With Art.sy, the Art World's Amateurs Become Connoisseurs

Even with the vast resources of the web, art collecting is a consuming task. Even if you know what you want, it’s not always easy to find it; once you have something you like, it’s hard to find things like. Art.sy is trying to change that. With its Art Genome Project, Art.sy is attempting to catalog the entire art world, with a focus on education. In addition to its academic initiative, Art.sy is putting buyers directly in contact with galleries across the globe.

The new platform may have some kinks to work out before fully taking over the art sector, but it sounds promising. It’s been called the Pandora of art, which is an apt comparison as Art.sy’s own genome works in a similar fashion. Being able to look at a piece and immediately get more information, as well as scroll through more work like it, is a huge help to burgeoning enthusiasts, clearly Art.sy has a vision for the next step of art buying. But how do art and technology work together? And what does that mean for art galleries today? I talked to Daniel Doubrovkine, the head engineer of Art.sy, to find out what needs to be changed about the world of art.

Motherboard: For starters, how did you get into the whole buying and selling of art online? Obviously it’s a relatively new frontier. What led you to explore it?
Doubrovkine: I’m an engineer. I’ve been in software for the past fifteen years and I’ve worked for some big corporations, and some small startups. But my entire family has a background in art. Carter Cleveland, who founded the company, is the son to an art historian who, from what I understand, is the world expert on Tonalism. So Carter’s passion has always been about art, but he majored in computer science. His background is very similar to mine — I have a lot of artists in the family and he has the same. He’s always wanted to create something online for art. The mission of Art.sy has been to allow anyone with internet to access all of the art.

Is Art.sy democratizing art by putting the art world on the web?
Well, I think that the world today is online, but art, by vast majority, is not. So we’re just following the internet revolution as it’s been going on in many other industries. We are focusing on creating the tools for the current generation. For them, the internet is something they were born with. It’s the primary medium to access information, no matter what it is.

So are you guys putting the haughty galleries in Chelsea out of business by going online?
Absolutely not. Art.sy is a company that enhances and helps the galleries. It’s a discovery engine. Clients find works on Art.sy, but we refer them to the galleries. The average distance between a collector and a gallery is over 2500 miles for the works that have been bought on Artsy. This is a big world, and there are amazing galleries in Chelsea, but also in London, Paris, São Paulo. So, with today’s communication it’s kind of silly to restrict yourself to the locality in which you are. The internet is the tool that really makes it happen.

But aren’t you based exclusively out of New York?
We are not based exclusively out of New York. We have somebody in London, Berlin, Los Angeles, and so on and so forth. The way that the team is structured is we really have three parts to the company. First the Art Genome Project, which is run by art historian Matthew Israel, and the team of individuals who have an art history background. Second, we have the engineering team, which I run. Third, we have ‘partner-liaisons’ or the gallery team, whose job is to work with the institutions, the galleries, to learn about their programs, look at their shows, and help them to get theirs works on Art.sy. And they need to be local, because that’s essentially a local business. You have to be where the actual, original art is. They’re very engaged with the partners. We are just a tool for the partners, to show their works.

Tell me a little about this algorithm that you guys employ on your site when people search for art. When you search for something on Google, it narrows its results based on your past preferences and tastes. Does the algorithm that Art.sy uses narrow the possibilities for collectors looking at art?
Our algorithm as of today does not take into consideration people’s opinions, tastes, or selections at all. The way [you] see the search results is in a range from art that is historically relevant to interesting. What you want in a discovery engine is to show works that are very relevant to the one you are looking at. We have a large collection of art that is not for sale, which makes the Genome really rich. A lot of people who don’t know anything about art start with the Mona Lisa, so you see results that you could see at the Louvre hanging next to it. And then you start seeing more contemporary, interesting works.

