Not Yet Netflix for Art

by Paddy Johnson on March 28, 2012 · 15 comments Reviews beta screenshot

Wouldn’t it be great if we had a Pandora for art? It would! You’d be introduced to hundreds of new artworks you actually liked and some you might even be able to buy. Cool. So, to accomplish that, we’re gonna remove the names of all the galleries you can purchase work from, make sure you can’t look at any video, and only include a fraction of the 20th century’s most important artists in our database. Wait, what?

WHAT IS IT? is an online database that recommends art to users based on their preferences. The site is still in beta but at present they list 3,000 artists from more than 200 galleries, museums and private collections — approximately 15,000 works in all. Most are for sale.

People care about the because, unlike the music or film world, in the art world there’s no site that makes it easy for users to find art they might like. Also, there are a lot of fancy people involved: Peter Thiel, the founder of Paypal and a board member at Facebook, and Jack Dorsey, creator of Twitter, are both investors, and the venture has enlisted high-profile advisors like Larry Gagosian, Pace Gallery President Marc Glimcher, and Joe Kennedy, CEO of Pandora.


So far, the answer has been has been VC funding, which isn’t exactly the same as making actual money. However, once the site launches, will work on commission-based sales. The company takes 15 percent on the first $10,000 of any purchase, and 10 percent of anything beyond that. In this way, it differs from the VIP Art Fair, which asks galleries to shell out for the limited-time use of their software, but charges no fee for any sales that occur.

Another difference between and the VIP Fair—and one we suspect will hurt in the long term—is the amount of verbiage dedicated to the art. Whereas the VIP Fair encourages galleries to upload oodles of press, videos, and CVs for the artist, almost none of this is made available on Only a small blurb under each artist tells users about their career; one 75-word biography begins, “American painter Agnes Martin built her oeuvre around the search for sublime beauty and serenity.” Frequently the “Additional Information” category is left empty or reads only, “Price includes frame”. Needless to say, there’s not much on this website that would sell a collector on a work that they weren’t already interested in.

Screenshot AFC


Overall, the site is fairly easy to use, which is something of a miracle when you consider the vague header menu. Want to search for an artist or art work? No problem! Just choose from one of the following headers, “Search”, “Browse”, “Artworks”, “For You” and “Collection”. We had no idea what any of them meant, but after a bit of trial and error we figured it out.

The site’s usability is helped by a design architecture that resembles sites we’re already familiar with. Their Netflix-style recommendation list is a good example of this, the working concept here being that if you liked this Mircea Cantor rainbow installation, then you might these six works using “symmetry”. As with the headers, the copy on this site could use some help; the words “if you liked X then you might like Y and Z” never appear, so it takes a minute to figure out that you’re looking at recommendations rather than just other related artworks.

One hopes that as works its way out of this early beta phase, they will develop better tags and search catagories. Right now, only basic sorting through color, size and filter combinations exists. You can search by one color, but not two, you can include add a filter such as “austere”, but can’t add an option like “emphasis on line” or “large brushwork”. As such, the search results that are brought back don’t feel that specific to the user.

There’s also a “taste test”, which asks users to identify seven pictures they like to help them determine their preferences. Again, it’s based purely on images, and never asks a number of crucial questions—chief among them a preferred size (a priority for many collectors).

Ultimately, the system seems wasteful. We take at their word that this is a goliath mathematical enterprise, and their job postings for developers, which refer to “real-time search in 800 dimensional space (one dimension for every art historical category)”, show that their algorithms are complex enough to be more specific. The user interface, though, fails to make the scale of that computation apparent.

Compare this to the taste preference tools at Netflix, which are known for their specificity despite being based on much simpler technology. AFC’s Will Brand has a recommendation category titled “exciting crime movies from the 1980s”, which seems highly specific but is merely a combination of three likely tags—exciting, crime movies, and the 1980s. Netflix looks a whole lot smarter than it is, and would do well to follow their lead.

