was. We’d fix the Bloody Finger. But for the moment, it was okay.

  After drying off and having dinner—mussels in some kind of sauce, a fish that Steve assured me wasn’t really endangered, all washed down by a wine I couldn’t pronounce the name of—we headed to Steve’s home theater, adjoining the living room. It had been a while since I’d seen it, and he’d made a number of modifications. There were huge leather chairs with massive arms (and cupholders), each separate from each other. Each of them was nicer than anything in my house—and he had twelve of them! He’d installed track lighting in the aisle, just like a real movie theater. The screen was easily eight feet across and took up a full wall, and the projection system hung down from the ceiling. Steve pointed out speakers: tall ones on stands at the front of the room, two massive ones at the back, and a center speaker in the middle of the room that Steve said was just for dialogue. Steve then gestured at one of the seats, second row, slightly left of center. “That seat? That’s the money seat,” he explained. “Everything is balanced, faded, and toned so that it sounds perfect in that one spot.”

  Karen started up the popcorn machine just outside the screening room, and I took a look inside the replica candy case that stood by a refrigerator stocked with sodas.

  Mounds bars. My favorites.

  “So, Marc, you got a movie for us?” Steve asked, once we all had our concessions.

  “Sure,” I said, fumbling in my backpack until I found it. “I don’t know anything about it, but it just came in today. One of this week’s featured releases.”

  Steve saw the cover. “Oh, right, Boogie Nights! I remember hearing that was good.”

  “Worth a shot,” I said. I felt good: relaxed, full of wine and seafood and the reassurances of a friend. I sat down in one of the front-row recliners, next to Lorraine. Steve took the money seat, next to Karen. Reed took the row behind them.

  The lights went down, the curtain went up, and we watched Dirk Diggler let it all hang out in crystalline, DVD-quality resolution, across an eight-foot screen.

  At first I was horrified. Then I laughed until I cried.

  “Let’s hope your content team knows more about your inventory than you do,” Lorraine said.

  I had to agree.

  That night with Steve Kahn taught me a thing or two about the virtues of preparation. But I’ve learned most of my lessons on that subject outdoors—particularly, in the mountains.

  It’s definitely not a place that you can take lightly.

  There are river crossings, where a single missed step can plunge you into water that was snowmelt only hours before. If the cold doesn’t get you, then it will rush you downstream and stuff you permanently beneath a submerged outcropping or felled tree or, failing that, trap your leg in the rocks and bend you backward, buffeting you up and down until you finally lose the strength to hold your head above water.

  There are snowfields. To cross them, you have to step with enough force to forge a solid platform. But it’s entirely possible that once you’ve committed your weight to a step, your platform will give out without warning, leaving you sliding downhill at increasing speed, hoping that you will be able to arrest your slide with your ice ax before you plunge at high speed into the rock-filled moat that forms the boundary between snow and earth.

  There are cliffs. To climb them, you must make a pact with the rock, promising to linger on each hold only as long as it takes to move to the next one, the cliff warranting that the tiny edge of rock you have grasped and staked your life on will support your weight. Until it doesn’t, and with sudden and unexpected consequences, you’re hurtling down, your fall unobstructed until it’s broken by the jagged piles at the base of the cliff.

  There are dangerous animals like bison, cougars, and grizzlies; poisonous plants, berries, and mushrooms; you risk infections, lacerations, contusions, concussions, and dislocations. There are avalanches, rockslides, mudflows, and icefalls. There are blizzards, downpours, hailstorms, and sudden freezes.

  There are countless ways nature can tell you that you are unwanted, alone, and far from medical attention.

  But probably the scariest risk in nature’s repertoire is lightning. When you’re in the mountains, weather moves fast. One moment, the sky is clear and cloudless, and the next it’s dark, filled with angry clouds. Is there anything more biblical than a bolt of energy that comes crashing down from the clouds without warning? In an instant, lightning can turn a towering Douglas fir into a blazing birthday candle. And when you’re up high, it’s certainly no consolation to know that lightning aims for the highest point around—whether that be a tree, rock pile, sailboat mast, ice ax, or head. Lightning doesn’t discriminate based on your religion, your educational background, your sexual orientation, how much money you have, or how many pounds you can bench-press. All it knows is that you are out in the open, unaware, and, for at least that particular moment, the fastest and easiest way to move 10 billion watts of potential energy in a single release from turbulent cloud to the ground. If it has to go through your head, down through your organs, and out through the soles of your feet in order to do so…well, then that’s just your bad luck.

  To maintain your sanity in the mountains, you can’t dwell on these things. But the best mountaineers aren’t quite sane. I’m no climbing legend or anything, but when I’m at elevation, I’m always asking myself, “What is going to go wrong?” If I have to cross a stream, it’s only after I’ve hiked a few hundred yards downstream to see if there is anything there that might trap me should I lose my footing and be swept down that way. I’m looking for tree limbs on the bank I can grab, areas where the current eddies out into a gentle swirl, so I know what to swim toward. And as I start wading across the stream—or start my way across the log that spans the creek—I’ll have loosened the waist belt on my pack. It makes it harder to carry, but infinitely easier to shed should I need to swim.

  That’s what it’s like being in a startup. You spend a lot of time thinking about what might happen. And preparing for it. Sometimes you actually put a backup plan in place, but most of the time you just think through how you will respond—you scout out the rivers for rocks, check out the cliffs for things to grab onto if you fall. Most of the time, the worst doesn’t come to pass. But when it does…when the shit really hits the fan? Well, you’re going to be the guy with the pail and the mop. And wearing a raincoat. And that’s the kind of thing that makes the difference between being a success and being the guy who is covered with shit.

  Sometimes, as we learned on Netflix launch day, there is no difference. You’re both.

  On the morning of the Netflix launch, I woke up early—around five. Lorraine mumbled in her sleep as I quietly slid on my slippers and shut the door behind me. The kids would be up in two hours or so, but until then I had the house to myself. In the predawn darkness, I dodged hammers and granite samples in the still unfinished kitchen. It was the last room in the house to be remodeled, and we hadn’t gotten very far yet. The décor was straight out of 1971: fluorescent lights, avocado-green cabinets, peeling linoleum over the wood floors.

  There was still some coffee in the pot from the previous day, and after heating it up in the microwave, I drank it standing in the kitchen, feeling my mind boot up. I made a new pot, scooping ground coffee into a filter and pouring water into the reservoir of our coffeemaker. It was ostensibly for Lorraine, but I’d probably drink half of it before she got up. I’d need every bit of the caffeine.

  In the six months since Reed had written that check, we’d done so much—we’d assembled an inventory, put together a website, built a company with a culture. We’d worked tirelessly to make our dream of an e-commerce site for DVDs a reality.<
br />
  But up until now, it still had the feeling of an unresolved dream. The site existed for us—but not for anyone else. The problems we anticipated—and we’d racked our brains anticipating them—were still in the future. We weren’t even sure if we’d identified the right problems. The successes, too, were in the unrealized days and months ahead.

  There are a great many stages in the life cycle of a startup. But a tectonic shift happens on launch day. Before you go live, you’re in the dreamy zone of planning and forecasts: your efforts are provisional. You’re making predictions about what can go wrong and what can go right. It’s a very creative, heady sort of work. It is essentially optimistic.

  The day your site launches, something shifts. Your work now is no longer predictive and anticipatory: it’s fundamentally reactive. Those problems you anticipated? You didn’t know the half of it. Your planned solutions? They’re a drop in the bucket. And there are hundreds—thousands—of issues that you could have never even imagined, and now have to deal with.

  That morning, watching the sun rise over the mountains, I was positioning the various teams in my mind, imagining what the day would bring for Jim Cook’s crew, for Eric’s programmers, for Te and the marketing squad. I ran through the day’s plan: the 9:00 a.m. launch, the morning full of press calls, the process of order to shipment.

  In other words, I was doing what I’d been doing since the summer of 1997: strategizing. Before you launch, you’re making a beautiful battle plan, coordinating the future movements of your troops.

  The second you launch, you’re in the fog of war.

  I got to the office at seven or so in the morning and called our standard daily meeting. Christina, Te, Jim, Eric, and I filed into the conference room to go over the day’s schedule.

  “We’ve got press calls starting at nine,” Te told me.

  For months, Te had been lining up reporters and news outlets who would be interested in writing a story about our startup, hitting her Rolodex hard so that when our launch day came, people would read about it. All morning I’d be on the phone with these reporters, giving them a pretty canned speech that I’d spent hours trying to make sound natural.

  Here’s an excerpt:

  With this morning’s launch of the nation’s first internet DVD rental store, every DVD owner—no matter where he lives, no matter how far he lives from a video rental store—is now guaranteed access to every DVD title available—to buy or to rent.

  “Who’s first?” I asked.

  “Steve Perez at the Santa Cruz Sentinel,” Te said.

  Starting with the hometown paper wasn’t coincidental. My strategy is always to start with a softball. For your first call, there’s nothing like having a friendly voice on the other end of the line.

  (And in this case it paid off. Unlike the San Francisco Chronicle or Yahoo!, two of the other outlets that covered us, the Sentinel gave us prominent coverage, with a photo. Somewhere in my files there’s a faded full-page newspaper clip from the day after our launch, featuring a photo of a very late-nineties iteration of me, complete with a pager clipped to my belt, standing next to a Gateway and a mess of cables and wires. The lede?

  Still trying to figure out how to program your VCR? Trash it then. Videotapes are as passé as Grandpa’s Polaroids.)

  “Great,” I said, running over my lines in my head. I knew that whatever happened, I’d have to project cheerful calm through the mouthpiece of my telephone. Bombs could go off, the servers could catch on fire, and the site could crash—but I’d just have to close my eyes and keep going.

  Netflix makes it incredibly easy to rent a DVD. There’s no driving. No searching for parking. No standing in line. We even make it easy to return it. And we’re open 7 days a week, 24 hours a day.

  We went over, one last time, the process for Jim’s team.

  “Order comes in,” Jim said, “and once we’ve got credit card authorization, it goes to the printer in the safe. My team finds the disc, slips it into the sleeve, and scans it once to pull it from inventory. Then it’s off to Dan. Dan inserts the promo sheet, seals things up, slaps the labels on, and scans it again to show it’s shipping. Then into the bin and ready to mail.”

  Jim was still smiling that stupid grin, but I could tell he was nervous. He’d spent weeks streamlining his process, checking it for flaws and inefficiencies. But there was only so much he could do without the pressure of real orders coming from the site. And one of the big problems was that we had no clue how many orders we’d get on launch day. Five or ten? Twenty or thirty? A hundred?

  Corey had been working overdrive on the message boards, pumping Netflix up to tech nerds and cinephiles, and he’d continue to do so throughout the day. But how many orders was that? I wasn’t holding my breath for big numbers.

  Eric and his team—Boris, Vita, Suresh, and Kho—looked inscrutable. Whether they were nervous or not, I couldn’t tell. Most of the stress of the day fell on their shoulders, of course. They’d anticipated all sorts of problems with the site, and they’d formulated any number of solutions to those problems. But they knew that things would go wrong that they hadn’t expected, and so the day, for them, was going to unroll in a flurry of Mountain Dew and pizza slices. Eric barked out a few largely incomprehensible reminders to his team, and I took that as an opportunity to look them over. Boris and Vita appeared the same as ever, unflappable and calm. Kho looked like he’d dressed up for the launch: clean black T-shirt, somewhat clean-looking black jeans. His hair looked combed.

  Christina was nervous. She’d planned for this day for months. She had hundreds of pages, in dozens of notebooks, detailing the site’s operation—how a user would interact with it, what would happen if they made a mistake. Her team had spent many hundreds of hours integrating our own movie content with Michael Erlewine’s back-of-the-box data, building informative, interesting entries for all 925 films in our archive. I could see her team through the conference room window, still manually scanning the cover images of the last few DVD boxes to be uploaded onto the site. To them, it was just another day of their usual work. But to Christina, whose understanding of the website’s logistics far surpassed anyone else’s, it was a stressful day.

  “You know,” she told me, “this is our fifth launch together?”

  It was true. We’d launched a whole series of PaperPort scanners at Visioneer together. And each of us, individually, had dozens of other launches under our belt. But that was then. After all, in software and packaged goods, when the actual launch day arrives, you’re already past the point of no return. The product has been finished for weeks—it’s come out of the factory, it’s made its way into boxes, it’s on trucks going out across the country. Launch day is just a press day.

  “I feel like this is going to go somewhat differently,” I said as we filed toward the bank of computers in the middle of the office workspace.

  “I think you’re right,” Christina said.

  We had no idea.

  It started well. At 8:45, everyone in the office gathered in front of Eric’s computer. The site was going live at 9:00, and we’d already made the rounds of preparations. Was there paper in the printers? Were all the DVDs tucked in their sleeves in the safe? Were all the i’s dotted, all the t’s crossed?

  There were actually two versions of our website. One resided on a server that wasn’t online. It was a duplicate version that Eric could use to test out new pages and features. Anything new was posted first to what was called a staging server. Then we would bang on it for a while to make sure that it worked the way we expected it to, and, more importantly, that the new additions played nicely with the rest of the site. Then, once we had some satisfaction that we weren’t going to have a disaster on our hands, we would push the new version over to what was called a production server, which was hosting the live site.

  Up until this morning, the distinction between the two sites was entirely academic. Although one was supposedly final and connected to the internet, it wasn’t visible t
o the actual public. Although we had practiced pushing things live, and pretended we had real customers using it, there weren’t any real consequences. This was all about to change.

  For the hundredth time, Eric idly scrolled through on the staging site, pretending to be a customer. “It looks good, it looks good,” he said, clicking on the links and filling in fields on our forms. Boris and Vita were acting nervous, too. They knew—as we all did—that things would break, and that they’d have to be on their toes to fix the site when it invariably malfunctioned. They’d planned for things to go wrong. What happened if a user entered his state abbreviation as NF rather than NC, ND, NE, NM, NV, or NY on the checkout page? What would happen if the credit card number didn’t start with a 4 (for Visa) or 5 (for Mastercard), or didn’t go through at all? Would we fail gracefully, or crash and burn?

  One final thread that I knew was still sticking out of the seams of our startup was the confirmation email. We hadn’t yet created an automated confirmation email function for users, one that would contact a customer after she placed an order and reiterate information about payment and shipping. We’d have to compose confirmation emails by hand for each individual customer. That wasn’t ideal, obviously, but I figured it would be workable.

  “Five minutes,” Christina said at 8:55. She was drinking coffee out of a huge mug and munching on a scone. That’s how I knew she was nervous—a gym rat like her usually stayed far away from buttery pastries.

  “How are the nerds?” I asked Corey. He’d been on the forums all morning, reminding some of the heavier users about the Netflix launch.

  He shrugged. “Hard to tell. I think they’ll show up, but who knows how many.”

  Jim had his hands on his hips. I could see his mind going through the logistics of shipping, replaying over and over how to fill an order, pack it, and store it until 3:00 p.m. That’s when the orders had to hit the post office in Scotts Valley, to ensure that they’d ship that day.