The Digital Museum: A Conversation between Loic Tallon and Julia Kaganskiy (Part I)
For a special two-part “Invention” interview, the Metropolitan Museum of Art’s chief digital officer Loic Tallon sat down with NEW INC director Julia Kaganskiy to discuss the recent developments in the digital role of museums and the future of their digital presence and influence.
In 2014, the New Museum founded NEW INC, the first museum-led cultural incubator. And in February of this year, the Met launched its Open Access policy, which made all the images of public-domain works in its collection accessible under Creative Commons Zero (CC0) licensing. That’s more than 375,000 images of artworks that are now available to use, share, and remix without restriction. (Read Part II of the interview here.)
Julia Kaganskiy: How did you end up at the Met?
Loic Tallon: I have been at the Met for about four years. I started there as senior mobile producer. That position really taught me a lot about helping people adapt to change: how to navigate colleagues through areas of digital discomfort by acknowledging what was and wasn’t being altered, and making scalable and sustainable strategic decisions within that context. Now, of course, no one needs to put the word mobile in their job title anymore, but mobile was one of the first disruptors in the digital space. I’ve been the chief digital officer at the Met for about a year. In that time I’ve restructured the Digital department into three functional teams: a collections information team focused on how we manage our data; a content team, to translate the stories of the museum into a digital space; and we’ve introduced a product management discipline to the Met with the formation of a dedicated product development team.
JK: What did you see as the big opportunity at the Met, the exciting challenge that you couldn’t wait to sink your teeth into?
LT: We have over 1.5 million artworks from 5,000 years of world history; our collection contains artworks that can inspire every person on this planet. The Digital team is privileged in terms of the collection we have available to us, and the team has the digital tools and expertise to make connections between our collection and people around the world. There are 3.7 billion people who we can reach online—that’s our audience—let’s go! That excites me. We reach about 31 million people through our website. We probably reach another 100 million through third-party platforms. We work with social-media outlets, Wikipedia, Google Cultural Institute, ArtStor and Khan Academy, and we want to expand our audience.
JK: So you won’t stop until all 3.7 billion internet connected people are connected to the Met?
LT: Until they’re connected to culture as a whole—why not? I think artworks are history. They teach us about who we are, about truths that I want to make accessible; I think that’s important, especially in today’s society. I love that someone might first encounter an artwork from the Met in an online space. They don’t need to know it’s from the Met; they just need to find the artwork and find the story. Over time, they’ll realize that this is an artwork from the Met, and that will help them find other stories which may be generated by the Met or enabled by the collection. But fundamentally, it’s about the connection between a person and an artwork. But no, I won’t stop until we reach all 3.7 billion.
JK: It sounds like this underlying mission is inextricably tied to the Met’s Open Access initiative.
LT: Other museums (like the Rijksmuseum, the National Gallery, and the Getty Museum) have done amazing things with their open-access policies. If an institution wants to increase its impact by, say, an order of magnitude, there are only a certain number of levers available to pull. One deeply interesting lever is foreign-language content. Another is making our content available on third-party websites and seeing that as success; for a long time, most museums saw success as traffic to their websites, but we’re redefining that charge as traffic to the museum’s collection, to the museum’s content, no matter where that content is. Another lever is building APIs to our content so third-party platforms can access it more easily.
And a fourth lever is simplifying and opening-up the licensing of the Met’s content to make it clear what people can do with it. I think our Open Access policy is one of the cleanest ways of doing that. Staff at the Met have been cataloguing and photographing the collection for over 120 years, beginning with writing library cards. Now, through our Open Access policy we’ve made all of that work available—more than 400,000 artwork records and over 300,000 images of artworks. The Creative Commons Zero (CC0) licensing is explicit: this is yours with which to play, discover, make, study, educate. The museum is really proud of the number of projects the policy has already made possible, and I believe we’re only just scratching the surface of ways people will use the Open Access content to meaningfully impact different communities.
JK: How did you get people on board with Open Access? I imagine that, despite the fact that there were other successful examples from [domestic and international] museums, you were probably met with skepticism, even resistance.