Max Jansons, Rainbows ( They’re Not Just For Gay People), 2010, Oil on linen, 7 x 13 inches

THE GOOD is a ridiculously accessible site for art newbies, with a vast treasure trove of large, high-resolution images. This is amazing. The site is fairly easy to use and does a good job at making its users feel a little more sophisticated after having spent a bit of time on it.

Although the architecture of the site still seems crude, manages to introduce visitors to works and artists they wouldn’t have found otherwise. I found Max Janson’s “Rainbows” simply by clicking through the Contemporary Art category, so the site works. Notice that the boobs with rainbow nipples double as a viewfinder to the ocean. It’s like they magically knew what would appeal to me!

Speaking of which, my recommendations have always led with Canadian artists, even though I have never shared that I’m Canadian with the site. I’m about 60 percent sure that this is just luck—there are no options to connect Facebook or Google Plus—but this user, for one, enjoyed it.

One final note on the site: looks like it was designed by a developer, which is to say, it doesn’t look designed so much as it does laid out. This isn’t so bad relative to most art websites, which are either overdesigned to the point of uselessness or haven’t been updated since the early 2000s. screenshot


Oh man, where to begin. feels like it’s still in the early stages of beta, which wouldn’t be such a bad thing if the company hadn’t launched back in 2009. That’s an awfully long time to be tweaking algorithms, but I guess when you’ve got 7.25 million in venture capital you can afford the time.

A few basic observations: At the risk of stating the obvious, the search functions don’t work well enough yet. I talk about a lot of those details in “The Use” section of this review, but past all that, it’s worth noting that the database is still too small to be useful. There are only 15,000 works in the database, which sounds like a lot, but feels particularly incomplete because if the art’s not for sale, there’s far less chance the work will be in the database.

To wit, Chris Burden has 15 entries, but “Shoot”, the canonical performance in which he had his assistant shoot him in the arm with a .22 caliber rifle as a response of the Vietnam War, is not pictured anywhere. I can read about it in his bio, but I can’t add it to my collection. Worse still, Vito Acconci, arguably the most influential performance artist of the late 20th century, has no entry at all. Arguably, there’s a fair rationale behind these decisions– why waste resources on entries that can’t lead to commissions — but the company shouldn’t be calling itself the art genome project if it’s not making an attempt at being one.

It’s hard to overstate how big this problem is, but it may not be the company’s largest. Oddly enough, the only way to find out what gallery represents what artist is to contact an representative about individual works. I suspect this is the site’s way of making sure that they receive their commission, but it’s needlessly complicated and bad for sales. It took more than 24 hours to return my request for additional information, a push button service on the site that doesn’t even allow me to ask a specific question. The email I received from the company simply asked me to specify the nature of my request.

I suppose that kind of customized service will make for a good spam filter, but it’s not worth it. Galleries do a lot of work to place their artists in museum shows and commercial venues and their names are tied to their artists’ success. Failing to list the gallery’s name is a little like selling shoes and removing all the brand names. I don’t know how anyone ever thought this was a good idea.

Luckily, this is a relatively easy problem to fix, as are many on the site. For example, it wouldn’t take too much for to offer embeddable video in the video art section. All they offer is high-res stills, which, given the lack of text on the site, ensures that no video art will ever be sold through Also, it’s unclear if social features will be built into the site—none exist yet—but they certainly encourage use. At present you can follow an artist for updates, but you can’t follow or share collections with your friends. And finally, there’s the buggy “view my collection” function, which rarely loads all the images, and offers no sorting options. I can’t figure out how the slides are currently being arranged, other than that it’s not alphabetically, chronologically, or by date favorited.

That basic sorting like this is a problem for a website branding itself as the master-sorter does not bode well. Many of the site’s current issues may have easy fixes, but there’s enough of them to suggest that there may yet be months of work ahead before the site is anywhere close to ready for the public.

screengrab afc


Not yet. We like what is trying to do here, but if the company wants to succeed, they’re going to have to make a lot more changes. That should start with a way to make money that doesn’t get in the way of user experience and end with providing better recommendations and search. People like to know why images are being suggested to them because it tells them something about themselves. does nothing of the sort.


