I started Soft32 in 2001, while I was still a student at the Bucharest University of Economic Studies (ASE). It was a software download site called Softrom.ro then, aimed at Romanian users.

The world it came from

Around 1996, my father bought a PC for our home in Sibiu. I learned to install and reinstall Windows, programs and, of course, games.

One of the best ways to discover software at that time was to buy a PC magazine. In Romania, that meant titles like CHIP or PC Magazine. They came with CD-ROMs packed with programs, updates, drivers, utilities and demos, whatever could fit on roughly 650 megabytes.

Internet access was slow and expensive. Dial-up had its own ritual: the modem sound, the pauses, the screeching, the failed connection, and then, if things went well, a very slow connection.

On a shared line, being online also meant nobody else in the house could use the phone. This was a good way to become unpopular.

A four-megabyte MP3 could take longer to download than to listen to. You played music in Winamp and opened compressed files with WinZip or WinRAR. You downloaded freeware and shareware. Sometimes you needed one program just to unpack another.

The first idea

In my first year at university, the library let students stay after hours to use the internet-connected PCs. That's where I came across a service that paid you to view banner ads through software installed on your computer. I installed it under my account on at least half of the PCs in the library, so whoever used them was viewing ads for me.

I think I reached the $25 threshold for a payout, but I never cashed the check. I realized the better model was to build something useful that people chose to visit, then make money from that attention with advertising.

As I explored the internet, I thought, why not build a site that would bring programs, utilities and games together in one place? A website could be kept up to date without waiting a month for the next issue of a magazine.

I spent the next six months teaching myself what I needed: HTML, CSS, ASP, databases, design, SEO and how to attract traffic. The first version was basic, but it worked.

I launched Softrom.ro on December 16, 2001.

Romania was only twelve years past the fall of communism, and the startup vocabulary I use today was barely part of my world. I didn’t think of myself as an entrepreneur, and I certainly didn’t think I was building a startup. I had a website to build and, later, a company to run.

That ignorance helped me focus. I cared about a few things: content, traffic, revenue and profitability. Almost everything else was secondary.

I had launched in the middle of the dot-com crash. The things that made the crash frightening to investors did not matter much to me. I had no investors, almost no money and, at the beginning, almost no traffic to monetize. I was just a student in Romania with time to build.

Finding a market and a business model

Softrom.ro started getting users, but most of them were from the U.S., and advertising revenue from the Romanian version was very limited. Both pushed me toward an English-language site. In 2002, it became Softrom.net. In 2003, I chose a more memorable name, Soft32.com, and incorporated ITNT, the company behind it.

We ran the business from Sibiu. I was from there, we had good people, and we didn’t need to move to Bucharest or Silicon Valley to run an international business.

I benefited from the early growth of affiliate marketing: earning a commission when someone bought something through a referral. At one point I even built Buy32.com, an affiliate storefront that listed what Amazon was selling, with purchases handled by Amazon.

In 2003, Google opened AdSense to self-service publishers. The program initially accepted English-language sites, so our move to English the year before had put us in a good position. Looking back, I was lucky with the timing.

Before AdSense, a small publisher like Soft32 had few realistic ways to sell advertising. The usual route was through advertising agencies or networks, and a small site from Romania was hardly at the front of that queue.

AdSense changed that. We could apply directly and place Google’s contextual ads on our pages. Google handled the advertiser side, and we earned money when people clicked the ads. AdSense became our main source of revenue in the early years.

By 2006, Soft32 was receiving 4.2 million visitors a month. In an interview that year, I was asked how much money I had started with. My answer was about $3 a month, plus internet access. The article’s introduction turned that into a business built with just $3.

In 2007, Google was sending ITNT a paper check by courier every month. We were running an international internet business, but the money still arrived as a physical check.

Our own advertising platform

The turning point was our own advertising platform for software developers, which we began building around 2010–2011. The business took off in 2012.

We worked directly with software developers and companies to promote their products to people already looking for software. The offers appeared both on the website and inside the installer. We were paid mainly when someone installed the promoted software, and occasionally for clicks.

In February 2013, for example, we partnered with AVG Technologies to include AVG Secure Search, a browser add-on, with selected software downloads.

We also earned commissions on software sales through the Avangate Affiliate Network, negotiating rates with developers and arranging discounts for users.

Higher revenue per user gave us more room to pay for traffic, mostly through Google AdWords. As long as we earned enough per user to cover the advertising and our other costs, we could grow profitably and reinvest.

Google had first been a source of advertising revenue. Now it was also where we spent money to attract users.

Google’s Dynamic Search Ads helped us advertise thousands of software pages without maintaining a separate keyword list for every program. In 2013, Google’s AdWords team reported a 15% increase in conversions within a couple of months. We were tracking software downloads as conversions.

In 2015, Google’s Romanian AdWords team wrote about ITNT’s expansion into 27 new markets over three years. For a while around that time, we were among Google's biggest advertisers in Central and Southeast Europe.

Growing the company

By 2013, Soft32 was averaging 10 million unique visitors a month, with a team of almost 20 people.

In 2014, ITNT ranked fourth in Deloitte’s Technology Fast 50 Central Europe, based on revenue growth of 1,433% between 2009 and 2013. It also appeared in the Inc. 5000 Europe rankings in 2015 and 2016.

This was Soft32 at its peak. I was asked more than once why I hadn’t sold. I still saw room to grow and wanted to keep the freedom to decide what to build next.

I knew more about attracting users and making money online than about managing a growing company. As I said in a 2018 interview, I gave people too little responsibility early on. For probably the first ten years, it didn’t even occur to me that learning how to operate and grow a company was something I should do deliberately.

I was learning constantly, just mostly about the product and the mechanics around it: SEO, caching, load balancing, servers and hosting. We started with dedicated servers and later moved our infrastructure to Amazon Web Services.

I did the same with marketing: AdWords, A/B testing, landing pages, acquisition costs and localization. We sometimes went surprisingly deep into local behavior. In Japan, for example, we thought pink was a good choice for a download button, based on studying local websites and how they used color.

I understood our niche very well. I understood much less about where Soft32 sat in the larger technology industry. I probably could have made some strategic decisions differently, especially around bringing in more experienced people earlier and planning our expansion.

When the world moved on

From around 2016, I could see the shift in the numbers. Organic traffic was declining; there were fewer downloads and fewer advertiser campaigns, and advertising revenue per user was falling. At the same time, AdWords was less economical for us as a way to acquire users.

More web browsing was moving to phones and tablets. App stores handled software discovery and installation on those devices. Attention moved, and advertisers followed.

Later versions of Soft32 included mobile sections, but the larger market was moving away from download sites.

The decline took years. Soft32 remained profitable almost until the end, but by 2025 the business was no longer worth continuing. My attention had also moved elsewhere.

In 2025, I closed Soft32.com. It had outlived its purpose, for me at least.