Читать книгу Winning with Data - Bien Frank - Страница 3

Silicon Valley owes its existence to a Frenchman living in Boston. Born in France in 1899, Georges Doriot graduated from the University of Paris in 1920 and matriculated at the Harvard Business School in 1921. Four years after graduation, he became the assistant dean and associate professor of industrial management at Harvard.1 Five years later, he would be promoted to full professor, in large part due to his beloved manufacturing course that graduated more than 7,000 students during his tenure through 1966. The year-long course tested the general management skills of second-year MBA students, and the final reports of students often exceeded 600 pages.2 In Creative Capital, Doriot biographer Spencer E. Ante summarized his interviews of former Doriot students:

“His lectures were so memorable and controversial – he once lectured students on how to pick a wife – that many former students who have forgotten most of what they learned at business school still remember Doriot vividly.” 3

A sinewy 5 feet 10 inches tall, with incisive blue eyes, a thin mustache, and a penchant for fine tobacco to stuff his iconic pipe, Doriot was highly decorated by the U.S. military. In 1940, he became a U.S. citizen to assume a military post created for him by a former student, Major General Edmund Gregory. Appointed lieutenant colonel and chief of the Military Planning Division, Doriot managed all the procurement for the U.S. Army, from trucks to uniforms to rations.

In the jungles of Southeast Asia, indigenous forces easily tracked American infantryman by their footprints. Unlike the barefooted natives, Americans left boot outlines as they marched through mud. So, Doriot contracted an anthropologist to develop molds of the feet of the locals and manufactured boots with these imprints on the soles. “If you ran down a muddy road you'd swear that was not an American, it was a native,” remembered Lieutenant Colonel William H. McLean.4

In addition to these tactical advances, Doriot and his team resolved large-scale logistical problems that supplied the Allied Forces with the ammunition, nourishment, and equipment to fuel their success. Doriot was ultimately promoted to brigadier general, received the Distinguished Service Medal (the highest U.S. military metal given to a noncombatant), rose to the rank of commander of the British Empire, and was awarded the French Legion of Honor.

After the war concluded, Doriot continued to change the world. In 1959, he and three of his students from Harvard Business School founded INSEAD (Institut European d'Administration des Affairs), the preeminent business school outside the United States.

In addition, he is widely regarded as the father of venture capital. His firm, American Research and Development (ARD), led the first institutional venture capital investment of $70,000 in Digital Equipment Corporation (DEC), maker of minicomputers, in 1957. Eleven years later, DEC went public and netted more than $355 million to ARD, for a 5,000-times return and an internal rate of return (IRR) of more than 100 percent annually. Among other notable investments, Georges Doriot financed the first company of future 41st U.S. president George H. W. Bush.5

American Research and Development's success launched the venture capital industry. A cottage industry through the late 1990s, venture capital exploded in size and impact during the dot-com era.

In the 1980s, venture capital firms in total raised roughly $10 billion per year. During the height of the dot-com era, that figure catapulted to more than $100 billion adjusted for inflation. Since then, in the course of a typical year, venture capitalists raise more than $25 billion to invest into technology, biotechnology, and other kinds of startups.

And the innovation fueled by this capital has transformed the world. FedEx, Google, Intel, Apple, Tesla, Genentech, Bed Bath and Beyond, Whole Foods, Starbucks, Uber, AirBnB: Is there an industry venture-backed startups have not yet disrupted? According to a recent study completed by Stanford researchers Ilya Strebulaev and Will Gornall, 43 percent of U.S. publicly traded companies founded after 1974 have been venture backed, accounting for 63 percent of the total U.S. stock exchange market capitalization. Further, 38 percent of American workers are employed by venture-backed businesses, including 82 percent of research and development employees.6

But, to hear my senior partners tell the story of the heyday of venture capital in the 1990s is to envision a completely different industry than the one we operate in today. One old-time venture capitalist recounted the ways of the bygone days: The 10 or so key members of various firms would eat lunch together on a weekly basis. Like trading baseball cards, they would swap information on the companies they'd seen and decide to invest with each other or not. The capital requirements of these startups outstripped these early funds, so they partnered to ensure the business would have enough runway to achieve success.

Of course, these syndicates competed. But even then, it was friendly. Whoever won the right to lead the series A, the first institutional round, would invite the firm that lost the opportunity to invest in the next one. However, this quid pro quo environment evaporated when the sums of money flooding the industry treated stiffer and stiffer competition from new and existing venture capital firms.

The secular increase in competition has continued over the last 20 years as the scale of technology companies has skyrocketed. Google is now worth nearly $500 billion. Facebook is worth $250 billion. And we venture capitalists chase the next one. The competition drives firms and partners within those firms to develop competitive advantages, and in our business that means information asymmetries, and that means data and relationships. The firm that finds the next breakout company first will often win the right to invest in that business.

There are many different means for venture capital firms to establish that information asymmetry. Some of them develop unique relationships with key angel investors, individuals who invest in very early-stage companies, with just two founders and a dream. Other firms rely on strong relationships with universities and professors who refer standout students to investors. Yet others specialize, focusing on financial services technologies or consumer subscription businesses. At Redpoint, we have tried to develop an information asymmetry using data. That initiative started almost a decade ago.

I started at Redpoint, a venture capital firm headquartered on storied Sand Hill Road in Menlo Park, in 2008. During my first week, I remember receiving a thick envelope in the mail from the National Venture Capital Association (NVCA). The envelope contained the NVCA's directory, a thick tome listing all the different venture capitalists across the country. They numbered more than 5,000. Looking out of my office over the Santa Cruz Mountains, I despaired; how would I ever differentiate myself in such a competitive industry? “What would Doriot do?,” I wondered.

I was very fortunate to work closely with three of the six Redpoint founders, Geoff Yang, Tim Haley, and Jeff Brody, three preeminent venture capitalists who financed billion-dollar businesses like Netflix, Juniper Networks, and HomeAway from their earliest days, and advised those businesses as they transformed huge industries. Over the next few years, they mentored me extensively, and boy did I need it.

As I started to attend board meetings with these senior partners, I began to realize how little I actually knew about startup management. Sure, I could help them with their Google advertising strategies. But founders would ask questions like “How much should I pay a VP of sales?” or “What is a reasonable cost per click on Google?” or “How fast will the business have to grow to be able to raise the next round of capital?” I was at a complete loss to answer these questions. I hoped no one in the room noted my silence.

But I knew, from my days at Google, this data must exist somewhere. So, each time a founder asked me a question about his business, be it revenue per employee benchmarks or marketing efficiencies compared to publicly traded companies, I searched for data.

Once, I found a data set containing startup IPO data dating back to the very earliest days of venture capital that Jay Ritter, a professor at the University of Florida, collected. Startups were surprisingly willing to share their internal data in surveys – anonymously, of course. So, I surveyed them. Friends working at investment banks showed me how to access the data reported by publicly traded companies.

Armed with those data sets and others, I began to answer the questions posed by founders, using the basic statistics ideas I studied in college. The data proved useful to a few of the CEOs I knew, and they asked me if they could share the data. Of course, I agreed. And one of them in particular suggested publishing the results on a blog.

I bought the tomtunguz.com domain, selected a simple blogging layout, and began to write. I jumped when 15 people read my first post. Fifteen daily readers grew to 100. One sunny summer day, I watched as my Google Analytics account reported 1,000 people had visited tomtunguz.com. In disbelief, I called my wife. All those hours spent on nights and weekends writing were finally showing some promise. That night we celebrated with some champagne.

Over the spumante, my wife asked which topics garnered the most interest. I didn't know the answer. So, I began to study the factors that attracted readers: title length, the number of subheadings, the presence of images, voice and tone, time of day to publish, and many others. I learned quite a bit.

I have 48 seconds with a reader. No pretty images, no witty title, no amount of social media validation from influencers will entice the reader to linger. Tweets sent at 8:54 to 8:59 a.m. Pacific Time generate 25 percent more views than those sent a few minutes after 9 a.m. But e-mail subscribers prefer to read content around 10 a.m., a nice midmorning break. Would e-mail readers like to read posts after lunch?, I wondered. A two-week experiment showed they most certainly did not! Open rates fell in half.

As I had done before, I published most of my findings and readers contributed experimental ideas. Over time, this iterative effort grew readership to more than 100,000 readers per month and more than 200,000 social media followers.

But what did all this content marketing ultimately create for Redpoint? A bit of a brand boost, perhaps. Could I justify investing five hours each week to this effort, especially in an industry where the most sought-after startups can raise capital in just a day or two?