Ashley ernest March 28, 2012 at 8:30 pm

Artists should also be featured on late night talk shows such as #conanobrien ….

Sam D. March 28, 2012 at 8:45 pm

Hit the nail on the head, Paddy. No Acconci?! I’ll “Vito” my invitation request.

Marius Watz March 29, 2012 at 11:01 am

Seems like a good summary (disclosure: I haven’t seen the site myself). But, ahem, what about the unfortunate name? It makes me snicker every time, sounding as it does like a Web 2.0 startup transparently trying the clever-funny route that made Twitter a brand name. And charging 15% of a $10k sale sounds like a non-starter, gallerists like keeping their money too much.

Paddy Johnson March 29, 2012 at 11:47 am

Thanks! It’s too bad more people can’t see the site right now in terms of the timing of this review, but they really aren’t ready for it. 

As for the name, you’re right, but I don’t think it’s bad enough to have any negative impact. People made fun of the ipad for days, but no one thinks about it now. 

They really have a problem with how they plan to charge people though. 15 percent commission isn’t that much — it’s a standard consultant fee — but collecting it is going to be a bitch. They might do better renting their software the same way artnet or VIP does. Since a lot of these larger sales take place off line, it’s not like you can install a cookie and collect commission on a sale that occurs 30 days after the first view. 

mariuswatz March 29, 2012 at 8:36 pm

Good point – hadn’t even thought of how hard it would be to collect the fee. They’re wide open to the Etsy workaround known as “agree price on site then handle payment off-site”. There’s no way to prevent buyers from contacting galleries directly and ask for a 10% discount on the online price, which still would be a win for the gallery. 15% is much for a traditional dealer / gallery cut, but I don’t know if galleries will consider equivalent to dealers.

ps. iPad? I giggle every time. Actually, no I don’t, but Still pretty bad.

Marius Watz March 30, 2012 at 8:42 pm

As it happens Leonard Tirulnikov (@speednoise:disqus ) from was kind enough to give me an invite, so I’m able to revise my comment. Well, actually my two sticking points remain – the 15% commission still sounds high and the name still sounds silly. But despite my skepticism I found myself liking the site as such more than any “art portal” I’ve seen so far:

The graphic style is pleasant and unobtrusive, the navigation is for the most part smooth and dynamic. The simplicity of the “save to collection” feature is smart, although I wonder what happens once you have “saved” more than 20 items. Browsing and navigating by genre or recommendation system worked well for me, I was able to find several pieces I liked but were unaware of.

You already pointed to the biggest drawback: The database is heavily defined by the limited galleries that are represented. Hence media art is basically defined by Bitforms’ inventory, and most mid-career artists simply don’t exist. I wonder how (and if) they plan to overcome that limitation. A Wikipedia-like database of artists would be welcome but likely too ambitious.

Ultimately, it does end up suffering from the same problem as art fairs, although it is one that hardly seems avoidable: The numbing sensation of seeing too much sales-friendly art in a single sitting.

As per Steve Lambert’s recommendation I took a quick look at Sadly, the fact that my artwork is already in Kapsul’s database makes me like them better by default. Artists are such egomaniacs.

Artplus March 29, 2012 at 11:31 am

welcome to check out Art+ in beta:

Steve Lambert March 29, 2012 at 4:03 pm is not trying to do the same thing, yet so much better.

Marius Watz March 31, 2012 at 11:51 am

Kapsul looks promising, although it’s a fundamentally different concept than Kapsul finds a lot of sources, but many of them are noise. Any good kapsuls I should follow to get an idea of “best practices”? Preferably media art related.

ian curry March 29, 2012 at 4:41 pm

If you view the site through the lens of being an comprehensive, encyclopedic resource, it fails in the ways you describe. And maybe some people will see it that way. If you understand it as a discovery engine for finding new works you might like, based on what you’re viewing, I found it absolutely wonderful. Like you, I discovered some great new names and works I was previously unaware of. Most of all, I find myself coming back to it. The site, unlike every other art index site I have used, is a pleasure to browse around in. The interaction is smooth and the structure is simple. I found the breadth of their metadata categories pretty surprising. Categories such as “art that plays with scale” give me interesting vectors to explore that time/medium and other standard categories don’t. 

