Our organization is reviewing its web content to ensure we comply with a statutory accessibility obligation. Our lack of information governance has meant that we are short on tools to manage the challenge. The fallback position, as it so often is, is to count pageviews and delete things that aren’t visited. It’s a short-sighted approach because it is using a measurement without considering the purpose of information. As it has been said before, “not everything that counts can be counted.”
Don’t get me wrong. I’m all for weeding both physical and digital collections. There is no need to create information resources that don’t create value. The question we should ask, though is this: is activity the only value that we can use to measure information access?
Rip Out All the Pages
I immediately thought of looseleafs. There are some volumes that get a lot of love and some that get none. Even within certain books, some chapters are more important than others. Should we throw away the unread pages because they’re unread? Or toss the volumes in a multi-part set that we have to pay to keep up to date, even if they aren’t used?
Can you imagine instructing loose leaf filers to only update – and remove – certain pages because no-one reads the other ones.
It sounds bonkers but that’s what a pageview approach does. It doesn’t look at anything other than (a) has something been used (b) within a given time frame. It doesn’t contemplate the value of an individual resource (page, volume, book, web page) within a broader context (a book, a set, a collection, a site).
I have often thought about the potential energy of a collection. A book on a shelf may have many purposes, one of which is its potential use in the future. There is a cost to having that book on that shelf. There’s rent for the space, initial cost, upkeep for a looseleaf, staff time to shift or maintain it. But it is selected because it plays a part in a collection.
I’ll return to cost in a moment.
Footfall and Visible Access
Foot traffic is another measurement that law libraries use that may or may not tell a story. When I think of foot traffic in the law library, I often think of the Inner Temple. They reported 150 visitors a day. That seems low to me, for a metropolitan law library. But who am I to talk? Our inability to accurately count foot traffic means that we don’t count it at all.
This gets at tools. If you’re going to measure something, are you just measuring for comparison or are you trying to understand. If visitors don’t swipe or sign in, how do you really know what your count means. We have gates that register people going in and out. A lawyer is the same as a self-represented litigant is the same as a wedding photographer is the same as a staff person. We can count it. But it’s meaningless.
I’ve touched on purposeful data so I won’t belabor the point. You need context around the activity itself. In the case of footfall or information access, you may need context around when or where the activity occurs. We’ve been closed for 2 and a half months. All of our normal usage data will be skewed if we use 2020 in our analysis in the future.
The Cost of Purpose
The accessibility initiative is a good and necessary one. Since the organization uses a CMS that will, in theory, create accessible templates, our focus is just on content that isn’t templatized. I’m talking mostly about PDFs.
Fortunately, our library only has one PDF on our site. Our archives department has hundreds. In the end, we’ve decided that there is no advantage to retaining the documents as PDFs. But we’re not eliminating them. We are converting them to HTML so that they can be placed within accessible templates.
I realize that not all archival material can be made accessible. These are finding tools and are easy to convert. Also, I generally oppose using PDF for anything that can be done in HTML.
The corporate alternative was just to delete anything that hadn’t been visited recently. It’s an incredibly short-sighted view of information. Even outside of the library and archive world, sometimes you need to keep things that haven’t been used recently.
There is a cost to making information accessible. A book has obvious acquisition and maintenance costs. A web page requires a server, perhaps a domain name, electricity. It’s a shifting scale. We do not tend to measure how much a web document costs to access.
There is a cost. If the document must be maintained and updated, you’ve got staff time. If not, there’s a nominal cost to maintain it. This blog post has a cost to create (my time, the cost of the device I’m creating it on). But there’s no additional cost to maintain it without change, as an archive, beyond the cost I’m paying for the other content I’ve posted.
We shouldn’t treat web pages the same as if they have the same costs and physical properties as books. We should consider the specific context of physical and digital items when we are considering adding or weeding them.
More importantly, bare usage isn’t sufficient to determine its usefulness. Ironically, we have many tools to understand online traffic flow. It is much harder to do in physical collections. We saw a change when we shifted a collection because the books were off the shelf more. On a web site, we can see how people move around.
There are many reasons a page may not be visited on a web site. It may not be findable through search. The navigation may hinder its browsability. It may not be intended as a page that is heavily used. As with our archival finding aids, it may have a role as part of a collection, some of which is used more heavily than others.
Operationally, the accessibility project is no big deal. We have a solution that will enable us to keep our content and make it accessible. But it’s a reminder that what we count may only reflect what we can count, not what we need to count.