The Method That Funded Google Could Not Fund a Solar Panel

IT HISTORY

The Method That Funded Google Could Not Fund a Solar Panel

In 1974, a twenty-three-year-old with a fresh master’s in electrical engineering from Rice wrote to a small firm on Sand Hill Road in Menlo Park and asked for a job.

The firm was Kleiner Perkins. Tom Perkins wrote back and turned him down. The firm kept an informal rule: before you invest in operating companies, you should have worked inside one.

So Doerr went to Harvard Business School. He also went to Intel, where he sold microprocessors under Andy Grove in the years when the 8080 was turning a memory company into the one that would define the personal computer.

He came back six years later. That time they hired him.

Over the next two decades, Doerr put money into Sun Microsystems, Compaq, Intuit, Symantec, Netscape, Amazon, and Google. He told audiences, without visible embarrassment, that what he was watching was the largest legal creation of wealth in the history of the planet.

He was right, which is the uncomfortable part of this story. The other part is what happened after.

Selling Chips Teaches You Something Papers Don’t

The thing to understand about Doerr’s Intel years is that he did not spend them in a lab.

He was in sales. He carried a bag. He learned the shape of a purchase order and the specific ways a deal dies: the procurement committee that stalls, the incumbent vendor who drops price at the last minute, the engineer who loves your part and cannot get it approved.

Most people who fund technology come from one of two places. They understand the technology, or they understand the money. Doerr had spent his twenties in the narrow strip between them, in the place where a working chip becomes a shipped chip becomes a paid invoice.

His questions were never only about whether something could be built. Who signs the check. What gets ripped out of the budget to make room for this. How long the sales cycle runs, and whether the company can survive it.

An engineer falls in love with a prototype. A salesperson knows the distance from prototype to purchase order is measured in quarters, and most companies die inside it.

Doerr brought both instincts back to Sand Hill Road in 1980.

Four Twenty-Somethings and Seven Pages

His early portfolio reads now like a table of contents for the 1980s.

Sun Microsystems. Compaq. Lotus. Cypress Semiconductor. Symantec. Quantum. Workstations, personal computers, packaged software, disk drives — separate industries on paper, one continuous movement underneath. Compute was getting cheap. Software was becoming a business you could sell on its own. Networks were starting to connect the machines to each other.

Sun is the one worth slowing down for.

The four founders averaged twenty-seven. Vinod Khosla, Scott McNealy, Andy Bechtolsheim, Bill Joy. Bechtolsheim’s workstation design had come out of a Stanford research project called the Stanford University Network, which is where the company got its name. Khosla wrote the business plan.

It ran seven pages.

Doerr repeated that detail for the rest of his career, and the reason was not sentimental. A seven-page plan is a filter. You cannot hide inside it. Either the machine makes sense in seven pages, or the thinking is not finished.

What Sun proposed was heretical for its moment: build the workstation out of parts anyone could buy, using standard networking, a commodity microprocessor, and Berkeley’s Unix on top. Against them stood Apollo Computer in Massachusetts, with better hardware, deeper experience, and a more respected management team.

Sun won. Apollo was acquired by HP in 1989 and disappeared.

Then the compounding started, which is the part outsiders always miss. Bechtolsheim would write one of the first checks into Google. Bill Joy, as Sun’s chief scientist, would drive the work that produced Java. Khosla would join Kleiner Perkins as a partner and in the 1990s back a set of telecom equipment companies whose returns ran into the billions.

One investment made in 1982 kept paying out through people for twenty years.

Kleiner had a word for this, borrowed from Japanese industrial groups: the keiretsu. The portfolio worked as a network the firm could point at problems.

What Actually Happens After the Wire Transfer

Money is the visible thing a venture firm provides, and rarely the scarce one.

A founder can write a browser and have no idea how to sell it into a Fortune 500 IT department. A researcher can design a chip and not know who the first ten customers are. Orders start arriving. Hiring, pricing, board governance, and the next round all land in the same month. Code solves none of it.

Netscape is the case where you can watch this happen in daylight.

Jim Clark had already founded Silicon Graphics. In 1994 he flew to the University of Illinois, interviewed the programmers who had built the Mosaic browser, and hired the whole group. Marc Andreessen was twenty-two.

