That Will Never Work

by Marc Randolph

  It was a funny moment, but as Patty and I headed back to work, we couldn’t help but wonder: If we were supplying our employees with fine dining, a fitness center, and an Olympic-size pool and they were still complaining, what are the factors that really drive employee satisfaction? Or more importantly, what does it take to get other people to sign on to help you with your dream—and be happy doing it? What we found was surprising. And surprisingly simple.

  People want to be treated like adults. They want to have a mission they believe in, a problem to solve, and space to solve it. They want to be surrounded by other adults whose abilities they respect.

  Years later, Patty would end up revolutionizing the field of HR at Netflix, and much of her philosophy can be traced back to the realization we both had that day at Borland: People don’t want hot tubs—not really. They don’t want free snacks or Ping-Pong tables or kombucha on tap.

  What they really want is freedom and responsibility. They want to be loosely coupled but tightly aligned.

  Following one problem through its various guises and iterations should give you a good idea about how we functioned pre-launch. This particular problem brought us into contact with another young company, one with its own unique culture—a culture that couldn’t have been more different from ours.

  When we started acquiring DVDs, there were only 300 or so titles available. By the time we launched, in April of 1998, there were 800. The relatively low number of DVDs in existence was, in some ways, a boon for us—it meant that we could feasibly acquire a couple copies of each one and accurately claim that we carried every DVD that had ever been released.

  But the relatively small DVD library was also a liability. Many of the titles available weren’t exactly blockbuster material. Sure, some of the studios were producing DVDs people might have actually wanted in 1997—The Birdcage, The Mask, Se7en. But the vast majority were decidedly less upmarket. It was a total grab bag: National Velvet next to Free Willy, Elmo Saves Christmas next to Sports Bloopers Encyclopedia. There were low-budget documentaries about trains, NASA, and World War II. There were dozens of nature films, and “magazines” with short video articles. Bollywood was big. So were karaoke videos, and concert films of orchestras and marching bands.

  Basically, the selection was all over the place. No one really knew if DVD was going to take off, and how it would work if it did. Would DVDs be a format used mostly for movies? Were they a good technology for listening to music? Would people want to use DVD five-channel sound to watch two-hour live performances by the USC marching band in their home theaters?

  So manufacturers and studios were kind of just experimenting. The library was a mix of somewhat newish releases, established classics, forgotten oldies, and discs that seemed designed to show off home theater equipment. If you looked at the library, you’d think that the audience for DVDs was mostly nerds, college sports fanatics, and anime fans.

  The distribution was similarly idiosyncratic. VHS was still king, so many distributors weren’t even carrying DVDs. And I don’t blame them. Why would anyone buy a library of things for which there was not any demand? As a result, DVDs became the province of the fringe distributors. These distributors were flaky. They didn’t return phone calls. In our quest to acquire at least two copies of every DVD, we’d spend days tracking down a title, only to play phone tag for a week.

  Through all of this process, Mitch Lowe was invaluable. He knew how to deal with distributors, even the small, elusive ones. He knew how to make them call us back. He was charming, persistent, and happy to call in the many favors he’d earned after his years as VSDA chairman.

  One of the most valuable skills he had was knowing exactly how many copies to buy. In those early days, there was no algorithm, no equation—just Mitch. He knew when to buy three copies and when to buy thirty. Two copies of every DVD was our minimum, but we bought many more for titles that Mitch thought would be in demand—far more than a video store would have for a comparable VHS title.

  The question of inventory was one that we thought would give us an edge over the video stores. There was a finite amount of space in a brick-and-mortar video store, so part of their job was learning how to deal with the inevitable shortages. Mitch and others in the business—most notably Blockbuster—called this managed dissatisfaction. What do you do when a customer comes into the store looking for a copy of Die Hard 2, but all of the copies are checked out? You try to rent them something else that they might like just as much. You try to make them happy enough so that they’ll come back. But even then, customers leave with a bad taste in their mouth.

  Since we weren’t tied to a physical location, we thought we could avoid managed dissatisfaction entirely. The closest thing to instant gratification that we could give our customers was always having the title they wanted in stock. So while a video store would have bought 5 or 6 copies, we bought 50 or 60. When L.A. Confidential was released, we bought 500. It was expensive, but we had four things going for us.

  First, $2 million in the bank. Let’s see Ye Olde Video Closet at the northeast corner of Boring and Main compete with that!

  Second, the cost of all those extra discs wasn’t expensive inventory, it was inexpensive advertising. It was like those old Mastercard commercials. A disc: $20. A reputation for having every DVD and always being in stock: priceless.

  Third, we were counting on the DVD market continuing to grow. What seemed like ten times too many copies today might very well be just what we needed when the market was ten times as big.

  And lastly? Well, if we really screwed up, we could always sell our extra DVDs.

  While Jim tried to figure out a way to squeeze all those DVDs into the bank vault, Christina and I wrestled with a different problem: How would our customers find a DVD? What search terms could they use? How would we group movies on our site? What information would they use to make their selection?

  When you rent DVDs, you have to have information about them. Think about the back of a DVD case: there’s a synopsis, a list of the cast, the director, and the producers, a few well-chosen (and often misleading) quotes from leading critics praising the movie you’re about to watch. In a video store, that stuff is all in place long before the movie hits the shelves. Sure, places like Blockbuster or Video Droid had searchable in-store databases for actors, directors, genres, and the like. But most of the information a customer wanted was right in front of them—all a potential renter had to do was walk in, see the picture of Tom Cruise on the cover of Mission: Impossible, and flip the VHS case over.

  We had a harder road. In our store, there was no case to flip over. What we were offering renters wasn’t just an alphabetically ordered trove of movies, separated by genre. We wanted the ability to filter: for director, for actor, for genre. We wanted renters to be able to follow their interests in a precise way. Did they like movies in which Andie MacDowell appeared? Were they really into the cinematography of Néstor Almendros, who was behind all those beautiful sunsets in Days of Heaven? Did they want a horror movie…and not only a horror movie, but a vampire movie…and not only a vampire movie, but a vampire movie with comedic elements?

  We wanted to be able to offer renters the ability to find exactly what they wanted. That required an immense amount of data—both objective data, like director/actor/producer/date of release, and subjective data, things like genre and mood. We also needed to be able to access data about awards and critical acclaim (or lack thereof). If a renter wanted to watch every Best Picture winner of all time, we wanted her to be able to do so.

  How could we go about building such a database? The obvious answer was to hire a few people to comb through the list of DVDs and pull data for all the categories we wanted. There were fewer than a thousand DVDs available, so it was possible.

  But we only had twelve employees. We had more than enough to do, so our time was a precious resource. But money? We had a bunch of that. So I started looking for other options. My thinking was that if we could buy it rather than bu
ild it, then we should buy it.

  Big mistake.

  Michael Erlewine is, according to his current Wikipedia page, “an American musician, astrologer, photographer, TV host…and Internet entrepreneur” who founded, in 1991, the All Music Guide (now known as AllMusic). Back in 1998, all I knew was the last part. I’d come across All Music Guide’s sister site, All Movie Guide, when I was looking for possible data sources.

  I didn’t know that Michael had hitchhiked with Bob Dylan, or that he’d been in a blues band with Iggy Pop. I didn’t know that he was the author (at the time) of five books on astrology, with titles like Tibetan Earth Lords: Tibetan Astrology and Geomancy. I just knew that he had something I wanted: data.

  I had something he wanted, too. DVDs.

  The goal of All Movie was to compile a detailed catalogue of every film ever made. Erlewine had dozens of employees tracking down, watching, and annotating movies around the clock. Visitors to his website could find the most arcane information about any movie they had ever heard of—and thousands they hadn’t.

  The problem was that he had no DVDs. An inveterate completist, Erlewine didn’t just want detailed information about every movie ever made. He wanted it about every format of every movie ever made. For DVDs, he wanted to know what special features were available, which languages were included, the screen ratios, and whether or not there was 5.1 surround sound.

  All the information he needed was right there on the DVD case, but distributors weren’t really selling to private consumers. So Erlewine’s problem was even more profound than ours as we tried to build out our inventory. Because there were so few retail stores that even carried DVDs, acquiring a library of every title would require thousands of hours, hundreds of miles of travel, and a good slug of chance.

  We began discussing a deal. We’d give him money—and DVDs—in exchange for his data.

  I love negotiation, and I’m pretty good at it. In large part this is because it’s easy for me to identify with other people’s needs. I’m able to understand what another party in a negotiation wants, what they need—and how they feel about getting it. Because I can quickly identify what the other side wants, I’m able to strategize more efficiently about how we can come to a mutually beneficial solution.

  With Erlewine, though, things didn’t go so smoothly over the phone. I knew what he wanted, he knew what I wanted, but I had a harder time understanding his reluctance to meet in the middle. Making a deal would help both of us, and our phone calls were cordial—but he kept stalling. He seemed reluctant to come to an agreement, and I couldn’t figure out why.

  So I flew out to meet him in person. One Tuesday that winter, I got on a plane bound for Grand Rapids, Michigan. When I landed, I rented a Subaru and drove north, to Big Rapids. I was expecting an office park, a corporate HQ, maybe even some good-sized rapids behind the building. But the address Michael gave me belonged to a big house in a residential neighborhood, with not a ripple in sight. Pickup trucks were parked in every driveway. Flanneled men were shoveling snow from their yards. I parked in a circular drive, in front of a three-story colonial that had been modified and joined to several other houses in the immediate vicinity. Covered breezeways connected several of the adjoining houses. I saw people hurrying from building to building, carrying cardboard boxes and, in one instance, a reel-to-reel projector. It looked like a commune, maybe a cult.

  I was a world away from our sterile office park in Northern California.

