Hi Martin thanks for participating in this interview. You guys interviewed us a couple of weeks ago right before you guys added .IconJar files to iconfinder.com which is absolutely great. Now that your users know who we are it’s time to let our users learn something about you guys.
I wrote the first code in 2007 while studying at business school in Copenhagen. I was working as a web designer mostly redesigning webshops – specifically with the goal of increasing sales and conversion rate. I was becoming frustrated with the lack of good sources of icons. There were just a few places such as iStockphoto and Shutterstock that had icons for sale but many of them were poorly designed and expensive. Also, you had to purchase a set of maybe 20 icons in order to get the one you needed.
In the open-source community Linux was distributed with different themes that also included icons. I checked the licence for these icons and they were mostly under the GPL license, which allows for commercial use while giving credit and passing the license along. These icons from Linux distributions created the foundation for the first version of Iconfinder.
I bought the domain Iconfinder.net because the .com was already taken. I wanted to explain the concept of the service in the domain and use a design as simple as Google’s. It should be a straight to the point site for finding icons — nothing else.
That first version got a lot of attention after trending on Reddit, Digg and Hacker News. However, this spike in traffic made the servers slower and slower until the hosting company shut down the site because the load affected their other customers. I desperately tried to get the site back up to get the benefit of having the attention of the Internet. This was in the middle of my exams so while trying to write my bachelor thesis I was juggling web server setup.
After the initial launch (and downtime), I moved to better server setup at Mediatemple and I started to regularly add new icons that I would find on blogs – basically spending one full day every weekend adding icons. All icons on the site were free and more and more users started using the site.
I decided to double down on Iconfinder and during the summer of 2009 I rewrote all the code with a much better architecture that could handle the increasing load. I reached out to a great company called TurboMilk and asked them to do a mascot for the site. I went for the robot — mainly because I love Pixar movies and Wall-E in particular.
With the new faster site and improved design the user base started growing quickly. The early spikes in traffic were suddenly dwarfed by the growth.
The growth of Iconfinder.net was noticed by people and the value of the .com domain Iconfinder.com was increasing for me. I wrote the owner of the domain in 2008 asking for the price. He wanted $5500 for the domain and at that time I was still a poor student who didn’t have $5500 for a domain name. Two months later the owner wrote me again and told me that he was receiving offers for the domain. He gave me one week to pay him the $5500 otherwise he would sell it to someone else. Yikes!
I desperately called the clients I had done freelance work for, asking them to pay me money in advance for any web design work they might need in the coming months. The owner of the company Clio Online knew about Iconfinder and agreed to purchase three web designs from me and pay me in advance. I had the money in my bank account by the end of the week and transferred the $5500 to the owner of the domain and got it.
We don’t really focus a lot on branding to be honest. But everyone on the team strive to do their best and want to succeed with our mission “Helping designers make a living doing what they love”.
There are still many areas where we don’t have any products e.g. icon fonts and services targeted at developers. So obviously we are not the leading in all areas. But we are fully aware of our position and have long term strategy for building products for both designers, developers and marketeers.
We are also investing a lot more into the search engine itself – it will improve quite a lot in the coming months. We are soon reaching 2 million icons and it’s becoming increasingly difficult to find the right icon. Luckily our data is very well organized so we have lots of options of using machine learning to automatically tag new icons and create personalized search results. That will come in 2018.
Yes, we are looking at some tools that will make it easy to use SVG icons without downloading them. Basically just using a URL for the SVG and slap some parameters to set e.g. the color. Developers love icon fonts because it’s very easy to use – but it’s possible to create an equally simple tool for using SVG icons. That’s an area we are exploring right now.
I still love to both code and design and I get to do that maybe 50% of the week. The rest of the time is mostly about helping the team stay productive. So that’s planning the road map, dealing with investors, running salaries etc.
I have two small kids – one is 4 and the other is 8 months old, so I try to be with them as much as possible. I often put in 7-8 hours during the day and then work again for 1-2 hours in the evening. Every day I prfioritize to eat breakfast and dinner with my little family.
The IconJar format was well-received by the users. Obviously the IconJar app is a great companion to the service we can provide. We are looking into wider support for serving single icons in IconJar format. And who knows what the future brings :)