The Manager of Collection Information at the Metropolitan Museum of Art came to speak to staff at the Art Institute of Chicago. Ms. Choi spoke about the history of data at the Met, and the long road to unifying information and providing broader access. This comes at an important time at AIC, when we are reevaluating our digital asset management system, and continuing to work on improving legacy data.
It was fascinating to hear the story of cataloging at an institution similar to ours in age and collecting scope, but which is substantially bigger. In the mid-twentieth century, there was one department tasked with cataloging museum collections. This changed in the 1970s with the disbanding of centralized cataloging, which was a result of both budgetary constraints and a lack of network infrastructure throughout the sprawling facilities. So up until recently, every curatorial department - 23 in total - had its own system for cataloging their objects and standards for this work. There was no unified search system in place for all collections at the Met. With new leadership at the museum, there came a push to integrate these databases.
By the early 2000s all departments had to make the migration to the collection management system TMS (if they weren’t already using it), and by 2010 these were merged into one master TMS database. Because each curatorial department was working in a siloed way, there were dramatic inconsistencies in how items were cataloged. Artist names had variations in spellings - sometimes spelled in English, sometimes in their native language - and sometimes information about the dates they were alive and cities they were born and died in were included, sometimes it was not. Another notable challenge was the object type, which is now a required field. Often, the most notable characteristic of an object - a card being a baseball card, for example - wasn’t actually included in the object record. Artists and object type are two important ways of searching for objects in collection management systems, so these issues needed to be addressed.
There was also a question of how to address cataloging as it pertained to nationality, culture, and geography. Different curatorial foci had different ways of approaching this, especially given the broad timelines covered at the Met. In the end, it was decided that this concept was required for cataloging, but this information could still exist in multiple fields. It is a fairly concept set of ideas, and a flexible approach was embraced so as to not force equivalencies where they did not exist.
Some field terms did not end up being standardized: medium, dimensions, culture, object name, and classification. A substantial effort went into the unification of these siloed cataloging systems, and in the end, the team working on this effort had to weigh impact versus effort. When dealing with tens or hundreds of thousands of terms, it makes sense to focus time and attention on the heavy hitters. Also, existing controlled vocabularies were leveraged wherever possible, so fields like object name were linked to Getty Art & Architecture Thesaurus. Why reinvent the wheel when there are already widely-adopted thesauri applicable to collections like this one?
All of this work was made possible because it was an institutional priority. The director decided that this was a key initiative - that all the databases be integrated and data unified as much as possible - and appointed a working group to decide on cataloging rules. When there were disagreements, the director stepped in as mediator. Through this collaborative effort, with a high-level individual stewarding through and removing obstacles, the team was able to make substantial progress.
Now that there is a unified collections management system in place for data entry, search, and discovery, work is ongoing to continue to improve the quality of catalog records. Each curatorial department is still responsible for the cataloging of their objects, as the museum still does not have a centralized cataloging department. This is challenging due to the variation in departmental staffing and structure - some have staff dedicated to this work but others only have curators or collection managers who must also attend to a wide range of other duties. The museum also contracted with a vendor to work on tagging catalog records on the museum’s collections portal on their website. The museum providing instructions and training: tag what you see and when in doubt, don’t tag. This type of more informal data generation is still very subjective, and after quality control, some tags ended up being removed by museum staff. Where this metadata creation proved useful was in cases where catalog records were scant and basic terms could be employed for improved discovery - man, dog, landscape, etc.
Much of this work was undertaken to improve efficiencies within the museum. Conservation staff no longer needed to search across multiple databases when dealing with objects from multiple departments, for example. However, these efforts have benefited the public, as well. This data is pushed to the museum’s collections portal on their website. Greater consistency in data entry and the use of controlled vocabularies makes searching more intuitive for users. Additionally, analytics are being generated and evaluated to inform ongoing collections accessibility efforts and cataloging work at the museum.
The collection landing page has been redesigned with users in mind based on information gleaned from analytics. It now features groups of collections based on ideas (color, artist, subject) rather than by search facets alone. These groups are periodically swapped so visitors might experience something new each time they visit. Looking at scroll mapping, it has become clear that certain parts of object records on the website are not viewed - related objects, for example - and that there is generally too much scrolling per page. The team is working on redesigning these pages for a more compact and user-friendly design to improve visibility of this information. The Met also employs more active feedback mechanisms through user surveys, user testing, and a/b testing. Analytics can only provide so much information about how and why certain aspects of collection records are being viewed. Having opportunities to reach out to individuals using the site more directly grants staff useful insight into how they can make improvements.
Analytics also helps the museum to understand how the public is finding their way to collections records. While some individuals navigate directly to the collections portal, and search via groups, facets, or textual search, many others are redirected to the site. Pinterest is the number one referral service for collection pieces, and it has contributed to some fairly low-profile objects having hundreds of times more traffic than would be expected otherwise. The same goes for social media sites like Reddit, an object can be viewed more times in the span of a few days than would otherwise happen over the course of years based on one post to one of these sites. While the origination of this traffic may not yield many visits that extend beyond a one or two second click in and out of the site, it does help raise visibility for the museum’s collections. It also helps inform folks about the type of objects that museums like the Met collect and make accessible.
It was inspiring to hear about Ms. Choi and her team’s work, and how enthusiastic she is about connecting people - both inside the institution and outside it - to these collections via useful data. She stressed that this work is never done, and that perfect shouldn’t be the enemy of good. This was reassuring to me, as our department faces hundreds of thousands of our own records which need some serious attention and unification. I also walked away from the presentation thinking more about how valuable it is to provide access to these materials, with the hopes that they’ll reach new audiences and allow for new contexts to emerge for these objects. I’m so thankful she came to speak at AIC!