Kleiner came in, and Doerr took a board seat. Then he went to work on the part nobody puts in the founding myth: staffing the company. Marketing, sales, and above all a chief executive. That turned out to be Jim Barksdale, who had run operations at FedEx and McCaw Cellular, and who arrived carrying the specific credibility a twenty-two-year-old cannot manufacture.

Netscape went public on August 9, 1995, sixteen months after it was founded. The stock was priced at $28. It opened so far above that price that trading could not clear for ninety minutes, ran to $74.75 intraday, and closed at $58.25.

Microsoft had taken eleven years to go public. Netscape did it in sixteen months, without ever having turned a profit.

The IPO is the scene everyone remembers. The work that decided it happened in the twelve weeks before, on recruiting calls.

1999, and a Number That Sounded Absurd

By the late nineties, Doerr had a thesis and a reputation, and both were about to be tested by two graduate students who had taken down a serious fraction of Stanford’s bandwidth.

Kleiner Perkins and Sequoia Capital both invested in Google in June 1999, $25 million between them. It was an unusual arrangement: two rival firms sharing a seed-stage deal, each taking a board seat. Doerr for Kleiner, Michael Moritz for Sequoia.

Doerr has told the story of that pitch many times. He asked Larry Page how big he thought this could get.

Page said ten billion dollars.

Doerr did the arithmetic any investor does automatically and assumed Page meant market capitalization. He asked for clarification, and Page said he meant revenue.

“I about fell out of my chair.”

Yahoo was the dominant portal that year and the obvious comparison. Its revenue for all of 1999 came to $588 million. Page was describing a revenue line seventeen times that, for a company that did not yet have a way to charge anyone.

He was low.

What Doerr did next is the part that generalizes. He did not simply wait for the technology to work. He spent eighteen months trying to find Page and Brin a chief executive and, by his own account, put roughly seventy-five candidates in front of them. All rejected. The founders wanted someone who could run a company and also survive a technical argument, and that person is rare.

Eric Schmidt, then running Novell, took the job in 2001. Doerr also brought in Bill Campbell, the former Intuit CEO who spent years coaching executives across the Valley, to sit with Page and Brin a few hours a week.

And in 1999 he brought the founders one more thing he had carried out of Intel: OKRs. Objectives and Key Results, the goal-setting discipline Andy Grove had drilled into him twenty years earlier. Doerr would eventually write a book about it, Measure What Matters, in 2018.

The mechanism is unglamorous. Write down what matters. Attach numbers that can be checked. Make the whole thing visible across the company.

For a young company, the problem is never a shortage of opportunity. It is that everyone is working hard on things that quietly cancel each other out. Writing the objectives where everyone can see them is how you find out where the money and the people are actually going.

Google’s success does not belong to OKRs, and it does not belong to any investor. It belongs to search quality, to distribution, to the advertising system, to infrastructure, and to a founding team that made a long run of correct calls.

What Doerr did was get a room full of extraordinary engineers to confront the boring organizational problems earlier than they wanted to.

The Ones That Went to Zero

Tell only the Sun-Netscape-Google version, and this becomes mythology. Here is the other ledger.

GO Corporation wanted to build pen computing, a tablet you wrote on with a stylus. Jerry Kaplan founded it, and Mitch Kapor of Lotus was involved. Kleiner invested. IBM and AT&T were both interested. The idea was correct and roughly two decades early: handwriting recognition did not work well enough, the hardware was not ready, and the company had to fight Microsoft over operating systems at the same time. It died, and Kaplan wrote a book about the experience called Startup.

Dynabook Technologies was an attempt at a thin, light laptop in the late 1980s, launched by Doerr and Khosla. The chip work came together. The display did not. The company failed.

These are useful precisely because they are boring failures.

An investor can supply capital, recruit executives, and open doors. An investor cannot make a component mature on schedule, and cannot make users want something before they want it. Markets keep their own clock.

Doerr’s framework of team, technology, market, and timing did not save him from either one. It was only ever a way of narrowing the space of things he could not see.

Technology is the part venture capitalists are best at judging. Market timing is the part nobody is good at.

When the Model Stopped Working on Him

Then came the decade that gets left out of the keynote version.

Kleiner Perkins made a large bet on clean technology. In 2012 it tied with Draper Fisher Jurvetson for the most clean-tech rounds of any firm that year, twenty-five of them. Doerr’s own summary was that some of the bets had been too big and too broad.

