Strategic forks: choosing between two paths you can't both take

The license Google turned down

By the end of April 2006, Sun had agreed to accept $28 million from Google over three years, to cover the Java licensing revenue it might lose to an open-source Android. Then the talks broke down, over control of the platform rather than price.

Google had two doors. It could take a controlled Java license and slow Android down, or build on Java's structure without a license, own the platform outright, and be ready to defend that choice in court. It chose the second.

The bill came due. Oracle sued in 2010 seeking as much as $9 billion, and a federal appeals court ruled in Oracle's favor in 2018. On April 5, 2021, fifteen years after the fork, the Supreme Court ruled 6-2 that Google's copying was a permissible fair use. Google owed nothing.

Kodak had stood at the same fork in 1969: license Polaroid's patents, or invent around them. It chose to go around. Polaroid sued in April 1976, within days of Kodak's launch. A permanent injunction forced Kodak out of instant photography on January 9, 1986, leaving 16.5 million owners of its cameras without film, and in 1991 Kodak paid Polaroid $925 million to settle.

Same fork, same choice, opposite endings. The difference wasn't the quality of the reasoning. In both cases someone outside the company got the last word, and only one of them said yes. That is the decision this guide is about: two paths you can't both take, and the people besides you who decide where the one you took ends.

What strategic fork is

Most strategic choices can be adjusted as you go. A fork can't: you license or you fight, you build the chip or you buy it, you split the company or you don't. The forks on record are rarely about what a company wants. They are about how to get it, and the how is written into a contract, a patent claim or a corporate structure that outlives the people who chose it.

Why it is so hard

On paper a fork is a comparison of two options. In a real company, three things make it harder than that.

The ways companies did it

The forks on record take four shapes. Each is the same question, which path, asked about a different part of the business.

What 44 companies did

Here is every decision on record, placed by the year it was made and marked by how it turned out. Hover a mark, or open the table, for the company and the outcome.

Of the 41 decisions that can be judged, 34 were never undone, 5 were reversed and 2 were kept only in part.

Never undone is not the same as right. Nokia's choice of Windows Phone in 2011 was never undone, and its smartphone share fell from 33% to 3%. Kodak's 1989 decision to protect film by not selling digital cameras held for decades, and it filed for bankruptcy in 2012. Mercedes-Benz's 2021 bet on premium electric flagships stood too, and in 2025 it sold 168,800 electric cars, down 9%, while BMW sold 442,072.

The walk-backs

Put the forks that were taken back side by side and the striking difference is speed. New Coke lasted 79 days and WeWork's listing a matter of weeks; Gap's split was called off within eleven months, HBO's name came back after two years, and Dow's move into specialty chemicals took more than a decade to unwind. The two kept only in part were reorganizations, trimmed from inside a few years later.

Coca-Cola and WeWork had chosen paths that only worked if someone else accepted them: drinkers had to accept a new taste, public investors had to accept a $47 billion valuation. Neither did, and both companies backed out within weeks. Warner Bros Discovery took longer to undo a change of name: it had taken HBO off the marquee to make room for everything else, and two years later it put HBO back.

The slow reversals were splits and mergers. Gap found that splitting off Old Navy cost more than keeping it, and paid $189 million for a separation that never happened. Dow's way back from its specialty bet ran through a merger with DuPont and a public breakup, and the specialty businesses ended up with another company. The more a path is written into a corporate structure, the longer and dearer the way back. Procter & Gamble and Unilever needed no one's yes to reorganize, and each kept most of what it had built.

Not every fork an outsider closed was taken back. Kodak kept selling instant cameras until a court ended the line in 1986. Amgen never narrowed its claims; the Supreme Court rejected them. Toshiba's board kept its split on the table until shareholders voted it down. Those choices stood until someone with a vote ended them, and what followed came from the choice.

At a fork, the question is not only which path is better. It is who else gets a vote, and what it would cost to walk back.

What the share price did

For many of the decisions there is a clean market record: the company's total return over the three years after the decision, against the S&P 500.

