← all stories
1993–1999sold for ≈$45M · 1999

Match.com

The first genuinely mass online dating service (1995). Gary Kremen, an engineer, was testing the idea of internet classifieds, found that personal ads paid better than anything else, and together with Fran Maier built the product around attracting women. Kremen was pushed out of the company within a year, the business was resold twice (1997/98 → 1999), and Kremen's second idea, the domain sex.com, turned into a twelve-year legal war.

Founders Gary Kremen · Peng T. Ong · Simon Glinsky · Fran Maier (calls herself co-founder & GM; Kremen calls her 'employee #6')
Domains match.com
datingtwo-sided-marketplacemarketplace-cold-startfounder-ousteddomain-names

In brief

Match.com was the first online dating service to become genuinely mass: it was thought up by Gary Kremen, an engineer and entrepreneur who was testing the idea of internet classifieds and found that personal ads brought in more money than anything else. The strategy that turned a niche bulletin board for single people into a brand was built by Fran Maier, who bet on women as the scarce resource of the market. In four years the company went from an apartment with a single workstation to a resale for ≈$45 million in stock, yet by that point the founder himself had spent a year off the board of directors with no real power, had received a miserable $50,000 for his stake, and had seen his second idea (the domain sex.com) turn not into money but into a twelve-year legal war.

How it started (the founders)

Gary Kremen was born in Skokie, Illinois, took a double degree in electrical engineering and computer science at Northwestern University (1981), and worked as a computer security programmer at The Aerospace Corporation, where he first ran into ARPANET. In 1987 he entered Stanford Business School, and at the orientation there that same year he met Fran Maier, who was then working as a membership marketer at Triple A (AAA). After business school he founded the security software company Los Altos Technologies with Ben Dubin (sold in 1996), and in parallel, as early as the beginning of the 1990s, he began buying up domain names for the future categories of classifieds: jobs.com, autos.com, housing.com, and sex.com, at a time when registering a domain was free.

Kremen explained the idea behind the test product differently in different interviews years later, and there is no single version: in one he recalls 1992 and a letter from a woman, "I could charge other men like me to have access to this woman's email and her picture"; in another, the idea of scanning photographs at Kinko's and his experience with paid 900-number dating lines at $2.99 a minute. Both versions converge on the same thing: monetize the shortage of female attention on the almost entirely male internet of the early 1990s (in Peng Ong's words, "about eight, nine out of 10 people on the internet were men").

Kremen met Peng Ong, an engineer from Singapore, at the Software Entrepreneurs Forum; they spent about six months discussing the idea at each other's homes before founding Electric Classifieds, Inc. in San Francisco in 1993. The concept was broader than dating: computer matching across classifieds instead of scanning by hand, or as Ong put it, "it doesn't make sense for a human being to look through hundreds of classified ads." Market analysis (Kremen was already studying for his Stanford MBA) showed that personal ads produced the largest revenue even for small publications, which is why they decided to start with dating specifically. They were not starting from zero: the domain match.com was already in use by a small email service, and Kremen bought it for $2,400–2,500. Ong wrote the database system and Simon Glinsky the business plan, and at the end of 1994 Fran Maier joined; she calls herself "co-founder and general manager, 1994–1998," while Kremen calls her "employee #6" and puts the split of credit for the idea at "I was 90%, [Ong] was 10%," a telling divergence in how the participants describe themselves. Kleiner Perkins offered money in exchange for a merger with the future Excite.com without keeping Kremen as CEO; he refused, and closed a $1.7M seed round from Canaan Partners.