The arithmetic underneath is worse than the summary. MiaSolé, a thin-film solar manufacturer, raised more than $500 million and was valued at $1.2 billion, then sold to a Chinese buyer for $30 million. Fisker Automotive raised $1.4 billion and went bankrupt.

Across the whole sector, $21.3 billion went into private clean-tech companies after 2000 and came back at a gross internal rate of return of 6.6% through late 2012 — venture money earning less than a decent bond.

Solar and biofuel startups carried the capital requirements of heavy industry and the timelines to match. Software did not prepare anyone for that.

Meanwhile, a different kind of company was being built in the same zip codes. Kleiner Perkins was not early in Facebook, LinkedIn, or Groupon. In 2012, a year when venture-backed companies were selling and going public at high valuations, the firm barely appeared on those lists at all. Sequoia was on nine of them. Accel on six. Kleiner on two, both small.

On the Forbes Midas List, Doerr fell from №12 to №26. He had not been in the top ten since 2009.

In May 2012, Ellen Pao, a junior partner and Doerr’s former chief of staff, sued the firm for gender discrimination and retaliation. The trial in 2015 became the most closely watched employment case in Silicon Valley’s history. A jury found for Kleiner Perkins on all four claims. By then the case had done something no verdict could undo: it put the internal culture of the most famous venture firm in the world on the public record, in sworn testimony, for four weeks.

Asked by Forbes how the year had gone, Doerr reached for the upbeat answer first, “Fantastic, really fantastic,” then paused and recalibrated.

“It’s also been a very challenging year.”

The firm changed its process. It built an internal system called Dragnet to track hiring patterns, social signals, and app store rankings across potential targets, an attempt to replace some portion of instinct with instrumentation. It revisited how deals got approved, after a period in which small partner teams could move fast enough that experienced partners never saw a deal at all. Speed had been the goal. Speed turned out to allocate capital by whoever moved first rather than by anyone’s plan.

The interesting thing is what did not change.

David Swensen, who ran Yale’s endowment, was asked about Kleiner during that stretch. His answer was that the firm had generated more dollar gains than any other partnership in Yale’s private equity portfolio, and had done it on far fewer dollars committed.

“I’m not going to bet against them.”

The Distance Between a Demo and a Company

Doerr’s line about the largest legal creation of wealth was accurate about its era. Personal computers, packaged software, and the internet did generate hundreds of billions of dollars in roughly twenty-five years, and the people who financed them captured a real share of it.

But the sentence usually gets quoted as though it were about money. Read the record, and it is about a much narrower claim: that at certain moments a technology stops being interesting and starts being purchasable, and that the distance between those two states is the entire job.

Sun’s seven pages. Netscape’s twelve weeks of hiring. Seventy-five rejected CEO candidates. Capital solved none of those. They live in the gap between a machine that works and a company that survives.

That gap has not moved.

Today the demo is an AI model, or a battery chemistry, or a protein. The demos are astonishing, and the videos are excellent. What happens after is the same set of unsexy questions: who deploys it, who carries the procurement risk, how it enters systems that were built before it existed, whether the team reaches repeatable revenue before the money runs out.

Doerr got twenty-five years of those questions right, and then hit a decade where the answers changed underneath him, and the framework did not adapt fast enough. Clean tech was not a failure of belief. It was a failure to notice that a method calibrated on software does not transfer to industries that consume capital like a refinery.

He built an operating system for turning inventions into companies. It worked well enough to look like prophecy.

Then the world produced a category it was not built for, and the same man who could see the internet coming in 1994 spent a decade funding an industry that was not ready. Both of those are the same story, which is what makes it worth telling.

Sources

• Google receives $25 million in equity funding — Google Press, June 7, 1999
• John Doerr’s Plan To Reclaim The Venture Capital Throne — Forbes, May 2013
• John Doerr Takes On His Critics And Talks Up Kleiner’s Prospects — Forbes, May 2013
• The Networker — The New Yorker, August 11, 1997
• Kleiner Perkins’ John Doerr: High Tech, High Finance — Institutional Investor
• When John Doerr Brought a Gift to Google’s Founders — Wired
• When John Doerr Met Larry Page — TechCrunch
• Pao v. Kleiner Perkins — Wikipedia
• Yahoo! Inc. FY1999 results, revenues of $588.6 million — SEC EDGAR
• Buyers Drive Up Netscape Stock Price — The Washington Post, August 10, 1995
• Microsoft’s 1986 IPO — Goldman Sachs, company history