Read it as a picture of the companies, not a score for the forks. Johnson & Johnson's purchase of Abiomed closed exactly as it was structured, with a tender offer that completed in December 2022, and its shares trailed the index by 65.7 points over the next three years. Microsoft's 2018 reorganization moved Windows a level down the org chart, and its shares beat the index by 109.4 points. Neither number says much about the fork itself.

How to run this decision

If you are standing at a fork, here is what the record suggests doing, in order.

1. Write both paths down to the same level of detail

   A license has a price, a term and a scope. A plan to invent around it is a hope with a budget. If one path is a contract and the other is an intention, you are not comparing them, and the intention will win because nothing about it can be objected to yet.

   Seen in: Kodak, 1969
2. List everyone who can overrule the path you take

   A court ended Kodak's instant cameras and Amgen's broad claims, shareholders voted down Toshiba's split, public investors turned back WeWork's listing and drinkers turned back New Coke. Name everyone who can do that to your path, and write down what each is likely to say. If the path only works when they agree, that agreement is part of the decision, not a risk to note at the end.

   Seen in: Kodak, 1969 · Coca-Cola, 1985 · Amgen, 2014 · WeWork, 2019 · Toshiba, 2022
3. Claim only what you can show

   Amgen disclosed 26 antibody sequences and claimed every antibody that worked the same way. Courts at every level declined to let it. A path defined by what it would cover, rather than by what exists, cannot be defended, costed or staffed.

   Seen in: Amgen, 2014
4. If you can't agree on value, split the price, and read the clause as if it will be tested

   Paying part of a price later can settle an argument about value. But Bristol Myers Squibb's all-or-nothing deadline paid zero after a five-week delay, and Sanofi paid more than $315 million to settle a fight over its contingent payment. Write the clause so nobody can argue about what it means.

   Seen in: Bristol Myers Squibb, 2019 · Sanofi, 2011 · Estée Lauder, 2021
5. Price the path you are refusing

   Google turned down a $28 million license and carried an exposure that reached $9 billion for a decade, knowingly. That is a legitimate choice when it is made with both numbers on the table. It is a gamble when only one of them is.

   Seen in: Alphabet (Google), 2006
6. If the answer is a new org chart, ask what was actually decided

   Procter & Gamble's six business units became five within about two years; Unilever's five groups lost one when Ice Cream was separated. A reorganization is cheap to adjust, which makes it the easiest answer to reach for when the real fork is harder.

   Seen in: Procter & Gamble, 2019 · Unilever, 2022

Warning signs

What it has looked like when this decision was going wrong.

• One path has a price and terms; the other is a plan to work around them.Kodak, 1969
• The path only works if a court or shareholders agree, and nobody has asked what they will say.Amgen, 2014 · Toshiba, 2022
• The claim covers far more than you have built.Amgen, 2014
• A payment is all-or-nothing on a deadline your own work can miss.Bristol Myers Squibb, 2019
• A private valuation is about to meet public investors for the first time.WeWork, 2019
• You are changing something customers think of as theirs: a formula, or a name.Coca-Cola, 1985 · Warner Bros Discovery, 2023
• A split is announced before anyone has costed untangling what the businesses share.Gap, 2019

All 44 cases

Each case is one company at one decision: what was on the table, what it chose, why, and what followed, with every fact sourced.

Questions

What is a strategic fork?

A decision between two paths that exclude each other, where taking one closes the other for years: license a rival's patents or invent around them, build a component or buy it, keep a company whole or split it.

What are examples of strategic forks?

IBM's 1962 choice of one compatible computer family, Google's decision to build Android on Java without a license, Kodak's choice to invent around Polaroid's patents, Apple's move from Intel chips to its own, and GE's split into three companies.

Why do strategic forks get reversed?

On record, the fastest reversals came after the people the path depended on refused it: Coca-Cola's drinkers rejected New Coke and public investors rejected WeWork's valuation. Others came slowly, when undoing a split or a merger turned out to be costly, as with Gap's canceled Old Navy spin-off and Dow's specialty-chemicals bet. Forks that a court or shareholders closed, like Kodak's instant cameras or Toshiba's split, were not reversed; they were ended.

How do you choose between two strategic options?

Write both options to the same level of detail, list everyone who can overrule the one you pick and what they are likely to say, price the path you are refusing, and write any contract terms so they cannot be argued about later.