The Revenge Against Oracle: The Origin Story of Workday

Generated by Master Biographer | Source for LinkedIn Content

I. THE HOOK

January 7, 2005. Pleasanton, California.

David Duffield, 64 years old, stands in a conference room inside the company he spent eighteen years building. In a few hours, it will no longer be his. It has not, technically, been his for several days — the deal closed on January 7th — but there are still employees who need to be told. People who trusted him. People who built their careers and identities around the idea that PeopleSoft was different from other companies, that the guy in the Hawaiian shirt genuinely gave a damn.

Duffield tells them.

The room goes quiet. Then it doesn't.

"We were in a conference room to explain the decision to the employees," he would later recall, "and there were people crying — men crying — in that room."

He calls it the worst moment in his life.

This is a man worth hundreds of millions of dollars. A man who grew up in Ho-Ho-Kus, New Jersey, earned an engineering degree and an MBA from Cornell, worked at IBM, built two companies before this one, and turned PeopleSoft into the second-largest application software company in the world. He has succeeded in ways most people don't dream about.

And watching those people cry is the worst thing he has ever experienced.

Within sixty days, he will have written a $15 million personal check and started building again.

He is sixty-four years old.

II. THE BACKSTORY: The Empire That Ellison Wanted

Walnut Creek, 1987

Dave Duffield founded PeopleSoft on May 1, 1987. He was 46 — already older than most startup founders — and he had a very specific, almost irritatingly simple idea: enterprise HR software shouldn't be miserable to use.

The systems that existed were mainframe relics. IBM dinosaurs. Architecture built for the back office, by people who had never met the people who worked in the back office. Duffield believed the emerging world of client-server computing and SQL databases could produce something different — software that looked like it was made for humans, software with a graphical interface and real-time reporting and actual, functioning help documentation.

PeopleSoft's inaugural product launched in 1988. Within years, it was winning.

The company chose "Excellence in Customer Service" as its north star, and Duffield wore that mission like he wore his Hawaiian shirts — constantly, visibly, without apology. He threw beer bashes for employees. He jumped out of a cake at a trade show. He let people bring their dogs to work. He didn't just talk about company culture; he lived inside it, loudly and personally.

By 1999, PeopleSoft had more than 4,000 client organizations. It had expanded from HR into financial management, supply chain, CRM. Duffield himself stepped back from day-to-day CEO duties to bring in fresh leadership — he remained chairman, the spiritual center — as the company grew into something he could no longer manage alone.

In 1993, a Stanford MBA named Aneel Bhusri showed up wanting a job.

The Beers That Started Everything

Bhusri had graduated from Brown University in electrical engineering, worked a stint at Morgan Stanley in corporate finance, and then done his MBA at Stanford. He wanted to work at a technology company. He chose PeopleSoft largely because of the person who founded it.

Duffield took the new hire out for beers.

This was not unusual for Duffield. He did this with a lot of people. But Bhusri — 27 years old, Indian-American, from Pittsford, New York, with the focused intensity of someone who understood both enterprise software and venture capital — filed it away as something meaningful. A CEO who had the time to take a junior person out and actually get to know them.

Bhusri started as director of planning. Within six years, he was vice chairman and head of product strategy. The two men built a working relationship that functioned less like an org chart and more like a long marriage: deep trust, complementary skills, shared language.

Then Larry Ellison decided he wanted the company.

The Bullet That Would Be for the Dog

June 6, 2003. Oracle announces a hostile takeover bid for PeopleSoft at $5.1 billion per share.

The timing was surgical and designed to humiliate. PeopleSoft had just completed a $1.7 billion acquisition of JD Edwards — the ink was barely dry. Oracle swooped in while the company was still digesting a deal, still reorganizing, still vulnerable.

Craig Conway was PeopleSoft's CEO at that point. He was not Dave Duffield — he was more polished, more combative in a conventional sense — but he knew what Oracle's bid really meant. Ellison had publicly threatened to fire thousands of PeopleSoft employees. He had described PeopleSoft's customers as having already bought into a platform with "no future." Conway's first public comment about the offer was that it exemplified "atrociously bad behavior from a company with a history of atrociously bad behavior."

Ellison responded in the way only Ellison could. When asked at an analyst meeting whether the hostility was getting personal, he said: "Trust me — if I have one bullet, it wouldn't be for the dog."

The gloves had been off before. Now they were incinerated.

Eighteen Months of War

What followed was the most vicious hostile takeover battle in enterprise software history. A prolonged, grinding, legally complex siege that consumed both companies.

PeopleSoft deployed every defense available:

A poison pill, promising to return two to five times license fees to customers if the company was acquired — effectively making Oracle responsible for hundreds of millions in customer reimbursements.

A legal offensive, filing suits in multiple jurisdictions.

A public relations campaign positioning Oracle as a predator intent on destroying a beloved company and dumping its customer base.

Oracle responded by hiking its offer. Then threatening to lower it. Then hiking again. Then suing to remove the poison pill. Then watching the U.S. Department of Justice file a counter-suit calling the deal "anti-competitive, pure and simple."

The DOJ challenge bought time. The European Commission voiced objections in March 2004. For a moment, it looked like PeopleSoft might survive intact.

Then, on September 10, 2004, Federal Judge Vaughn Walker ruled that the merger would not be anti-competitive. It was described in the press as a "stunning victory" for Oracle.

Three weeks later, PeopleSoft's own board fired Craig Conway.

The official reason was loss of confidence in his leadership. The deeper truth was that the board had concluded the fight was over and Conway — the defiant general who had turned defense into a personal war — had become an obstacle to the only real exit left: negotiating the best possible price for shareholders.

December 13, 2004. PeopleSoft accepts Oracle's offer of $26.50 per share. Approximately $10.3 billion in total.

Eighteen months of fighting. Eighteen months of defiance, lawsuits, public campaigns, poison pills, and personal insults traded in the business press.

Oracle won.

The Conference Room

The deal closed January 7, 2005. Within days, Oracle announced it would cut 5,000 of PeopleSoft's 11,000 employees — nearly half the company, gone.

Duffield gathered what remained of his people and told them what was happening.

Men were crying.

He called it the worst moment of his life.

After the announcement, Larry Ellison called Duffield personally. "He was very gracious," Duffield remembered. "He said, 'You have built a great company, and we will take care of it.'"

What "take care of it" meant in practice: half the workforce out the door within weeks.

Duffield reached into his own pocket. He set aside millions of dollars to give every laid-off PeopleSoft employee a check for $10,000 — for medical bills, for transition expenses, for breathing room while they figured out what came next. He didn't announce it as a PR move. He did it because these were his people.

III. THE GRIND: Building at 64

Three Months on a Rocking Chair

Duffield tried retirement.

He has described this period with characteristic self-deprecation: "I tried to make model airplanes and failed. Rocked away on the back porch, sort of failed at that too."

He owned a $35 million Dassault Falcon 900EX private jet. He had a compound on Lake Tahoe — 5,000-square-foot lodge, movie theater, pool house with a slide, quarters for his four parrots. He had a foundation, Maddie's Fund, into which he had poured $200 million to support animal rescue efforts nationwide. He had ten children, seven of them adopted.

He had, by any rational accounting, enough.

But Duffield is not wired for enough. He is wired, as one cloud computing consultant put it, the way a fish is wired to swim. "A fish has to swim, a developer has to develop," the consultant said, describing Duffield as someone driven by a compulsion to create — "to make the world a better place, to take the drudgery out of work."

After three months of rocking and model airplanes and failing at both, he picked up the phone and called Aneel Bhusri.

The Breakfast

They met near Duffield's Lake Tahoe compound, at a diner in Truckee.

Bhusri, 38 years old, had been a partner at Greylock Partners since 1999 — one of Silicon Valley's most storied venture capital firms, the shop that would back Facebook, Dropbox, Airbnb. He had watched the PeopleSoft saga from both sides: as a former executive who had given six years of his life to the company, and as an investor who understood the mechanics of capital.

Over breakfast, they built the thesis.

The cloud was no longer theoretical. Salesforce.com had proven that enterprise software could live on a server you didn't own. The Salesforce model worked for CRM — for managing customer relationships. What if you applied that same architecture to the most critical software in every large organization: HR and financial management?

But there was a deeper, harder insight underneath the business case.

Oracle and SAP — the dominant players in enterprise software — could not make this transition even if they wanted to. They had built their entire business models on perpetual licenses and maintenance fees. Moving to cloud would cannibalize their installed bases. They would fight it, delay it, hedge it with "hybrid" offerings that were really just on-premise software with better marketing. They were structurally paralyzed.

Workday would have no such paralysis. They would build everything cloud-native from day one. No on-premise version. No legacy code. No upgrade cycles. Just a single, continuously updated platform that every customer ran on simultaneously.

"We had all the ingredients for a potential success," Duffield said. "The market was springing up. We had the technology. We had the cream-of-the-crop talent. We had a good reputation right out of the gate. We felt we would be foolish not to try."

By March 2005 — sixty days after Oracle closed on PeopleSoft — Workday was incorporated.

The Recruitment

The founding team was not assembled from job boards. Duffield and Bhusri called the people they knew: former PeopleSoft engineers, product managers, designers, salespeople who had been laid off by Oracle or who had quit in disgust when the acquisition closed.

There was a loyalty to this network that defied easy explanation. It was not simply that these were talented people who happened to know each other. It was that they had built something together, watched it be taken, and now had the opportunity to build it again — under a structure that could never be taken.

Duffield and Bhusri personally interviewed the first 500 employees. Not to interrogate credentials. To find people who shared the belief that enterprise software could be better. That work should not feel like punishment. That the people who maintained a company's payroll data and financial records deserved software as thoughtful as their consumer apps.

Greylock Partners led the first institutional funding round, alongside Duffield's personal investment. The founding capital was a mix of Greylock's institutional credibility and the personal conviction of a 64-year-old billionaire who had decided that rocking chairs were not for him.

The Product

Workday launched publicly in November 2006. The HR suite was first. Financial management would follow.

The decision to go cloud-only was not universally applauded. Enterprise CIOs were deeply skeptical. HR data — salaries, social security numbers, performance reviews, terminations — was among the most sensitive information a company maintained. The idea of putting it in someone else's data center, on servers accessible over the public internet, struck many of them as reckless.

Workday's counter-argument was architectural and practical: a multi-tenant cloud model allowed for security investments at a scale no individual company's IT department could match. Updates rolled out continuously, automatically. Every customer was on the same version, meaning bugs were fixed for everyone simultaneously and new features arrived without implementation projects. The eighteen-month "upgrade cycle" that had consumed IT budgets across the enterprise world simply ceased to exist.

It was not an immediate conversion. The early years were slow, grinding, skeptic-by-skeptic work.

IV. THE BREAKTHROUGH: The Enterprise Goes to the Cloud

310 Customers and a $9 Billion IPO

By the spring of 2012, Workday had 310 corporate customers — a mix of mid-sized companies and Fortune 500 names that had placed an extraordinary amount of trust in a seven-year-old startup founded by two people who had lost their last company.

The company had raised additional rounds. It had attracted investors including Michael Dell and Jeff Bezos — the kind of validation that carried meaning with skeptical enterprise buyers. The vision Duffield and Bhusri had laid out over breakfast at the Truckee diner was, by the numbers, working.

October 12, 2012. Workday goes public on the New York Stock Exchange.

The shares, initially priced at $28, surged 74% on opening day. The IPO raised $591.5 million — the largest cloud computing IPO in U.S. history at that point, second only to Facebook's deal earlier that year. The company was valued at over $9.5 billion.

Dave Duffield, 72 years old, had built his second multi-billion-dollar company.

His 85 million shares were worth approximately $3.9 billion.

At the moment the bell rang on the New York Stock Exchange, he was no longer just the man who had lost PeopleSoft. He was the man who had built something larger.

Larry Ellison Calls Workday "A Frail Toy"

Oracle watched this with the institutional anxiety of a company being outflanked.

At the AllThingsD conference, Ellison dismissed Workday as a "frail toy" — the same contemptuous language he had deployed against every competitor who eventually beat him to a market shift. Simultaneously, Oracle began building Oracle Cloud, a belated attempt to build the thing Duffield and Bhusri had started in 2005.

The irony was architectural. Oracle's cloud product had to accommodate its enormous installed base of on-premise customers. It was, by necessity, a bridge product — not a clean break. Workday had no such obligation.

The "frail toy" currently generates over $7 billion in annual revenue.

V. THE AFTERMATH: What Revenge Actually Looks Like

The Takeover Defense

Duffield and Bhusri had learned one lesson from the PeopleSoft disaster above all others: public company shareholders will trade long-term vision for short-term premium. Oracle didn't just beat PeopleSoft through superior strategy. Oracle beat PeopleSoft by offering shareholders enough money that the board had to say yes.

At Workday, they designed the capital structure to make this impossible.

Class A shares — the ones traded on public markets — carry one vote each. Class B shares, held by the founders and their affiliates, carry ten votes each. The result: Duffield and Bhusri control approximately 68% of Workday's voting power, regardless of what any outside shareholder accumulates.

No hostile bidder can force a sale. No board can capitulate to a premium offer. The founders decide.

Bhusri was direct about the lesson: "Dave and I control the votes, we have a dual-class [stock] structure. We learned our lesson."

Both men took annual salaries of $35,000 — a number so small it became a statement. They were not here for the paycheck. They were here because, as Duffield once said, a fish has to swim.

The Character of the Man

Craig Conway — the man who had gone to war against Oracle, who had called Ellison a sociopath, who had been fired by his own board in the final stages of the fight — said something instructive when asked whether Workday was Duffield's revenge.

"It's just not in his character."

Conway said he "never heard Dave invoke Larry's name" in the context of Workday's founding. The motivation, as far as Conway could discern, was simpler and less cinematic: Duffield missed working. He missed building. He missed the particular satisfaction of making something that helped people do their jobs better.

Duffield himself offered a version of this, in his own words: "I wanted my kids to see me leave in the morning and go work."

Not revenge. Redemption. And perhaps a more personal one than the business narrative usually captures — a father who wanted his children to see him choosing purpose over comfort.

The Numbers That Don't Lie

PeopleSoft, at its peak, was the second-largest application software company in the world. Oracle paid $10.3 billion for it in early 2005. That was the number that was supposed to define the end of Duffield's story.

Workday's current annual revenue: over $7 billion.

Workday's current market capitalization has reached, at various peaks, over $60 billion — multiples larger than what Oracle paid for PeopleSoft.

Dave Duffield, who started building again at 64, who wrote his own check, who interviewed the first 500 employees personally, who structured the company so it could never be taken — has built something Oracle cannot acquire, cannot replicate on the same architecture, and cannot overtake without dismantling itself to do it.

At 85, he remains a founder emeritus on Workday's board.

Maddie's Fund has distributed over $300 million to animal welfare organizations across the United States — more from a single couple than any other donors in the history of animal welfare philanthropy.

The foundation was named after a Miniature Schnauzer who was, in Duffield's words, "the lighthouse during the stormy period."

In the end, that metaphor holds for Workday too.

The storm was Oracle. The lighthouse was the conviction that if you build something with genuine care — for the people who work there, for the customers who depend on it, for the culture that sustains it — it will outlast anyone who tries to take it from you.

The worst moment of his life turned out to be the preface.

KEY FACTS (for LinkedIn content)

LINKEDIN CONTENT HOOKS (ready to develop)

1. "The man who started his second billion-dollar company at 64 spent three months first trying to make model airplanes."

2. "Oracle fired half of PeopleSoft the week after the acquisition. The founder wrote personal $10,000 checks to every laid-off employee. Then started building the replacement."

3. "Larry Ellison called it a 'frail toy.' The frail toy now does $7 billion a year."

4. "They structured the entire company to make what happened to PeopleSoft impossible. Two founders, $35K salaries, 68% of the votes. The lesson from the worst hostile takeover in software history, written directly into the corporate bylaws."

5. "He didn't say revenge. He said: 'I wanted my kids to see me leave in the morning and go work.' Same result."

6. "Men were crying in the conference room. Within 60 days, he'd written the check and started over. At 64."