  Michael gave me a tour of the premises. At the time he was lean and wiry; undoubtedly from a diet high in kale, yogurt, and granola. He wore an open-neck shirt that showed some kind of necklace. He was direct, soft-spoken, and leaned in when he made a point, listening carefully to what I said in response. I expected at any moment for him to shake his head slowly and murmur, “Ah hah, just what I thought. Taurus with Aries ascendant.” But then he could shut off quickly if things weren’t going the way he wanted. Conveniently, he seemed guided by an external force, a higher power that was actually in charge. Negotiating a compromise was not something he could do, since he was, in fact, only the messenger. He would have to check with his boss before deciding.

  If the exterior of the All Movie HQ resembled a commune, the interior was like being inside an OCD record collector’s head. Every square inch of wall space was covered by floor-to-ceiling shelves filled with LPs, compact discs, and tapes. The rooms had clearly been improvised into working spaces. The first room we poked our head into had three desks, one in each corner. The woman at the desk closest to the door was holding the liner notes to an LP up to a small lamp, a foreign-language dictionary open in front of her. “Norwegian folk music,” she said.

  In the next room, a man was going through a huge stack of Daily Variety from the 1930s.

  “What are you looking for?” I asked.

  “Filming announcements,” he said. “I’m trying to correlate shooting dates.”

  “He’s capturing data on films that never even got made,” Michael said proudly.

  The tour took nearly an hour. I saw three or four buildings, all filled with these detail-obsessed scriveners. One building, a former garage, was particularly loud. “That’s the woodshop,” Michael said, opening the door. Saws, pallets, stacks of lumber. And dozens of identical six-foot bookshelves.

  They acquired so many books, records, and films that they manufactured their own shelving.

  We didn’t reach a deal that day or in the weeks that followed. We parried over the silliest details. As soon as I had met one of his demands, he would come up with another one. In one sense, I knew exactly how to appeal to Erlewine. For all his sites’ rhetoric about being the world’s leading repository of information about music and movies, my visit to Michigan had taught me something valuable: Erlewine was a hoarder. His real motivation wasn’t information: It was collecting. He’d cannily found a way to monetize his obsessions, and he needed my DVDs more than he needed my money.

  But even with DVDs in hand, I suspected that there was another reason Michael was holding out: He was paranoid about his data. My proposal was that we would use his movie data—release date, cast, and so on—as the bedrock for our own entries. We would append all the DVD-specific information and send that back to him. But Erlewine insisted that he be the one to add the DVD data and then send it to us.

  I liked the idea of him doing more of the work, but the deal breaker was his insistence that, in the end, he would own all of the data. Even though it was our participation that had made this data possible. This was unacceptable to me. We would be building our entire website infrastructure on top of that data, and were he to decide, in a fit of astrologically induced pique, that he didn’t like us or our terms, or the way Aries was ascending, he could easily take his crystal ball and go home, leaving us with nothing. We would be royally screwed. And I didn’t need to read any tea leaves to make that prediction. An early-stage startup is a delicate ecosystem, in which competing pressures—investor expectations, market realities, and plain-old plausibility—push in from all sides. I didn’t need yet another outside force dictating our progress.

  I recognized Erlewine’s anxiety. It was the same anxiety that many people felt during the internet boom. His services—both All Music and All Movie—had begun as print publications. They’d been analog. The transition to digital made him uneasy. He was sitting on his hoard, and he didn’t want anyone to take it away.

  Eventually, I started to get exasperated. Under Erlewine’s plan, we’d be paying him for the privilege of learning about DVDs we technically owned. But I also knew that he had us over a barrel. We were the ones under a time crunch, and he had the information we wanted. The site was launching in just a few months, and we needed the data he had to even begin to build the database that would power our site.

  I stalled. I hemmed and I hawed. Every day, as January flowed into February, Christina and Eric badgered me. We needed the data to be able to build a model of the database. We were writing our own blurbs,
adding the movies to “collections,” and making other editorial decisions, but all of our content needed to be attached to a root record. Even if we finished the content well before launch date, we knew there were days—if not weeks—of effort to actually connect our content with Erlewine’s. We couldn’t just plug everything in the day before launch.

  There was another problem. Even if we could get a deal, the amount of data he would be sending our way was so massive—especially in 1998—that it couldn’t be done over the internet. We’d receive the data in a pretty analog fashion: reels of magnetic tape. Forget email: this data would have to be shipped to us in a box. That was another reason we needed all the time we could get: once we had the tape in hand, we’d have to “translate” it, teaching our site how to “read” it.

  The contract Erlewine had drawn up was completely unacceptable. I hated it. But I had to sign it.

  He won.

  But the second I signed that contract, I started thinking of ways to get out of it.

  I have to hand it to the guy: Michael Erlewine found a way to monetize his obsessive-compulsive hoarding. That company ran according to his principles—it resembled him. Sure, walking through their “offices” was like taking a tour of the inside of an obsessive record collector’s skull—but the place had an ethos. It had an identity.

  It was not anywhere in the same ballpark of how I wanted my company to look, feel, or behave—but it worked for him.