1. Birth of a Notion

Surfin’

Experienced entrepreneurs tell you to be prepared for the emotional roller coaster that launching a new venture entails. I think this evokes the wrong image. A roller coaster has a fixed track. You can see when you are climbing to the peak or diving into a valley, and you can see the bottom. You are also strapped in, and insurance companies have signed off on the risk. Your path is precisely the same as that of many who have gone before and will come after. Starting a new venture is nothing like that.

Surfing is a more apt metaphor. Before you even get wet you can watch how the waves break, look for submerged obstacles, and wax your board to minimize slipping. For the most part, you can choose your wave from how its early form looks. You can see which way the wave is breaking, and opt to go left or right on the wave. You can choose whether to kick out – if the wave walls up, cut back and let the wave reform– or ride through. The experience of a good ride on a strong wave is exhilarating, but you can also wipe out. A bad choice, or just a bad break, can send you flying into a storm of white water, crushing you under its weight – leaving you unsure of which way is up and whether you will get there in time for another breath before being sent down again. No two waves or rides are the same.

The uncertainty inherent in surfing parallels the new-venture process. Sometimes great rides are abundant and sometimes the waters are flat. Waves come in sets, as do entrepreneurial opportunities. The key is to recognize the opportunity in the early stages – when the wave is forming – ride the curl while the break is good, and kick out before the shore pound crashes you to the bottom. Easier said than done.

I hung up my homemade surfboard many years ago, when the famous storm surf of Christmas 1962 brought 30-foot waves to Malaga Cove in Palos Verdes. I watched from the bluff as veteran surfers Greg Knoll and others rode these massive forces of nature. I knew I didn’t have the skill or bravura to join them. The next biggest wave I saw was almost four decades later, when the digital revolution and the Internet craze built toward a crest. I jumped on the mythical seventh wave of the seventh set. It was quite a ride. I tried to recapture the thinking that went on during the experience, the feelings both good and bad, and the emotional texture from clarity and joy to confusion and anger. For the feelings and emotions, the writing has to stand on its own. I hope I have been revealing enough to prepare you to encounter some of the vicissitudes I faced. I am eager to share the thinking that took place during the ride this book describes. I hope it aids understanding the process of new-venture initiation, particularly for university-based technology, in which radical innovations can change the way we live in and experience the world. Ultimately, this is a story about the tension between a world of technological genius and a world of business. The masters of these worlds don’t know how to talk to each other. Yet, so much of the magnificent prospects for our future depend on this communication. I think managers need to extend their thinking at least enough into the technology that the basis for decision making is not opaque. And the technological geniuses of this world need to understand that, while others may be the best judges of the practicality of markets and opportunities, they more than anyone else are the best judges of the technological limits of their innovations.

I start in the middle of the ride, with the story of the first live test of the new technology, before rewinding to the beginnings of this adventure.

July 28, 2000

Jason Kapp picked me up at 8 a.m. from my Santa Monica home for a 9 a.m. meeting at Idealab Capital Partners, the Pasadena-based Internet incubator that spawned eToys, CarsDirect, Cooking.com, Overture, and others. Jason, our VP of client services, had played a key role in helping me start this venture. We had pitched our approach to technology-enabled marketing many times before in our successful $5 million B Round. We knew the story: The Internet dangled the prospect of huge potential returns for those who could monetize its promise for personalized shopping and browsing. But the preparation this time was different. We had heard exaggerated rumors about successes in our arena by one of Idealab’s portfolio companies, as well as tales of their acumen at taking other people’s ideas. So the focus centered on what to reveal and what to conceal about our approach. But the dark presence in the car, and over all discussions for almost a month, concerned when we would launch our first real market test – a test that would determine the near-term fate of this startup.

Through the B-Round representative on our board, we had arranged with iPlayer.com to purchase 20 million banner-ad impressions designated for registered users on its popular Internet versions of video games. We would use our segmentation method to learn about the preferences within each segment of its customers; we hoped to use that machine learning to increase the abysmal (and getting worse) click-through rates on the company’s banner ads. Using the site’s own customer data to improve ad targeting fulfilled the marketing maxim to know your customers and circumvented all the privacy issues concerning public-policy makers at the time. June 15 was set as the starting date. At $.75 per thousand, $15,000 to get a live test of the extension of our personalization technology in the Internet ad space seemed like a good idea. When the planned test failed to start on time, Scott Sellers, the VP for business development at iPlayer.com, told us that June inventory was sold out at $5 CPM (cost per thousand) – a much higher rate than we had negotiated. We arranged to start July 1 at $2.50 CPM, and waited…and waited some more.

