Henrik de Gyor: [0:01] This is Another DAM podcast about Digital Asset Management. I’m Henrik de Gyor. Today, I’m speaking with Monica Fulvio. Monica, how are you?
Monica Fulvio: [0:09] I’m very good, Henrik. How are you?
Henrik: [0:12] Great. Monica, how are you involved with Digital Asset Management?
Monica: [0:15] I’m the senior taxonomist at National Geographic Partners, formerly National Geographic Society, before our merger and reorganization. I worked with other area stakeholders to establish a single taxonomy, a set of topics hierarchically managed, and locations that match our cartographic policies, a list of the people that features our explorers, organizations we’re interested in, events, other topics or concepts that we discuss frequently across our content, regardless of media type, and manage this in a central taxonomy management system.
[0:58] We use Mondeca since we also have some ontological elements that we’re using, we use of these with their auto‑tagging system. This is served to our internal clients through a taxonomy through a business layer API that we’ve built internally, the taxonomy service.
[1:20] This in turn is consumed by a growing number of our content systems. Like a lot of organizations out there, we have a lot of content management systems. We are working on consolidating on our alpha content management system is AEM, but we have a number of other systems that we use and we’ll probably continue to use.
[1:44] This is true, if in an earlier phase with Digital Asset Management and any large number of systems, and less that we’re going to necessarily end up with a particular alpha system, but maybe a smaller pack of them at the end of this and little more consistency between them. I work with various internal clients to help them standardize their metadata.
[2:12] I’m mostly focused on this set of taxonomy field. So I give them a schema, a set of documentation about how to interact with this internal web service. I train users, often producers, many writers, or art editors, whoever is expected to do the tagging or view the tagging. I help train them.
[2:33] I often tag archives of content, especially when we’re doing large migrations. I tune and refine the auto‑tagger. I manage the taxonomy itself. I also work with our digital product team and other people who are planning how our end user facing product may be using this taxonomy.
[2:55] Right now, we are using a text‑only auto‑tagger. We’ve definitely talked about starting to use some kind of an image auto‑tagging service. I will say I’ve long kept an eye on those technologies and been a little skeptical about them, achieving the level of accuracy and granularity that we really need for describing our content. We don’t need to know that it’s a bird. We want to know that it is a snowy owl or a bird paradise. We need a certain level of granularity.
[3:30] I’d say that I’ve suddenly seen leaps and bounds in the accuracy of image auto‑tagging tools over the last year and we are looking around at them. I hope to be able to add that to our suite of metadata enhancement tools here, sometime soon.
Henrik: [3:54] Monica, how do you maintain the taxonomy for one of the largest nonprofit, scientific, and educational institutions in the world?
Monica: [4:02] It’s a lot of work. It’s one totally fair answer. I’m also assisted here by a fairly flexible system. Each node in the taxonomy is not anchored by name, it’s a UUID. That is its core element, where that UUID follows the terms through its lifespan, in all the systems that it’s used in, it stays that if we rename it, if we have to move it within the hierarchy, if we merged terms, things like this.
[5:55] We architect it, our taxonomy system, and the way the client systems interact with it for flexibility. That is one thing that helps us. Also, the system assists us because Mondeca, when it’s processing text will extract suggested terms…candidates. We also allow users in the various client systems to themselves post candidates, post it through the Web services and add it to the candidate queue inside of Mondeca, which I can review and, if appropriate, add to taxonomy or discover that I’ve missed a synonym or any number of suggestions that I can confer.
[5:18] Some of them are junk and some of them are great. I’d say also about the users…I work with internal stakeholders in all of these various topic areas. I’ve invited in a number of other central data owners around the organization to help manage their data sets where relevant in the taxonomy system. I frequently review sections of the taxonomy. I’ll put down in review the travel section with the folks who work on Traveler magazine and our other travel products, and make sure that these terms make sense to them, that I’m talking about what’s important to them, etc.