LT: Change is challenging, and there will always be doubters. A lot of it entails structuring the conversation, focusing people’s attention on how this helps us better serve the museum’s mission. Thirty years ago, museums were completely non-digital institutions, with probably a couple of computer terminals in a basement somewhere. Getting a non-digital institution to adapt to and feel comfortable in the digital space is no small task. The Met has an incredible history and collection and has been successful at what it does for a long time. Transitioning that success into the digital space just takes time. The Open Access conversation became meaningful when we started talking about how people access content today: every day, we all go to the same few websites; how many website URLs have you typed today?
JK: Like, two.
LT: Yeah, and the Met Museum was probably not one of them. Only a certain number of people are going to type in “metmuseum.org.” And God bless the several million who do. But most people come to our site through a Google search. In today’s digital age, it’s not realistic to just have a website and hope people will find us. We’ve got to put ourselves in the spaces where people already are; we can’t expect them to come to us. Articulated in this way, and understanding that the Met’s mission is collecting, preserving, studying, and presenting objects of art, the context of that presentation can include many different online spaces.
Fortunately, we benefited from the earlier museum examples. Like any big decision, it took a leap of faith. In an August blog post, I pointed to some early successes. I’m convinced the best ideas that will come out of the Open Access policy are the ones we’re not yet aware of. I always tell the team: Now that we have made the data available, now that the images are available, the best ideas are probably going to evolve from the community, and our role is to make sure that the data has integrity, and that it’s as accessible and consistently available as possible—and then seeing what people can build with it. How did Open Access appear on your radar?
JK: I’d been waiting for it. I remember learning about Piotr Adamczyk’s [the Met’s Associate Analyst from 2007-2011 and current Director of Image Content and Museum Partnerships at ITHAKA] early experiments with parsing the Met’s collection to reveal blind spots, from his workshop at the 2009 “Museums and the Web” conference. It felt like it was time [for the change], but I am aware of the difficulties faced by organizations when they endeavor to do something of this scale. I thought, “Good on Loic and his team for getting that shit done.”
LT: Ultimately, generating the best digital ideas from within the museum space is going be challenged. We have many highly-talented digital people working at the Met, but there will always be many more highly-talented digital people working outside of museums. We’ve got to work hard to get these people’s attention. If someone [in the digital realm] wants to do anything close to playing with art, playing with data, playing with machine learning, I want to make sure they see the Met’s Open Access flag waving and think, “Oh, that sounds like a cool starting point.”
JK: What are you doing to recruit those people?
LT: I’m doing interviews like this, getting more people aware that the Open Access data is out there. But remember, the policy is only six months young. We’re still learning how to build awareness, how to get people to understand the policy, and how to make the data as accessible as possible.
JK: I imagine that curators have nightmares or break out in hives over handing materials over to the community, because museums are tasked with stewardship, which is a protective stance. What you’re talking about is letting the chicks out of the hen house to go play in the woods, where God knows what might happen to them. We want information to be freely distributed, but this can be threatening to someone who espouses the responsibility of stewardship.
LT: Yes, and the curators at the Met take that responsibility very seriously, but the chicks have already left the hen house, to use your analogy. [laughs] Visitors’ personal photos of collection artworks are already available online: that is the reality. Through the Open Access policy we’re responding to that reality by making available the highest quality images of those artworks. There’s no doubt that some people will use these images for spurious purposes, but that will happen whether there is an Open Access policy or not.
People expect easy access to content. No matter if the Met made the images available or not, there will always be images of works from the Met’s collection online. But here we have the opportunity to put out the highest quality images and to work with partners—such as Google, Wikimedia Foundation, Pinterest, and Artstor—to make those the primary reference images that people have access to. By going Open Access, we are actually maintaining our standards as stewards of the collection: we are representing the collection in the best way possible, in a way that’s meaningful in a digital age.
JK: You said that you’re interested in seeing what the public does with Open Access, that some of the best ideas are going to be revealed through listening to the crowd. What is the process for feeding some of that thinking back into the institution? Where do great ideas come from? Does invention necessarily come from within? Where do you look for inspiration, when it comes to the next move?
LT: Our department is proactive in learning from the community and building awareness of different practices inside the institution. Even a simple thing, like improving how our search function works on a website, is an important discussion about what it means to uncover an object in the collection. We also invite speakers to talk to us about things happening in other sectors, other parallels we should be aware of. Most ideas, by the time they’ve reached us, were invented somewhere else, inspired by a myriad of things that happened over a period of time. The number of inspiration points are many, and while we might wrap our stories neatly (“because of this and this, my idea is as follows…”), they’re often much more complex than that.