Year-by-year timeline

Lesser-known but significant facts

  1. The access fee was introduced not against the strategy of attracting women but as part of it. In 1996, when the company began charging for a subscription, the share of women in the base rose from about 10% to nearly 20%; by Maier's account the fee "qualified the guys," screening out men who were not ready to invest seriously, and that, rather than free access, is what made the platform more comfortable for women.
  2. Match.com was not an idea built from scratch; it was somebody else's micro-business, bought. The domain and a tiny email-based dating service on it already existed before Electric Classifieds; Kremen bought them for $2,400–2,500 instead of launching a product on a blank page.
  3. A co-founder left after watching five CEOs come and go in six years, and took the churn as a symptom of a deeper problem. Peng Ong left the company for exactly that reason and built his next startup, Interwoven, which reached an IPO with a market capitalization of as much as $10 billion estimate.
  4. As early as January 1998, right in the middle of the change of owner, the site was physically moving to a new design. A Wayback Machine snapshot shows a redirect to "www-new.match.com" under the corporate shell "Match.Com, Inc.," a rare piece of direct evidence of how quickly the new owner began rebuilding the product.
  5. There is no single version of the founding story, even from the participants themselves. Kremen puts his own contribution at 90% against Ong's 10% and calls Maier "employee #6," while she calls herself "co-founder and general manager." Credit for a startup's success is almost never divided evenly and almost never remembered the same way.

The first growth lever

Match.com's growth in 1995–1996 came not from a single mechanic but from a bundle of three decisions, the most important of which runs directly counter to today's common recipe of going free for the scarce side of the market. First, the product was designed around women as the bottleneck of the market from the very start: the positioning as "safe, anonymous, fun," the refusal of the word "personals" in favor of "online dating," anonymous handles, a separate employee for site safety, and Maier's personal veto on a weight-in-pounds question in the profile ("No way… lets use body type"); the whole product design was subordinated to the logic of "if we got the women, the men would follow." Second, growth was kept geographically concentrated: instead of spreading itself across the whole internet, the team deliberately built critical mass in several key cities, starting with the San Francisco Bay Area, so that a real user had a chance of finding a match nearby, an explicit strategy that Peng Ong describes. Only after reaching some 65,000 users in those cities did the company introduce a fee for access for the first time, in 1996.

The third and counterintuitive element: it was the introduction of the fee, and not its removal, that raised the share of women in the base from 10% to nearly 20%, because the fee cut off men whose intentions were not serious. Before that, monetization ran through a charge on the outgoing message ($5–8 from the sender), which was also a tax on the surplus male side rather than a formal policy of free access for women: there is no direct confirmation of gender-differentiated pricing in the sources (see Discrepancies) estimate. A separate PR lever worked specifically for fundraising: Kremen's personal story, that the founder of Match.com could not get a date himself, run with a photograph and flowers, helped close a $7.5 million round.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. For the scarce side of the market, going free does not always work; sometimes the exact opposite does. Match.com raised the share of women not through free access but by introducing a fee that filtered out the men who were not serious. Before you copy-paste the recipe of one side of the market for free, check whether your real problem is the low quality of contacts, which is solved by a barrier rather than by its absence.
  2. Critical mass is local, not global. If the value of the product depends on the density of users nearby (a city, a niche, a language), squeezing one market dry pays better than spreading resources across the whole world at once; thirty years later one of the projects in the catalog arrived at the same choice on its own (Poland instead of the world).
  3. The idea the money is raised for rarely matches the one that takes off. Electric Classifieds was conceived as a platform for ALL classifieds; dating was only the first test vertical, and it became the whole company, while the ambitions in cars and real estate were never realized.
  4. Being the founder and having control of the company are different things, and the conflict may be about strategy rather than product. Kremen was not fired for poor numbers, since Match.com was growing; he was pushed aside for wanting to expand into other classifieds when the investors wanted to double down on what already worked. A year and a half later the business was sold for $7–8 million, and a year and a half after that resold for ≈$45 million; the founder got $50,000.
  5. A founder's personal story is a working fundraising instrument, if it is true and on topic. The headline that the founder of a dating service could not get a date himself was not a random piece of viral luck; it directly helped close a venture round, because it strengthened trust in the product through the vulnerability of its creator.

Discrepancies and what we could not verify

Sources (primary first)

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

Secondary (reference works, retellings, books):

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 →