Henrik: [6:03] What are the biggest challenges and successes you’ve seen with Digital Asset Management and taxonomy?
Monica: [6:08] It’s easier to start with the challenges. There are many of them. I think particularly, we wrestle with I think a lot of organizations do, we have a number of different systems. For us, this is the inheritance of a lot of groups making their own individual selections using their own separate workflows.
[6:31] So in the fall of 2015, National Geographic reorganized itself. Previously, there was the [National Geographic] Channel, which was a company that was primarily owned by 21st Century Fox and about a third owned by the National Geographic Society. The magazine, the book publication, the map publishing, all of that was part of the Society as well as what I described as more traditionally, nonprofit activities, giving out grants, running educational programs, running a museum.
[7:03] They struck an expanded partnership with 21st Century Fox, which moved all of the media portions of the organization or consolidated in National Geographic Partners and the National Geographic Society owned a third of interest of that, but otherwise, a free‑standing organization that’s very focused on their nonprofit activities.
[7:27] There’s a lot of change and upheaval, in fact, there always is at the end of any kind of big company reorganization like this. Personally, I’m suddenly interacting a lot with the folks from our Channel and talking to them about how we manage broadcast content, both domestically and around the world. The Channel particularly is directly and…I think it’s forty odd markets.
[7:56] They’re working on consolidating themselves, which is exciting, to see the work that they’re doing. Broadcast media, especially for me, is a whole new world of systems and workflows that I’ve previously really been very deep in the magazine, book side and associated assets, and workflows, and systems, and challenges. It’s a little bit of a personal challenge, but it’s exciting.
[8:22] As I may be alluded to with the Channel stuff, I’m being asked to look more at how do we represent the taxonomy and how do we manage our content in global setting. Personally, this is super exciting, looking at making our taxonomy multi‑lingual, which will help us better serve our readers both in the United States and around the world.
[8:47] This is maybe a little self‑biased, but I’d say that acceptance of the taxonomy has itself been really quite successful. It allowed us to build consistent experiences that had alluded us previously. Things like consistent ad targeting. We’re using it for analytics as well now, dynamic map experiences, and we have some interesting work upcoming using the taxonomy for personalization and using the ontology models for… let us say light linked data to provide experiences.
Henrik: [9:30] What advice would you like to share with DAM professionals and people aspiring to become DAM professionals?
Monica: [9:35] My main piece of advice that I would share, both for taxonomists and aspiring taxonomists, and DAM professionals is to maintain the balance between standards‑based approaches, what is the best practice, how have other people solved this problem, can I lean on external schemas and standards and pragmatism.
[10:00] Particularly, working with ontologies and linked data, there’s this temptation to build this beautiful, complex, perfect model that is ultimately more for you to maintain and, in some cases, can be more than you will really use.
[10:18] This is both a piece of advice and something I’ve struggle with myself, how to best strike that balance of what is actually going to be used, what can we really get people to accurately and consistently apply in terms of metadata, what is the core data that we really need, which really at heart, involves answering the question, “What do we really need to do with this content, with these assets and with this data?”
[10:44] My other piece of advice, and something that is constantly helpful and fruitful for me, especially when I feel like I’m stuck is reach out to other people in the field. Conversely, any of you feel free to reach out to me. I’m always happy to talk to other professionals in the field.
[11:02] I think it’s reassuring that the problems that we’re often wrestling with, which can feel enormous, and maybe are enormous, but very frequently your colleagues in another organization are wrestling with remarkably similar challenges and sometimes coming up with answers that you haven’t thought of. I always find that enormously refreshing.
Henrik: [11:24] Great advice. Thanks, Monica.
Monica: [11:26] Absolutely. Thank you, Henrik. It was great talking on this topic.
Henrik: [11:29] For more on this and other Digital Asset Management topics, go to anotherdamblog.com.
For this and 170 other podcast episodes, go to anotherdampodcast.com.
If you have any comments or questions, please feel free to email me at email@example.com.
[11:46] Thanks again.