At about the same time, I read Aaron Ross's book Predictable Revenue, which describes Salesforce's processes and tools for growing from zero to more than $6 billion in revenue. The former director of corporate sales, Aaron described Salesforce's process of finding potential customers, educating them through sales efforts, and cajoling them through the sales funnel into a satisfied, paying customer. The heart of this software process was, naturally, Salesforce's software, which catalogued the journey of all the potential buyers.

Predictable Revenue inspired me to create a sales funnel from my blog. Read by many startup founders, the blog generated leads – startups in which Redpoint might want to invest. If I could consistently and quickly identify those readers, I might be able to grow Redpoint's network of great entrepreneurs and pinpoint the next great business idea. I decided to call it Scour.

Here's how the system works. I write a blog post. That post is distributed on the web page and through e-mail, social media channels, and some other websites. This content marketing engages a broad network of people. Some of those readers elect to fortify their relationship with the content by electing to receive blog posts by e-mail.

Scour captures their e-mail address in a database. Using that e-mail address, Scour determines who the reader is by looking across the Internet: Where do they work, do they belong to a startup that could be a good fit for Redpoint, whom do we know in common, are they influential in a particular sphere like open-source software or consumer product design? This research process concludes by prioritizing a list of people to meet for us to build our network and find new startups.

Unlike the late 1990s, when the startup ecosystem encompassed perhaps 1,000 founders, today more than 4,000 technology businesses are financed each year. And, again in contrast to the previous era, today those 4,000 businesses leave digital footprints all over the Internet.

Two young computer science students might launch an experimental mobile application for iPhones. The app's success is recorded by Apple. The data is freely available for anyone to download and analyze.

As founders recruit a team, they open requisitions on job boards all over the Internet. One of the founders might decide to blog in order to build an audience of like-minded people who might eventually work for the business and also generate early demand for the product they are building. Twitter accounts, LinkedIn profiles, Facebook interactions, comments in public forums, job listings – with enough data, we have found it possible to identify very early stage startups with promise consistently.

Consequently, we have built data infrastructure to aggregate all these signals scattered across the Internet. We store them in a cloud database and continue to grow the size of that database in the hope that all this data will eventually help us find the next great business before anyone else. With this repository of information, we can experiment and explore investment hypotheses.

Some firms like First Round Capital publish their results on these kinds of trends.7 For example, in their 10-year analysis of their investments, they found female founders outperformed their male peers by 63 percent in terms of returns generated. And founding teams with an average age less than 25 at the time of investment generate 30 percent more returns to the firm than other demographics. But the average age of all founders within the portfolio is 35. Understanding these data points is key to debunking some of the biases that lurk within the Monday partner meetings.

With this kind of data, investors can consistently make better decisions and generate more compelling returns. Again, an information asymmetry manifested in better decision making.

From its modest beginnings with American Research and Development, the venture capital industry has grown in size and sophistication. From marketing to deal sourcing and selection, data has infused every key process of a venture capital firm. And it was that data that led the Redpoint team to Looker.

In 2012, I met Frank Bien and Lloyd Tabb, the CEO and CTO of a Santa Cruz startup, Looker. Jamie Davidson, a friend and colleague from Google, and now a partner at Redpoint, had been using Looker technology at his startup HotelTonight. Another Redpoint portfolio company, Thredup, had been using Looker to manage the operations of more than 100 employees. And they raved about it.

When Lloyd demoed Looker's technology, I fell out of my chair. I knew he had built something unique, a product that would solve the data access problem that plagued nearly every business.

The race to win the opportunity to invest in Looker was on. Over the next week, we gathered as much information on the company as possible. We called existing customers, prospective customers, former coworkers, and industry experts. They all concurred: “Looker is special.”

July 8, 2013, was a Monday, a partner meeting Monday. I remember sending Frank and Lloyd access to our database a few hours before the 1:30 p.m. pitch. The database contained all the information we had aggregated on mobile startups. Lloyd told me later he modeled the data in the car, typing in the copilot seat, while Frank negotiated the conifer-curbed curves of Highway 17 from Santa Cruz to Menlo Park.

During the pitch, Lloyd showed us our data in a completely new way – the way a modern startup explores data, the way businesses create lasting information asymmetries data, the way companies win with data.

That was the beginning of our partnership.