← all stories
1994–1996IPO · 1996

Yahoo!

Jerry and David's Guide to the World Wide Web

An internet directory that two Stanford graduate students assembled by hand as a hobby in a university trailer: lists of favorite sites sorted by topic, anything to avoid writing their dissertations. Out of that hobby grew Yahoo, a free reference site that, thanks to free placement in the Netscape browser, became one of the main front doors to the internet of the mid-1990s and went public in April 1996 with a 154% rise in the stock on its first day of trading.

Founders Jerry Yang · David Filo
Domains akebono.stanford.edu/yahoo (the historical version, 1994–95) · yahoo.com
directorysearchadvertising-modelmedia-companyportalcurationnetwork-effects

In brief

Yahoo began not as a business but as a way for two Stanford graduate students to avoid writing their dissertations: they gathered a list of interesting sites by hand, sorted it by topic, and put it on the net simply because that was more fun than research. In two and a half years that personal list turned into the most visited site on the emerging web, thanks to the largest browser of the day making it the default "Directory" button, and in April 1996 the company went public with one of the best debuts of the decade.

How it started (the founders)

Jerry Yang, a Taiwanese immigrant who moved to California at ten and grew up in San Jose with his mother and younger brother, and David Filo, a native of Moss Bluff, Louisiana, from a semi-communal settlement where several families shared a kitchen and a garden, met in the electrical engineering department at Stanford: Filo was a teaching assistant on a course Yang was running, both were writing doctorates on electronic design automation (EDA), and together they taught students in Kyoto. Back from Japan, they settled into neighboring cubicles in the same university trailer, which a friend described as oxygen-deprived and which someone called a cockroach's idea of Christmas.

The idea was not a startup plan. Mosaic, the first genuinely mass-market browser, appeared at the end of 1993, and both of them began spending hours wandering the net instead of working on their dissertations. They kept their bookmarks in an ordinary Mosaic favorites list, but the list grew so quickly that the browser could no longer handle sorting it. So they wrote their own scripts in Tcl/TK and Perl to group the links by topic, and called the resulting page "Jerry's Guide to the World Wide Web." Yang himself states the motive without dressing it up: "We just wanted to avoid doing our dissertations." They set themselves a target of reviewing and categorizing at least 1,000 sites a day, by hand, with no automation at all: "No technology could beat human filtering," Filo insisted.

By September 1994 the directory already held more than 2,000 sites and was taking about 50,000 hits a day. Growth kept accelerating: by the end of that same year the figures, depending on the measurement, range from hundreds of thousands to several million hits a day estimate, and Yang's dissertation adviser, noticing the abnormal traffic, told them outright that the hobby would have to come off the university servers. By Yang's own account, Stanford's system administrator arrived at the same conclusion independently: "You're getting this big sucking sound — all this traffic is coming from this corner of the building." A few months later Filo and Yang abandoned their academic careers without finishing their dissertations, by one account with barely half a year left before the defense estimate.

Year-by-year timeline

Lesser-known but significant facts

  1. The first specialized category in the directory's history was created personally by Jerry Yang, in one evening, off one letter. In December 1994 the entrepreneur Eric Ward wrote to Yang asking him to add a "web promotion" category, which did not yet exist. Yang answered the next day, "i will create a category for web promotion and place you in it," and Ward's company became the first listing in the new category.
  2. The slot Netscape gave Yahoo for free in 1994, as a favor, turned within three years into a whole ladder of ever more expensive paid contracts, and Yahoo was itself paying in order to spite a competitor. At the beginning of 1996 Yahoo agreed to pay Netscape $5 million a year to become one of five services (along with Excite, Infoseek, Lycos, and Magellan) in the rotation on the "Net Search" page, a slot that had previously belonged exclusively to Infoseek. As CEO Tim Koogle admitted: "We did it to hurt the competition. After the first month, Infoseek's traffic dropped in half." And in March 1997, per the company's own 10-K, Yahoo signed two separate new agreements with Netscape: "Premier Providers" (one of four, $3.2 million in cash plus $1.5 million in advertising services) and a joint product, "NETSCAPE GUIDE BY YAHOO!," with its own "Guide" button on the Netscape Communicator toolbar; for the right to that, Yahoo paid a one-time trademark licensing fee of $5 million and guaranteed Netscape $10 million in advertising revenue for the first year and $15 million for the second. That is, the free "Directory" button of 1994 and the paid "Guide" button of 1997 were not one and the same contract that had simply gone up in price, but different products at different stages of an escalating relationship.
  3. The investor literally threatened to tear up the check if the founders dropped the unserious name, and the founders later repeated that same principle with a candidate for CEO. When Sequoia Capital was closing the deal, Yang proposed changing the name to something more businesslike. Moritz answered, "Well, I'm tearing that check up that I just gave you if you change the name." Later, choosing their first professional CEO, one of the candidates announced that the first thing he would do was change the company's name, and Yang and Filo turned him down (in Fortune's phrasing, "showed him the door").
  4. Yahoo's Japanese subsidiary was working three months before the parent company itself became public. The Yahoo Japan joint venture with Softbank (60/40 in Softbank's favor) launched in January 1996, and the American Yahoo's IPO would only happen in April.
  5. Yahoo pulled pornographic sites out of the directory not on ethical grounds but because of dead links. Interviewers in May 1995 put the question directly: "Is it true that Yahoo stays away from sex-based URLs?" Filo answered: "Well, we don't have pornography… But the real reason that we actually took pornography out was because any time we listed a site that had it, the next day… it would be down." Such sites went down instantly under the influx of traffic, leaving dead links in the directory, and user complaints about the content itself were only a secondary motive.

