Founder story · 2000–2006
I started an online alumni network in 2000, before “social network” became everyday language.
It began with a domain name I noticed while building Toutelachine.com and a printed student directory I was helping produce at school. It survived the dot-com crash, found a business model, grew into a national network and was acquired by Classmates.com. More than an old startup story, it is where I learned most of the things I now ask founders about: focus, distribution, network effects, financing, monetisation and knowing when to change direction.
Where it started
The original observation
The school network already existed socially. The directory existed on paper. But contact information was changing faster than a book printed once a year could follow.
In 1999–2000, while working on Toutelachine.com, I was looking at domain names and noticed that trombi.com was free. The word came from trombinoscope, the familiar French name for a school directory containing names and photos.
The coincidence was perfect because I was part of the team dealing with the printed trombinoscope at Reims Management School. Email was spreading quickly, but not everyone had an address yet, and a paper directory was already outdated the moment somebody moved, changed a phone number, got a new email address or wanted to add another school.
The first product idea was therefore almost embarrassingly concrete: put the trombinoscope online, let people add an email address, and let the information stay current. Profiles, multiple schools and the more recognisable social-network behaviour came later.
A month after the idea there was a beta. The technology would make a modern engineer smile: an HTML form called a CGI script, the script wrote a new line into a text file, and a query searched those files to return results. It was fragile and completely sufficient to answer the first question: would anyone use it?
Version zero lesson
THE FIRST SYSTEM DID NOT NEED TO SCALE. THE IDEA NEEDED TO PROVE ITSELF.
The first real Trombi.com opened to Reims Management School in May 2000, then to other business schools. A simple viral campaign brought in hundreds of students. It was only a few months old when I took it back to someone I had met because of Toutelachine.
The first backer
At the Clics d'Or ceremony for Toutelachine.com I had met François Guilmoto, the entrepreneur behind Minitelorama. We stayed in contact. We first explored whether there was a business to build around China and concluded there probably wasn't.
Then I came back with Trombi. The project was roughly three months old, already showing the kind of school-to-school traction that made its virality visible. Guilmoto understood the network effect immediately.
Minitelorama helped finance the incorporation, provided office space and gave the project access to advertising inventory. In exchange it received a significant minority position. From today's perspective the terms were strongly investor-friendly; as a student founder with little negotiating leverage, I prioritised getting the company started.
The relationship later became a professional chapter too: I joined Groupe A Vendre A Louer as marketing director while continuing to build Trombi, working on the avendrealouer.com brand, real-estate marketing and the integration of legacy digital-classified activities.
The product kept changing
While working on Toutelachine.com I discovered that trombi.com was available. I was helping with the printed trombinoscope at Reims Management School, and the mismatch was obvious: a yearly book could not keep pace with email addresses and changing contact details.
The beta was an HTML form calling a CGI script and writing registrations into flat text files. It worked. That mattered more than elegance. A few weeks later the first real version launched for Reims Management School and then other business schools.
I had met François Guilmoto of Minitelorama through the Clics d'Or ceremony for Toutelachine.com. When I returned with a three-month-old Trombi showing early traction, he saw the network effect and Minitelorama helped turn the experiment into a company.
The first vision drifted toward a student portal: music, movies, jobs, roommates and other services. Then the bubble burst and many student websites disappeared. We narrowed the product instead of broadening it: from students to alumni, from portal to directory, from content to a very specific social job-to-be-done.
By 2002 the service was fully focused on former classmates. Profiles, school affiliations and the ability to update personal information turned a static directory into a living network. The product became more useful every time another person joined.
For years Trombi.com was free. Meetings with other alumni networks around Europe and the US made the weakness obvious: a network can grow and still not be a business. In late 2003 we began charging members who wanted to contact old friends. The core directory stayed accessible; high-intent communication became the premium value.
By 2006 the network had roughly half a million members. Classmates.com acquired Trombi.com in July. I stayed with the group, moved to Berlin and continued working on the French business before taking on broader European marketing responsibilities.