We’ve worked on the algorithm so it’s a good source of discovery, and a few results are relevant and more results are more interesting. It doesn’t take into consideration who you are — it’s the same for everyone. But we’ve explored, and continue to explore how we can make the interesting works more appropriate for what your tastes are. We’re still working on that, and it will improve the genomic breakdown of the works you see. We work very much with experiments. We build a lot of software that never ships, that we never show to everyone. And we try it. We build something, we deliberate on it, we show it to users, they love it, we put it on the site. They hate it, we don’t put it on the site.

What’s one example of an experiment that didn’t work out?
We had an early recommendations engine, and our users clearly told us that it wasn’t good enough. It was a bit confusing, and they didn’t like it. So, we took it down. It was called the ‘Taste Test,’ and the way it worked was you would go through a set of works. For example, we’d show you ten works at random. Then you would pick one. We would use that work as a base and find other works that are somewhat similar to that one, and show you another set of ten. You’d pick another one. Then we’d continue refining. So what happened at the end is that you end up in a small corner of works because you’ve been narrowing down your tastes, and that stops being interesting. So, we spent a lot of time perfecting the algorithm to actually work, and it did exactly what it was designed to do, but that’s not what people wanted.

Art.sy has been getting compared to Pandora frequently in the media. In fact, Joe Kennedy, chief executive of Pandora helped to create the site. Would you say that this comparison is appropriate, based on how you’re describing Art.sy?
We have looked at how Pandora does things, and we are very grateful for Pandora opening its doors to how they created their recommendation engine. Their engine is very driven by individual taste and preference. There is a small percentage of music that is heavily liked by everyone — the very popular stuff. We really want to avoid that. We don’t want to narrow down the world to very few, very iconic pieces. We want this to be a balanced discovery engine. So the comparison ends there. And certainly our current search is very far from Pandora. But the concept is similar. If you like something, what else should you be looking at that you may potentially enjoy? What else do you want to learn about? I don’t know how much Pandora is used for educational purposes, but we have a lot of interest from academia.

So you guys are doing more than just putting art collectors and gallery in contact?
Yes. If you want to have a successful educational mission, you have to have a working business as well. We really believe that in order to create a sustainable, long term, amazing company, we have to do all of this. There is the very noble mission of art accessibility, and there is also the commercial mission of selling more art. I think that both work very well together. If we can succeed as a business, we will spend more time and efforts on the noncommercial missions of the company, because that’s part of our DNA. We have assembled this genome of art, some two dimensional art, some three dimensional art. It’s been used by academics. The quality of the data is extremely high. Some I know have been teaching classes by showing examples of an art movement, for example, from Art.sy.

Art.sy is infusing art and technology in an amazing way. Do you think that this enhances or detracts from the traditional art buying experience? Or is it a completely separate entity?
I think we all wish we had the time, the ability, and the finances to spend the entire day going to galleries around the world. So, I think it’s more of a necessity today to have an art collection on the internet as opposed to a deterrent to the experience of art.

I collect art. My price range is low, a few thousand dollars at the most. The first pieces I bought were kind of random. I went gallery hopping, saw something I liked, couldn’t get it out of my head, went back, dragged my wife to the place, and we ended up buying it. The thing I missed in this experience is understanding what I’m looking at. It’s because of the intimidating factor of these galleries. That’s one of the reasons people don’t buy as much art: people are afraid of asking and learning. Also, it’s overwhelming. I wish I had something like Art.sy when I first started collecting.

The current search covers the basics. I can look at the genes that explain the which movements this work is part of, I can look at some of the original content Art.sy has produced. The kind of information that you get from Art.sy, the environment around which you are looking at the pieces is very different online than at the gallery where it’s not sure if you belong. Once you learn the basics of a piece, then you want to talk to the gallery rep, and you want to maybe buy the gallery piece. We are filling the learning gap and making the process of buying much more successful. It’s hard to understand from people you don’t know, don’t trust, whether they’re trying to sell to you or if they’re actually telling you something you need to know before you acquire something.