Two VA-Linux machines with our software had been co-located within the test site’s Web-server racks at Exodus. We had two similar machines in our own half rack at PSINet in Marina del Rey, and four more tied to a T-1 line at our office in Santa Monica. Tests on our end were fine so far, but within iPlayer.com’s network we couldn’t read the cookie – that tiny piece of text code that gives the originating site so much information about the customer at the other end. We needed the site to change six lines of code to share cookies just within its complete subnet. Since my technical expertise covered not cookie logic but the conceptual and analytical models that drove our learning and optimization algorithms, I was barely grasping onto these problem areas. Early in July, the client’s CTO promised to make the simple change we required, but it didn’t happen. On July 19, the CTO left for a year, supposedly to recover from an unspecified illness. I suspected burnout. It was July 25 before Sellers got his team to put the new cookie code in place. We found and fixed a small bug in our database agent, and sent a ready-to-go message to Sellers. Ravi Srinivasan, the head of our technical-implementation team and an Anderson MBA student, heard that AdForce, the firm serving iPlayer.com’s ads at that time, had been notified to start sending our ads first thing on July 27…and still we waited.

Jason and I agreed to show Idealab much the same demo that had been developed in record time for our first board of directors meeting in early February. The demo featured our product-recommendation suite, called PersonalClerk, that incorporated the Internet advertising optimization only as a simple, natural extension of the personalization solution. We wouldn’t talk about the ad test with Idealab.

The 45-minute meeting went nowhere, but went nowhere smoothly. We learned Idealab’s efforts were mostly vapor – great customer data, but no sense that the company knew how to use it effectively. Technology-enabled marketing operates at the boundary between intelligent information systems and that obscure area called marketing science. Without strong capabilities in both areas, the problems are at best half solved and the solutions, consequently, are half-baked.

Once back in the car I called Ravi. He sadly reported, “No data yet.” Then I heard Chuck Yu, our hardware guru, in the background yelling that the first ads were being sent. It was 9:52 a.m. and our labor pains had just begun. The drive back felt like a rush to the hospital to be in time for the delivery of my first child.

When Jason and I arrived back at the bullpen on the second floor, where most of the technology group sat, everyone was gathered around Chuck’s desk. One window on his Linux notebook tracked the number of open sockets – one for each active user. The theory was that when a user’s browser requested a page, our HTTP agent that handled the raw traffic opened a socket as a file description, delivered an ad, sent the needed information to our learning algorithm and a session log, and closed the socket. We watched as the number of open sockets increased toward 1,024, blocking the thread and crashing our system. Our design didn’t threaten to bring down the client site or interfere with its basic operation, but when our system crashed, the ads we paid for weren’t being delivered. The data that were crucial to our machine-learning algorithms weren’t being collected. Chuck’s fingers flew over the keyboard each time the thread was blocked. If the sockets were unblocked (closed) when the process crashed, a simple chain of commands would restart the process immediately. If the sockets remained open – blocked from accepting a new-user request -- Chuck had to reboot our Linux box inside the client’s rack at Exodus before restarting the process. Then learning would begin again, until the next blockage.

Why was this happening? What could we do about it? I didn’t have a clue. Chuck was madly typing away to keep our downtime to a minimum, and I had no insight into the problems we faced. Thirty years of designing and building medium- and large-scale analytical and statistical systems for squeezing meaning out of market data, and I had never been this clueless. Others had done major chunks of many of my earlier projects, sometimes for efficiency and sometimes for learning. When stumbling blocks were encountered, I had always been there to solve the problem. Not this time. After more than three decades on the faculty of UCLA’s management school, I was now outside the ivory tower. I needed to step back, micro-manage less, and grant other people control over a problem-solving process that transcended my expertise.

The intellectual resources available were substantial, but incomplete. Giovanni Giuffrida, a UCLA doctoral candidate in intelligent information systems, was our CTO, and had developed wonderfully into a leader of the technology group. An ever-reshuffling handful of program developers would gather around the conference table in his office next to the bullpen while Giovanni worked with them through whatever was the greatest barrier to our technical progress. Giovanni and I previously had worked together for several years on large-scale forecasting projects[1] and research-oriented datamining projects that had led to the first datamining article published in the mainstream management literature.[2] Our big-data experience mostly concerned retail scanner records. We had to deal with about 25 million records at a time and be prepared to create up to 800 million forecasts a year. Those are very small numbers compared to what we faced on the Web.

Our main Web expert, Fabrizio diMauro, an amazing code hacker, was stuck on a trans-Atlantic flight returning from Italy – and grounded in Newfoundland when his girlfriend became extremely ill in flight. When Giovanni, Jason, David VanArsdale (VP/admin) and I had started this venture together, Fabrizio was the first person hired in the technology group. Giuseppe Blanco, a Web-design specialist, C programmer, and the third part of the Sicilian Connection, was in the bullpen. Not only did these three grow up in the same small town in Sicily, they all reconnected in computer-science graduate programs at UCLA. Brandon Davinski, the UCLA computer-science undergrad who worked with us full time in the summer, brought an almost scary knowledge of Web programming. He already had found a major security hole in another client’s cash-register program. He could change the price for anything in his or anybody’s current shopping basket. Many times I mumbled to myself that I was glad Davinski was on our side. Wesley Rhim was a database expert with deep SQL and Perl skills. Murilo, a great C programmer with a PhD in physics, had just started earlier that week. Nick, a computer-science undergrad from MIT, worked with us that summer. So did Jonathan, the best computer mind in my older son’s cohort (then 20), who had programmed the common object module (COM) needed for dealing with the Microsoft servers that the client used. In supporting roles were a group of very talented college students in their first important summer jobs.