The messy middle
In 2000 the temptation was to become a “student portal”. Portals were what internet companies built: news, music, movies, jobs, classified ads, services. We even experimented with online roommate listings. More features felt like more value.
Then the internet bubble collapsed. Student sites closed. Capital became scarce. The broad portal logic stopped looking like ambition and started looking like distraction.
The strongest behaviour inside Trombi was much simpler: people wanted to find other people. They also added former schools to their profiles even when we had designed the service for current students. Users were telling us what the product wanted to become.
We followed them. By 2001–2002 Trombi shifted from current students to alumni, and from portal to directory/network. That is the version of Trombi people remember.
Business model
By 2003, I had met people behind several classmates-style services in Europe and the US: Friends Reunited, Passado, StayFriends, Classmates.com and others. Trombi was unusual because it was still free.
Those conversations changed how I thought about the product. If the value of the network was reconnecting with someone meaningful, then communication itself had willingness to pay. In November 2003 we began charging members who wanted to contact friends.
Today we would place this somewhere in the freemium/subscription vocabulary. At the time it felt more practical than theoretical: keep enough of the network open that it grows, and charge for the moment of highest intent.
More people → more chance of finding someone → more reasons to return → more valuable conversations → more willingness to pay.
Growth
A social product has an awkward cold-start problem: you can build the best search experience in the world, but it is useless if the person somebody wants to find has never registered. So growth was not a layer added after product. Growth was product quality.
We tried almost everything available at the time: school-by-school viral loops, student events, Paris Métro campaigns, an early Eurostar partnership, search, email, and large distribution partnerships. France Télévisions later linked its “retrouvez vos amis” services to Trombi. The goal was always the same: increase the probability that a search produced a meaningful human result.
By the time of the acquisition in 2006, Trombi had around 500,000 members. After the acquisition I stayed inside the Classmates group, moved to Berlin and continued leading marketing. The network then accelerated dramatically: Trombi publicly reported two million members in 2007 and four million in July 2008.
Clubic reported two million monthly unique visitors in 2008 and described Trombi as one of the leading French social networks. The demographic was also distinctive: the overwhelming majority of members were over 25, with a sizeable population over 50 — a useful reminder that social technology was never only a youth phenomenon.
The acquisition
Classmates.com acquired Trombi.com in July 2006. I could have treated that as the end of the founder chapter. Instead, I joined the acquiring group.
That decision took me from Paris to Berlin and from one French product to a European portfolio that included StayFriends in Germany, Austria and Sweden. I spent the next decade learning the other side of startup life: operating at scale, managing large acquisition budgets, coordinating teams and seeing what happens to a product after the founder transaction is signed.
It also changed my life outside work. Berlin became home. The founder story that started with a domain name in Reims became the bridge into the European technology career I still have today.
What I still use today
Almost none of Trombi.com’s original stack matters now. The operating questions still do. They are the questions I bring into startup mentoring and teaching today.
The durable idea was not ‘build a portal for students’. It was ‘help me find people I used to know’. Every time we moved closer to that job, the product became stronger.
The most important product decision was removing things. Movies, music, jobs and generic student-media ambitions made the product broader but less distinctive. The alumni focus made it legible.
School-by-school virality, student events, Metro campaigns, partnerships and later search acquisition all mattered because a social product only works when the people you are looking for are already inside it.
The first outside backing kept Trombi alive, but I entered it as a student with little leverage. I still use that experience when founders ask about dilution, timing and investor terms.
Afterword
What I like about the Trombi story is not that every decision was right. Many were not. The portal was too broad. We were late to some forms of promotion. Competitors had bigger media machines. The technology had to be rebuilt repeatedly. The business model took years to become obvious.
But the team kept learning from behaviour instead of defending the first plan. The project survived long enough to find its real shape. That is usually the part of entrepreneurship that gets compressed out of the retrospective version.
Twenty-six years later, when I work with researchers or founders, I am still interested in the same question: what is the smallest piece of real behaviour that tells us this should exist?
Sources & further reading