The first growth lever

Yahoo's first and decisive growth lever was not a single mechanic but a bundle of two: the product itself (a directory a human curates by hand and that you can therefore trust) plus free distribution through somebody else's already warmed-up platform. Filo stated the first part as bluntly as it can be stated, "No technology could beat human filtering," and hand curation was exactly what set Yahoo apart from competitors like WebCrawler or from wholly automatic search indexes: the user got not a million irrelevant matches but a list that two specific people had judged worth the attention. Curation alone was not enough, though. Traffic to a directory running on a couple of university workstations was growing exponentially before any distribution at all: from ~2,000 sites and 50,000 hits a day in September 1994 to figures that different sources put in a range from hundreds of thousands to several million hits a day by the end of that same year estimate, and it was this organic growth that became the main argument in conversations with investors and partners.

The second part of the lever was free placement. When Stanford demanded that the service come off the university servers, Marc Andreessen wrote to Yang and Filo personally: "Why don't you come on into the Netscape network? We'll host you for free and you can give us some recognition for it." Later, when Netscape shipped its Navigator browser, the "Directory" button in the menu bar led to Yahoo by default, which meant that every one of the millions of new internet users who installed the most popular browser of the era got a direct, free path to the site of two graduate students. By May 1995, according to an interview with the founders themselves, the number of links in the directory had grown to roughly 39,000 and traffic to on the order of 2 million hits a day and 200–250 thousand users, at a rate of growth they themselves described as a doubling every month. It was this free placement, and not paid marketing, that secured Yahoo the position of de facto front door to the web long before the company sold its first advertising banner.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. A product can be built out of boredom rather than for a market, and the market sometimes shows up on its own. Yahoo did not start with demand research: two people were solving their own problem (the chaos of links in a browser), and only when other people's traffic to their hobby grew until the university servers gave out did it become clear this was a business.
  2. Manual, non-algorithmic labor is not a temporary stopgap until automation arrives; sometimes it is the product itself. Filo's position, "No technology could beat human filtering," was not an excuse for a shortage of resources but a deliberate bet on curation as a source of trust, and that is exactly what set Yahoo apart from the purely mechanical indexes of its competitors.
  3. Free distribution from someone else's platform is the cheapest growth channel, and also the least stable asset. The Netscape button gave Yahoo millions of users without a dollar of marketing, and then that same button turned into a competitive auction at $5 million a year, because the platform had grasped the price of its own distribution. You can build a business on top of someone else's free channel, but you have to count in advance the day it gets expensive.
  4. A decision about monetization can and should be postponed, but not forever. Yahoo deliberately dragged its feet on advertising, discussed a possible backlash, and even polled users before the first banner, but in the end it launched advertising less than a year and a half after the founding, as soon as organic traffic was big enough to sell against.
  5. The name and the brand matter more than they seem to a technical founder. Yang himself wanted to replace the unserious Yahoo with something more solid ahead of the investment, and it was the investor who insisted on keeping the whimsical name. Sometimes the right brand instinct sits not with the product or technical founder but with the person who looks at the company as a media asset rather than as a tool.

Discrepancies and what we could not verify

Sources (primary first)

Primary (period documents and the founders' own words):

Secondary (academic case studies, retrospectives, and aggregated sources):

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →