Q&A for Putting the Pieces Together: A Study in Digital Content Management

In 2018 the Princeton University Art Museum embarked on a project that leveraged an open-source headless CMS in order to incorporate a backlog of unstructured digital content into the museum’s existing data and publishing architecture. This presentation will discuss the methodology of the project touching on initial planning, decision-making, systems architecture, metadata schema, workflows, and future roadmap.


Unknown Speaker 00:00
Welcome, everybody, feel free to let us know where you're coming from in the chat. This is putting the pieces together for study and digital content management with Dan Brennan and Sarah Brown from the Princeton University art museums. My name is Susan Edwards, I'm just introducing you to this session. I am a now former board member of MC and I was on the board for three years. And I'm now stepping down. And I'm here to remind you that MC n is a nonprofit, volunteer run organization for growing digital capacity for museum professionals. We have a really deep active community engaged in year round conversations. And if you want to be part of that, you can become a member of MC n if you haven't already and join our special interest groups participate in our mentorship program. And you can even become a leader within MC and by being a cig special interest group chair or even join the board. I also need to thank Microsoft, our registration Assistance Fund sponsor, axial b Ignite sponsor from last week, I hope you all went to Ignite that was really fun. And all the other sponsors that are listed in the program schedule for making this really unique and really special conference

Unknown Speaker 01:07

Unknown Speaker 01:09
So today's session is a zoom meeting, we can all see each other. We're going to use the chat bot for questions. I'm going to try to stay on top of that. And you all should have watched the, I hope, the recorded session that Sara and Dan recorded in advance of this session, but if not, I think we can still enjoy the conversation. I'm gonna let Dan and Sarah introduce themselves. Go for it you guys.

Unknown Speaker 01:35
Thanks so much, Susan. Thanks for moderating this conversation. I'm Dan Brennan. I'm a developer at the Princeton University Art Museum. I work on all manner of things that have to do with collection data integrations between systems architecture. And in the context of this project that we're talking about today. I really manage kind of the systems behind it, sir.

Unknown Speaker 01:57
Hi, everyone. My name is Sarah Brown. I'm a collections information specialist at the Princeton University Art Museum. For this project, my role was primarily to work on the data model within our headless CMS system. And my day to day job at the Art Museum ranges from exhibition and publication support to records management projects having to do with the collection. So for those of you who have not been able to view our pre recorded session, I will do a quick summary. So you know what we're talking about. Um, so as Susan had introduced, this is all about digital content management at the Princeton University Art Museum. Um, so with this project, what we sought to achieve were a few different things on the first being finding a centralized storage for all of this content and resources that our museum has generated in the last 10 plus years primarily in digital format. We also needed a content management system to index these and relate them back to our collection and establish relationships to one another. We also needed to better surface these resources on online for our visitors and, and enhance and increase access to these resources. Also another part of this project, which I should also add, this was funded through IMLS. In 2018, we also were given funding to generate new content around the collection. So when I talk about content, this is wide ranging. It varies from audio tours, videos, and these videos can vary from artists talks to exhibition openings, panel discussions, an abundance of learning resources that our education team has created over the years. And these learning resources vary from family and gallery activities, or they're focused in K through 12, educational resources, higher education. And we also had quite a bit of analog content on these primarily were newsletters and magazine articles that were written by our publication teams in the last 40 years. So we had quite a bit of interpretive text around the collection and exhibition history. Um, so let's see, we also had quite a bit of semantic groupings around our collection that also had interpretive text talking about these sets of objects. So what we also summarized in our presentation, is how we began to put the pieces together of content, and how we found our content management system, and how this content management system already fit within what Dan, our museum application developer had already built up. You Museum, we talked a little bit about the relationships and data model that we came up with to accommodate these resources. Um, and lastly, we summarized where we were in this project. Now that we have cataloged the backlog of resources, we are now in the design and search piece of this project. So I will open this up to all of you. We're very excited to answer your questions and continue the discussion about content management. Oh, Dan will also drop our link into the chat. We also have a sandbox environment for anyone who wants to explore strappy, which is our headless CMS system. So you can explore around and see the data model and some of the digital resources we've already cataloged in there.

Unknown Speaker 05:55
Yes, I just dropped that into the chat. We have links to all the resources Sarah mentioned, and also an actual version of our CMS that you can interact with delete things, you know, because we like to live dangerously.

Unknown Speaker 06:12
Does anyone have a question? Feel free to just turn on your mic and ask it if you like. was asking if there's a password for the sandbox,

Unknown Speaker 06:24
there is I probably did not quite explain it's what's in brackets behind the link. But I will drop that into the chat as well.

Unknown Speaker 06:35
Oh, I see. MC n 2020.

Unknown Speaker 06:37
MC m user is the user MC n 2020 is the password.

Unknown Speaker 06:47
So Brenda is asking if you can search for all content across types that are associated with given artists and objects.

Unknown Speaker 06:55
So I'll tackle that first. Within the context of the CMS system that we're working with, no, the searching that you're doing is based primarily on content type. But the nature of the headless CMS is such that you can then rely on the API of the CMS to pull things out, create your own query routes, and then create queries across different content types, which is really why we went down this road is because that type of thing is what we need to do on our web platforms. So again, within the context of a content creator, or Content Editor, they kind of have to know how to create the relationships between content and we assume that they know where those things exist, they know how to access those things. It's really kind of programmatically on on the output side where we need to start querying across different types.

Unknown Speaker 07:48
And that was me and I just as a follow up, so the thing that we're sort of grappling with is a curator may change scholarship, retributive an artwork, do something, and it would be great to see, okay, well, what are all the audio stops that are associated? What are all the video? What are all the essays? And what are all the things that we need to do in order to reflect that change? So that was the reasoning behind it?

Unknown Speaker 08:15
Yeah. And and that's, that's a super common use case. And it's a use case for us as well. Right now, and again, like I was saying, the way that we would approach that right now is to sort of start from from the outside in, if that makes sense. And then, you know, there's some institutional knowledge that would go from inside out, and we'd meet somewhere in the middle, what you're talking about, though, in the long term sense, and I think from like a systems stability and staff turnover sense, you know, makes is incredibly compelling use case.

Unknown Speaker 08:56
Anyone, anyone? Well, I'll ask a question

Unknown Speaker 09:03
about other tools for content, assessment and inventory and auditing that you did that you needed to use in association with this project. And aside

Unknown Speaker 09:12
from or so I can talk a little bit about that. So part of the IMLS funding we received, we were able to hire a production assistant to do an entire sweep of what existed where these pockets of content were, but she also had an education background. So it was so great to have one person be able to not only inventory but also assess this content. Is this still relevant? Do we want this to show up in the new search and do we want to still make this available? So since we only had you know, one or two people working on this, Excel spreadsheets was just enough but after attending the air table session at MC n someone showed us their use for air table to do a whole content assessment. And I think tools like that would be so great. If, you know, you had so many more people in the process of assessing and approving, you know, certain pieces of content? Definitely. So, for us for such a small team, you know, spreadsheets was sufficient enough.

Unknown Speaker 10:27
Um, there's a question here in the chat from Elizabeth Peters. And I'm, I don't really understand what you mean, by shaping the package. But he, Elizabeth, do you mind?

Unknown Speaker 10:36
Can you turn on your mic? And sure, sorry. Um, so I feel like I heard a resonance of an issue that we have here at the Botanic Garden, which is we have a lot of very particular stories. And then we have like, data streams. So we have the all of the data about our plant collection. And that's very, it's kind of regimented, and it's used in a particular way. And then we have stories, that, you know, a content editor has specifically thought, you know, this is the meaning and the way it might resonate with an audience. And I'm going to, we're going to have some text here, but then we're going to have some a series of photos with deep captions, and then we're going to put in a video and it's like, it's very shaped. So so we have a bunch of like, kind of very deep, well shaped pieces, and then a bunch a bunch of very large, comprehensive volume have very shallow pieces of information. And, and really like kind of how do we meet in the middle. So I really appreciated your model as a way to be kind of like, here's a place where we're putting this all together. So that's a little more like the middle. But But if it's just, if it's created, because you've like pre determined a set of rules that will lay out a page with these attributes, then you lose that kind of storytelling hand in at the other end. So I'm not sure if it's really a question, but have you guys had thoughts around?

Unknown Speaker 12:10
Oh, definitely, I can speak a little bit about what you're talking about, like, we have very short interpretive texts that might just be, you know, here's, um, you know, interpretive text about a grouping of objects. And that's it. But then we also have these very layered talks where there's comparative illustrations or other relationships to different things. And that brings us into this whole realm of not only just, you know, indexing interpretive text, it's like we're getting into the digital publishing world of Yes, we need to capture how this original presentation was. But we also need to be flexible enough to be able to put this, you know, in a different way. And I think Dan can talk a little bit more to about finding the balance.

Unknown Speaker 13:01
Yeah, I mean, that's that, that's such a good question, because it's really at the core of this whole project, basically, is on and, you know, the idea of will things ever become fully API driven and automated to the point where no one is ever actually putting a human touch on these kinds of things. I don't think that's ever going to be our goal. But we understand that the nature of scale is such that we have so much content in our backlog now that we will never have enough people to actually assemble those handcrafted stories. So we need to enable a system whereby, in addition to the composite parts coming together in something of a narrative form, we can also pop out to a collection object page, and you can see all of those composite parts from all different stories that relate to that object without someone having to actually make that connection on both sides. If that makes sense.

Unknown Speaker 14:03
I think this point about digital digital publishing versus like a website is really interesting one and Aaron Richardson brought up this related question about whether you can decide to not use content without removing it. And this, which brings up the question of who decides, I mean, the website starts from the catalog of your organization.

Unknown Speaker 14:24
Sure. So Dan, do you know when the published flag was established in the strappy upgrade, was that when we did that just a few months ago, that they are

Unknown Speaker 14:34
sure, surprisingly recently.

Unknown Speaker 14:36
Yeah. So that opened up the doors, you know, we started this project, you know, almost two years ago not having this option to flag content as published or unpublished. And I think we do need to go back now and probably add in maybe some of our old content that we did not deem necessary to add into strappy So one example is, you know, we have a lot of interpretation and learning resources that are gallery based. And for those of you who don't know, our museum will be continuing to be closed, starting at the beginning of the year to build a new museum on the current site. So all of these very gallery directional driven content, you know, we thought, you know, there is some, you know, important text in there. But overall, do we really want to, you know, put this into strappy? So, yeah, having now this option to essentially archive it, I think has opened the door. But Susan, you bring up a really great question about what is the workflow now of who all approves this. It's definitely something that we're working on. And luckily, we have such a small working team involved, that we are able to, you know, bring in, you know, we know who to bring in from education and the production assistant for, you know, approving content.

Unknown Speaker 16:11
Chad Weiner has a question.

Unknown Speaker 16:15
Yes, yes. First off, Dan, and Sarah, I love this project. And kudos to you both. It's a problem. And it seems like, you know, almost every museum has, there's this explosion of content around our collections. And ever since, ironically, since digital channels have have have gotten so popular and so influential, a lot of that content is no longer managed, you know, that the opportunity is is is to really gather the information the the life of an object. And so Dan, you know, your mention of that object page right at its at its core? Wouldn't it be great to hold on to the context around it? You mentioned in the presentation that the cataloger was key for this project, to go back. And to find those bits of context, and to essentially wrangle them back into the into the system. on projects I've worked on, in the past, certainly, the project at Williams College, we actually had a good bit of that content as data, either in spreadsheets or in databases. One of the things I was looking at, as I was going through the sandbox, I was wondering if if there's a mechanism, even within strapping for bringing in essentially importing, you know, from a CSV from a JSON, what have you some of that, some of that material to get a head start. Before we would go in, you know, and do the catalog, but is there that capability?

Unknown Speaker 18:17
Um, hi, chat. Yes. Short answer. Yes. And we did do a fair amount of that, like a lot of what you see in that system, especially as it pertains to collection objects, Core Collection, object data, data about artists is the same. It's the same JSON that we use that we've been using for API for a long time, just you know, ingested not even ingested, but moved into a database that sits behind strappy. Um, there are also plenty of functions for things like spreadsheet imports, batch importing. And the reason we have not done quite so much of that really pertains to that kind of evaluation process that Sarah has talked so much about, where if an individual was combing through that content, we might as well train them on the CMS and have them imported directly. Sarah, do you want to add anything else to that?

Unknown Speaker 19:10

Unknown Speaker 19:12
I don't think so. But, um, yeah, it was definitely a long task. And I don't know about other people. But you know, as people come and go at our institution, certain people are involved on projects that were, you know, so vital, and then they just kind of get bogged down and lost in whatever learn tab. It's been. And so that took quite a bit of time, and quite a bit of quick interviews with people to Yeah, compile a whole list of what resources already exists.

Unknown Speaker 19:42
I mean,

Unknown Speaker 19:44
since I've been at the museum for my gosh, probably eight years now. The amount of yet digital projects and content that have been created in my time it's definitely big, and especially in the last eight months,

Unknown Speaker 19:59
right? I suppose the follow on question then to that is moving forward as new projects arise and need to be cataloged. Is there is there a structure to manage that? Is there a governance to around that?

Unknown Speaker 20:17
It's definitely something that we're all talking about, you know, our production assistant is still on a part of this project for the next few months. But exactly, that's such a, an issue that I know every institution deals with, you know, how do we properly delegate this new process to staff? How do we get by in? How do we, you know, show them the importance that this needs to be cataloged? And strappy now that it's a part of the workflow? So anyone has great input? What is it like free snacks, you know, for a walk through?

Unknown Speaker 20:49
This is how you do it.

Unknown Speaker 20:50
But it's definitely it's a challenge that I think lots of institution face, faces. How do we get everyone on board?

Unknown Speaker 20:59
It also, Sarah, if I can chime in, it gets to a thing that we didn't, you know, is sort of underlying all of this, which is that now we have to CMS is at least right now. Um, and so you know, there's a question before from Eric, that was about like, is the plan just to move off Drupal and move everything to this JavaScript framework? And presumably use strappy? As a CMS? We haven't fully decided yet. If that's the the endgame here. It certainly could be potentially. But you know, for now, we do have to take into consideration when we're starting new projects and creating new things. Like, what's the law? What are the legs on this? How far does it go? How far we expect it to go and make decisions based on that? Yeah,

Unknown Speaker 21:42
I mean, I think one thing that got, you know, our staff really excited about scrappy was, you know, we did, you know, a staff presentation and just showed them, you know, how transformative This is to have all of our content in one place and how it's connected to, you know, our API. So we did get, you know, some excitement. And you know, we showed them the ease of putting information in and like Dan said, it's a real balance of how much going forward? Do we change certain processes for adding content, in strappy versus just in Drupal? exhibition pages is a great example. You know, everyone has exhibition landing pages with a little summary of the exhibition and the credit line and things like that. So you know, with Drupal, you know, people who can get into Drupal and edit that is very limited versus strappy, where it's a lot more intuitive and easier to add an edit text. So that's just one example of where do you draw the line?

Unknown Speaker 22:40
So we have a question about analytics. How do you measure engagement with this new system? And I just want to also remind us, we have a little more than five minutes left, and

Unknown Speaker 22:54
I'll defer

Unknown Speaker 22:55
to Dan about the analytics.

Unknown Speaker 22:57
This is I was really glad to see this question. Because analytics engine in general, sort of the relationship between this CMS and streaming media and rich media and storing and delivering that is a question we've grappled with quite a bit. I mean, if you notice, in the CMS, our video content points to probably a variety of different hosting services at this point. And in terms of outputting, that and being able to present it on a page that doesn't, you know, conditionally that could work, but it's not the cleanest solution. So we really need to wrap our heads around kind of what we want to do in terms of delivering video. And then once we are able to do that I think we build backwards to an analytics based solution for the CMS. But the reality is, we're just not there quite yet.

Unknown Speaker 23:51
And Aaron, do you want to elaborate on your question about maintenance and governance? What what aspects of

Unknown Speaker 23:57
that? Are you?

Unknown Speaker 23:59
Sure, so I'm working on a very large scale project with involving Library and Museum and archive metadata. But so my question is, so when someone decides, oh, the content in that video is inappropriate or incorrect, or whatever? Who decides that? Who do they tell? Where does it go? How many people have to agree and and then, you know, if, you know, if the, I guess you're controlling where all all of this is published, but you do have like, a lot of micro sites. So I guess my my question is, like, if things move around with

Unknown Speaker 24:51
Yeah, I mean, that's that's a very interesting point. And thankfully, it's one that we haven't had to grapple with in reality, a whole lot. But there are You know, when it comes to content and governance of content, this is one of those things where because we are something of, first of all, we're couched within a larger university. And so in a lot of cases, we can kind of defer to the university infrastructure on what we should do in those cases. I didn't take our cues from that. Um, but you know, our content creators are really considered the owners of the content, they divorced from any notion of like the system or the platform on which that content is published. And so it's a human thing when we have to work through these issues rather than a systems issue. And the system solution always follows behind the human decision. Not sure that that was the most thorough answer to that question. But that's not

Unknown Speaker 25:55
to say it like

Unknown Speaker 25:56
love. Questions are coming back to like the human workflow questions. Um, so Chad has another question about our reliance on archive.org. Is that a good archiving solution?

Unknown Speaker 26:10
That's one of the things that we've actually that we tried. Um, because, you know, there are a couple of reasons we tried it. It was a free hosting solution. It allowed for embedding a player that we could modify and theme to the designs of our different platforms pretty easily. Um, there are some other issues where archive.org is obviously a nonprofit. There are some stability issues, not in terms of things being available, but in terms of streaming speed, and things like that, that we've noticed. And so it's one of a couple different solutions that we've tried, it probably will not ultimately be the one we landed on, honestly, we were looking at different solutions. Because we don't have we didn't have a CDN for delivering rich content at the time that we started this project. Since then, we're using cloud front in front of s3 buckets for a whole bunch of other things such that we could probably put a bunch more video content in there and be happy with it. That might be what we end up doing.

Unknown Speaker 27:15
We have about a minute and a half left. Are there any other questions? Brian, are you stretching? Are you racing? Okay, Instagram stories,

Unknown Speaker 27:27
or hashtag What? page I

Unknown Speaker 27:31
need you to? Could you mind coming on mic and explaining? I don't understand your

Unknown Speaker 27:35
question. Sure. Thank you. First, thank you so much for this important topic. for casual illness, or accessibility for viewers, is just Instagram stories or hashtags ever enough for museums for categorizing and archiving. Does any museum rely on just stories or social media for certain events, or experiences or collections?

Unknown Speaker 28:18

Unknown Speaker 28:18
I would, I would say, you know, we leverage those platforms for creating content, but from a long term content retention perspective. As long as those things are locked up in that platform, there's not much, you know, we can't carry on with the expectation that they'll be available to us. If we talk about migrating things out of those platforms, that's maybe a different story. But we in practice, we don't do any of that right now. So this is all hypothetical. All right. I

Unknown Speaker 28:47
think our time is up you guys.

Unknown Speaker 28:48

Unknown Speaker 28:49
Thank you so much for coming. And thank you so much for listening in and adding great questions to the discussion about content management.

Unknown Speaker 28:58
Thanks all.

Unknown Speaker 28:59
Dan, and Gina, you