At first, the core technology team tried to force the sockets to close – fix the symptom and ignore the problem. We didn’t find a way to do this. Maybe the level of simultaneous traffic was too heavy for only 1,024 sockets, but this seemed unlikely given the short duration (milliseconds) that each request required. So Giuseppe and Jonathan pulled a 10-thread version of the HTTP agent from the program repository and tested it. Chuck copied it over to the machines at Exodus, compiled it – keeping the other system going to the last second -- and then started the 10-thread code. The inexorable rise in blocked sockets foretold the outcome. Blocked sockets led to crashed threads. It took 10 times as long to crash the whole program, but this try eliminated traffic volume as the possible cause. To eliminate bandwidth issues, the banner ads were copied over to the extra machines at PSINet and served from there. No discernible difference.

By mid-afternoon, someone suggested that possibly old or strange browsers contributed to the problem. Our session logs were full of requests from newer Internet Explorer and Netscape browsers, but none from the much older versions of these or from WebTV browsers. Jonathan grabbed a WebTV emulator that we used to log into iPlayer.com. No ad was delivered and no record of the user showed up in the session log. This signaled that the problem was in how our HTTP agent handled old or odd browsers. The HTTP agent was our interface to Internet traffic – in essence, a Web-server operating system stripped down for speed. We could handle about 1,200 requests per second, per box in a system we knew how to easily parallelize, if additional speed was needed for high-volume commercial sites. But apparently, too much of the browser handling had been stripped out. The fix required putting a whole APACHE Web-server operating system in front of our HTTP agent to take over the browser handling. Giovanni understood the best talents of his core team and assigned David and Jonathan to the APACHE tasks, Giuseppe to the modification of the HTTP agent, and Nicholas to the required CGI scripts. Chuck was still typing rapidly and constantly to keep the processes almost continuously available. Murilo poured over APACHE support documentation, regretting that as a “newbie” he couldn’t be more central to the excitement of the problem-solving process. His time would come.

My role shifted from quietly making sure that the discussions, diagnoses and possible remedies made sense to even more quietly making sure the group had enough pizza and sodas to keep them working. As long as I stayed there, I knew they would stay. That much mutual respect we had built in the eight months of working together on the multitude of technology problems startups must solve to endure. Unfortunately, the office space we leased wasn’t set up for the long hours of startup companies. The air conditioning shut off at 7 p.m. on that hot July Friday. I had the unassigned troops getting all the fans from our 6,000 square feet of offices into the technology bullpen and the corridor outside in order to usher the cooling evening air into an overcrowded work space.

Close to 10:30 p.m., the pieces began to come together. Chuck started sending chunks of code over to Exodus and bolting them in place: copy, compile, start, check the logs, try the WebTV browser, and check the logs again. The system held. By 11:10 p.m. we knew we had succeeded. Chuck had spent more than 13 hours keeping our baby alive as it worked its way through the birth canal. We all felt like proud new parents; the sense of accomplishment was palpable. We made bets on what final click rates our optimization would ultimately achieve. Chuck set up the monitors that would ring the cell phone and send emails if the baby sneezed, and we all prepared to go home – having wasted fewer than 150,000 ads from the 20 million we had purchased.

The bonding that went on that day-turned-night had a lasting effect on the performance of the technology group members. Their individual skills in problem diagnosis and code creation proved instrumental to achieving a desired outcome. They all felt a deep stake in the performance of the software and the company. In important ways, their work became a mission rather than merely a job. The focus, tenacity and dedication they continued to exhibit, I attribute in no small degree to what we all experienced that night.

Something special happens when people are connected to the mission of the organization, value what the company wants to accomplish, and sense an alignment between their skills and efforts and the sought-after organizational goals. I learned this lesson 25 years ago when I directed the UCLA Arts Management Program – an MBA program that trained managers primarily for not-for-profit performing-arts organizations, museums and arts councils. The mission of a not-for-profit arts organization has value – both conceptual value and monetary value. The monetary value is easy to see when people volunteer their discretionary time and money to help the ballet or opera, or when others work for much less compensation because they are “working in the arts.” The conceptual value is less tangible. Caring about the organization is a partial antidote to the bureaucratic barriers and careerist silos that trap so many other organizations. While not a cure for ineptitude, the conceptual value of an organization’s mission helps control behavior within the organization.

I once thought these organizational advantages accrued only to the not-for-profit sector.[3] My experience in founding Strategic Decision Corp. taught me, however, that entrepreneurs have the potential for gaining much of the same leverage. One of my goals for this book is to share my experiences and offer perspectives that can help entrepreneurs, practicing managers, and management students build successful organizations in which people can invest their hearts along with their minds. Another goal is to aid entrepreneurs inside and outside the ivory tower who need to understand the venture-initiation process along with the opportunities and traps that lie in wait. A final goal involves describing a modern approach to strategic marketing planning for new ventures. Along the way I will suggest how the mantra of segmenting, targeting and positioning that has characterized marketing education for the last 25 years can be updated to reflect the realities of technology-enabled marketing in the digital world.

If you are interested in a copy of the entire book email Lee Cooper