At the Met, we’re trying to build a rich content platform from which other people can invent new experiences. With Open Access, someone could literally build the Met’s collection online. They could use a significant amount of CC0 content and make a competing product, and that keeps us on our toes. That will push us in terms of what we’re doing internally, which products we build, where we should be investing, what we think is unique about the museum, and what we can do better than anybody else.
Building digital products is difficult. Even major companies like Apple still struggle with building a map, for example; they will get products wrong. Doing so inside a museum is equally challenging, but we can tell good stories about our collection because we have access to the curators, and we have the objects. I would be very concerned if I felt the Met wasn’t creating some of the best content about the collection. I think others can create collection-related content, but we have a privileged position.
But if someone’s managing our data better than us, then something is fundamentally wrong. The data is the core asset: we need to carefully structure and manage the object records and the relationships between them. We’ve got to get the data right: only then everyone else can build from it. Right now, we use the data to build our own products, but I would not be surprised if we have a smaller product footprint in a few years time. Hence the value in letting others build from that same data.
JK: You mentioned Wikipedia, Google Cultural Institute, and Google BigQuery. What are the advantages and disadvantages of these collaborations?
LT: Right now, we are collaborating with the Wikimedia community to encourage the uploading of Met content onto the platform, but they could do it without us. With Google BigQuery, it’s the same; they are making CC0 data available on their platform and could do it without us. The advantages of these partnerships is that we can talk them through how the data currently works and let them know when we put updated data on GitHub, so they can run a query the day after and make sure their data is fresh. We digitize over 20,000 artworks every year. With BigQuery, the work was about making sure they have the freshest and most up-to-date data and that they have a workflow for accessing it. Now they can do what they want with it, and people can build upon that.
JK: What kinds of things do you wish people were doing with the data or the images? I’ve seen some interesting machine-learning experiments—which, depending on who you ask, will generate either a grimace or enthusiasm. What do you see as the untapped potential of releasing this data?
LT: The machine-learning component—how machines will manipulate large quantities of data and help us understand it differently—will be enormously influential. There have been two studies of the Met’s datasets, from which people have [extracted] trends of collecting patterns—by FiveThirtyEight and one of Google BigQuery’s analysts—and we learned something new from both. The implication is that if you put the data through a machine, there are trends we’re not even aware of—trends in artistic habits around the world—that a machine is able to extract, to explore, on a scale that is impractical for humans. I can’t wait for when image recognition can be truly combined with machine learning. Image-recognition [capability] is [becoming highly accurate] with 2D images, but the majority of the Met collection is 3D. When a machine can read those 3D objects easily—a piece of furniture, the Temple of Dendur, a Fabergé egg—that will be a milestone for us.
JK: What else are you working on?
LT: We’re focused on those big levers. We’ve made a good start with licensing. We’re still working to release the best data possible, so we’re building our partnerships and awareness that this data is available CC0. We’re doing a lot of work around our data structures, making sure that we’re managing all our assets well—audio and video content—and making that as accessible as possible. We’re also re-engineering the backend; we have a number of content-management systems that we are making talk to each other through a micro-service architecture, and that will will help us overcome some of our current technical challenges.
Foreign translation is an area that really interests us: how can we [ensure] that when someone in Russia is Googling for something of cultural relevance, they will find the Met’s content about that culture in their language? That’s a deep goal: to build our international audience.
JK: This merges with the accessibility-focused trends that I see in museums, [concerning] not just foreign languages but also the hearing-impaired and the visually-impaired, in a society with an aging audience that we hope to maintain. We’re leading a two-fronted battle: On the one hand, we’re trying to capture the younger audience and stay technologically relevant. On the other hand, our core demographic is getting older, and we have to adjust to their needs and abilities.
LT: Part of it is doing the fundamentals brilliantly. Twice in recent meetings at the Met, I’ve said, “Digital is not sexy.” I think there’s an assumption that the word digital means augmented reality or machine learning. But nothing is more important than doing the fundamental layer well. I believe we can build the Met’s online collection—the encyclopedic hub for discovering and navigating 5,000 years of world history—better than anyone else. Our goal is to fulfill that goal in the most rigorous, compelling and accessible way possible, based on the resources we have.
Continue reading this interview in “The Digital Museum: Part II.”