In September 1997, two Stanford graduate students needed a better name for the search engine they had been running on borrowed university servers. They chose a mathematical term for an unimaginably large number, then checked whether the matching domain was available.

What appeared on the screen was not the word they had intended to search for. The altered spelling was available, Larry Page liked it, and one of the most recognizable names on the internet was registered within hours.

The mistake seemed minor at the time. Years later, it would prove unexpectedly useful.

The search engine started with a really terrible name

BackRub, and yes, there was a hand involved

Credit: Ehud Kenan / Flickr

Before Google was Google, it was called BackRub. Larry Page and Sergey Brin began developing the search engine in January 1996 while studying for their PhDs at Stanford. Their project was built around an unusual idea for the time: a webpage’s importance could be estimated from the number and quality of other pages linking to it, rather than relying mainly on how often a keyword appeared on the page.

They called the ranking system PageRank, while the search engine itself became BackRub because it analyzed these incoming “backlinks.” That link-based approach was central to how Google eventually came out on top in web search. The name was technically accurate, but it sounded less like a revolutionary search technology and more like a questionable service advertised on a bus-stop flyer.

The early branding did little to improve matters. BackRub’s homepage, hosted on Stanford’s servers, featured a crude black-and-white logo created from a scan of Page’s own left hand. As the project expanded, its crawler also placed growing demands on Stanford’s network, making it increasingly obvious that BackRub had outgrown its identity as a graduate-school experiment.

The project needed a name that could express the scale of what Page and Brin hoped to build. Their search eventually led them to mathematics.

The word they wanted was invented by a nine-year-old

Googol came from a walk in New Jersey, not a lab

In 1940, mathematician Edward Kasner, with James Newman, published Mathematics and the Imagination, introducing readers to the name given to 10¹⁰⁰, or a 1 followed by 100 zeros. Kasner had not invented the word himself. According to the story he later published, he was walking through the New Jersey Palisades when he asked his nine-year-old nephew, Milton Sirotta, to suggest a name for an unimaginably large number. Milton answered with “googol,” casually creating a new piece of mathematical vocabulary.

Milton also proposed “googolplex,” which Kasner later defined as 10 raised to the power of a googol. Written in full, it would be a 1 followed by a googol of zeros. Even a googol is vastly larger than the estimated number of atoms in the observable universe. A googolplex is so large that writing it out in full would be physically impossible.

For two Stanford graduate students attempting to organize an expanding internet, the word was an almost perfect thematic fit. A googol suggested a scale far beyond anything Page and Brin could realistically expect to index.

The misspelling happened inside a domain search box

Sean Anderson typed it wrong, and Page liked it enough to keep it

A widely cited account of the naming session comes from Stanford researcher David Koller, who reconstructed the story using recollections from colleagues in the office where Google was born, including Sean Anderson.

According to that account, Anderson suggested “googolplex” while brainstorming with Larry Page. Page preferred the shorter “googol,” so Anderson turned to a computer to check whether the corresponding domain was available. He accidentally searched for google.com instead. The misspelled version was available. Page liked its appearance and registered the domain within hours on September 15, 1997.

There’s an important caveat. Koller recorded the story in 2004 from Sean Anderson’s recollection, several years after the brainstorming session rather than from a contemporaneous account. The precise exchange may therefore be impossible to reconstruct, although the mathematical inspiration behind Google’s name is not disputed.

After investor Andy Bechtolsheim wrote a $100,000 check made out to the still-unformed Google Inc., Page and Brin incorporated the company in September 1998, beginning Google’s transformation from a search startup into a much larger technology company. They soon moved into a Menlo Park garage owned by Susan Wojcicki, who would later become CEO of YouTube. By then, the misspelling had already become the company’s permanent identity.

The typo gave Google a name of its own

A real word would still have worked, but the invented one carried less baggage

You see, under US trademark law, brand names are often discussed in terms of a spectrum of distinctiveness. Generic terms cannot function as exclusive trademarks, while descriptive terms generally require evidence that consumers associate them with a particular company. Suggestive, arbitrary, and fanciful marks can receive stronger protection because they are inherently distinctive.

Real words are not automatically weak trademarks. An ordinary word can become a powerful arbitrary mark when it is used in an unrelated market, as “Apple” does when used for computers. “Googol” would therefore not necessarily have been unprotectable. Because it suggested enormous scale without directly describing a search engine, it might still have qualified as an inherently distinctive mark.

The misspelling nevertheless gave Page and Brin a name with little pre-existing meaning attached to it. “Googol” already referred to a mathematical quantity, while “Google” had no recognized meaning in the search market before the company adopted it. That potentially made the altered spelling easier for consumers to associate almost entirely with one company.

Whether this produced a substantial legal advantage is impossible to prove, since the correctly spelled name was never tested as Google’s brand. The typo did not rescue the company from an unusable name, but it gave it a more self-contained identity than “Googol” would have provided.

Turning Google into a verb put the trademark to the test

763 domain names and a run at taking Google’s name away

Even a highly distinctive trademark can become vulnerable if the public begins using it as the ordinary name for an entire category of goods or services. This process, known as genericide, turned once-proprietary names such as aspirin, escalator, and cellophane into generic terms. Google therefore encouraged people to write “search with Google” rather than treating “google” as a general synonym for searching the internet, although controlling everyday speech proved far harder than controlling a corporate style guide.

The company’s concerns eventually reached federal court. An entrepreneur named Chris Gillespie had registered 763 domain names combining Google’s trademark with other terms, including googledisney.com. Google recovered the domains through arbitration, after which Gillespie and David Elliott sued to cancel the trademark under the Lanham Act. Their argument was that “google” had become the generic term for searching the internet and could therefore no longer belong exclusively to one company.

The argument had intuitive appeal because people routinely use “google” as a verb, sometimes even when referring to one of the alternative search engines that can find what Google cannot. However, the US Court of Appeals for the Ninth Circuit rejected the claim in Elliott v. Google [PDF] in 2017. The court explained that using a brand name as a verb does not automatically make it generic. The relevant question was whether consumers understood “Google” as the name of an entire category of search engines or as a brand identifying one particular company.

Elliott and Gillespie showed that people frequently say they “googled” something, but they failed to prove that the public regarded every search engine as “a Google.” Someone might say they googled a subject using Bing, but they would not normally describe Bing itself as a Google. The Supreme Court later declined to review the case, leaving the Ninth Circuit’s decision in place.

Google therefore kept its trademark despite becoming part of everyday speech. The case did not prove that “Googol” would have failed as a brand, but it showed that consumers still associated the coined spelling with Google itself.

The word they meant is still standing in the parking lot

Google’s Mountain View headquarters has long been known as the Googleplex, a name that combines “Google” and “complex” and echoes “googolplex,” Milton Sirotta’s second invention. The mathematical name Page and Brin originally reached for never disappeared. It simply migrated from the search engine to the surrounding buildings.

A spelling mistake did not make Google inevitable. PageRank provided the breakthrough, while funding, infrastructure, and years of relentless execution turned the project into a global business. The typo gave the founders something they had not deliberately sought: a distinctive spelling with no established meaning in the search market.

Sean Anderson went searching for a number and accidentally found a trademark.