I was skeptical when I heard of the site’s grandiose computational claims, the high-profile supporters, and the piles of venture capital – but after trying it I really find it to be lovely and useful. Hoping others will give it a shot before passing judgement. 

Paddy Johnson March 29, 2012 at 10:50 pm

People will view the site through the lens of being a comprehensive, encyclopedic resource because that’s what they’re telling people it is. It’s not all bad, I just think it has a ways to go. 

Carter Cleveland March 29, 2012 at 6:03 pm

Carter (’s founder) here. Thanks for the feedback, Paddy. Our goal of making all art accessible to anyone with an Internet connection is very ambitious, and so balancing the desire for a perfect product against putting something out there and getting feedback is always a challenge. 
Getting more artists on the site is one of our highest priorities. The reason has taken so long just to get where we are today is because we do everything legally and in partnership with the artists and institutions providing the art. Rest assured, we are also big fans of Acconci and are working hard to add him (and many others) to the 3,000 artists currently on 

Recommendations are at the core of what we do but are very challenging to get just right. Netflix has been working on theirs for 15 years and likewise we see the Genome as a long-term investment. But what’s exciting is seeing computer scientists and art historians collaborating to build something that has never existed before and could potentially unlock the art world for a far greater audience. Already we are seeing art world and non-art world people alike spending lots of time on due to our recommendations. 
We will also soon be adding more information about the artist and their artworks. Having grown up as part of the Youtube generation, I’m particularly excited about video.Regarding the revenue model, our short-term focus is on creating value for galleries, museums, and collectors, and making monetization a long-term priority. Many of our investors’ companies such as Twitter, Google, and Facebook, have followed this model with great success.

Paddy Johnson March 30, 2012 at 10:21 am

Hi Carter,

Thanks so much for your comment. I take it the legal concerns here are image reproduction rights? Just curious what the those processes are, since it seems like it is such a big part of what you’re doing. 

On the subject of computer scientists and art historians working together, it’s not mentioned in the post, but one of the things I think your project is doing right is the individual attention that’s given to each entry. Going forward, I hope that approach will help. I also hope the site is able to make the computations its doing more transparent. The site is doing all this work and it seems a waste that so little of it is made transparent. 

As for the revenue model, I think that needs to be sorted out. That said, Artnet offered their services for free for a couple of years before they started charging, so I assume there maybe long term development plans in the pipe. 

I’m curious, do you see yourself as a competitor to any pre-existing art site? What websites would you identify as the models for your own?

Dan March 29, 2012 at 7:52 pm

The site looks like it was designed by a developer? That’s a low blow. I haven’t used the site yet, but judging from the screenshots, they have done an excellent job of emphasizing the work and allowing the interface to function as a tool in the background. User testing will dictate if they need to change their menu structure at the top. They are simply following good UI/UX patterns. 

Who cares about their monetization goals? They raised $7.25million. 
A lot of what was mentioned in the article are easy product fixes. They seem to be on the verge of solving something exponentially more complex and significant. 
This article is bleh. 

Paddy Johnson March 30, 2012 at 11:03 am

I think we can disagree about the site design without my comments being a low blow. There isn’t much design to speak of on the site, so I don’t think there’s anything wrong with that assessment at all. 

Also, 7.25 million in VC funds is not a guarantee for success. Colors raised 40mil and people know the powerpoint lecture that makes fun of the app more than they do the app itself. 

I’d be interested in hearing, what specifically, makes you think is on the verge of solving something significantly and how. This seems like an awfully bold assertion considering you haven’t used the site. 

Comments on this entry are closed.

Previous post:

Next post: