Since we've been running with virtual browse for a very long time now, I
thought I'd give a shot at a best answer to some of these questions....
> Where is the feature demand originating? Staff? Faculty? Students? Grad
> students? Undergrad students? (Not to exclude publics or special
> libraries, but this seems to be an academic catalog feature, when it shows
There is a twofold demand for a feature like this. The first is that with
the move to remote / inaccessible storage, people want the ability to be
able to browse the books. We have almost half our collection in an
automated storage retrieval system at our second main library. Users don't
necessarily say they want to "browse the shelf" but there is a sense of
need to explore the collection in a more tactile way than is available via
standard catalog search.
The second is a pretty well-expressed demand to find related titles. Again,
not necessarily expressed as browsing the shelf, but more of a desire to
get interesting titles suggested, just like Amazon and Netflix and about
every other content service you can think of already does.
Call number is an excellent way to make recommendations of possibly related
titles. Shelf browse lets user explore and cherrypick the collection in a
way that transcends explicit search terms.
Also, I will say that we implement TONS of services that are useful to
patrons that they never explicitly ask for. If we waited for patrons to ask
for things specifically before we tried them, we would not be very
> What is the level of familiarity with library/library services/library
> systems for those that request this feature?
They don't really request it, so I don't think this question matters much.
We do deal with faculty and grad students who are genuinely concerned about
loss of shelf browsability (esp. those in the humanities). Those folks do
understand library collections and call numbers.
> Is implementing shelf browse an attempt to work around some other catalog
> deficiency (e.g. weak subject cataloging)?
No, absolutely not. It's an attempt to present alternative forms of
exploration and browsing beyond keyword searching. I should note that we
have tested and are currently working on a feature that will be similarly
presented that suggests related titles based on subject heading terms. Our
plan is to show them both as 2 suggestion widgets on the full record for a
title in the catalog -- if Amazon can recommend related items via 3 or 4
mechanisms, then why can't we?
> Does the corpus have the cataloging data to support such a feature? (A lot
> of ebook packages do not have call numbers, for example.) What¹s the
> percentage? Is that reasonable?
This is a *great* question. Ebook packages are a challenge. When we first
looked into this question back in 2012, we saw about 60% of our ebooks
didn't have call numbers. However, that number doesn't mean as much as it
might, since about 90% of the records lacking call numbers were from 2 very
specific ebook packages -- Early American Books and Early English Books
Online, which are not evenly distributed throughout our collection. We do
get call numbers in our YBP ebook records. We are currently working with
automated tools to bring call numbers into our 360 MARC record loads using
OCLC Classify. Without enriching those records, we were seeing maybe 20%
that had classification. I think initial testing showed that we might be
able to jump that up to at least 60% using Classify, but we are still in
the testing phase.
> How do you plan on tracking use of the feature? What would you consider to
> be a success rate? 20% of sessions? 5%? 1%?
> At what point do you sunset the feature? Expand upon it?
> How long will the feature take to implement? How many staff will be
> involved? What is the ROI?
All difficult questions. I'm not sure I could answer them for many of our
services. We would probably only sunset the feature if it started costing
too much staff time to maintain. We first implementing virtual browse as a
replacement for the call number list browse that was the last step to
getting rid of our old OPAC. So we had justification beyond just thinking
virtual browse was cool.
> Will all of your users understand the visual implementation on the page?
> How do you plan on testing it?
> We have tested it, and no they don't understand the visual implementation
(at least in our cover view implementation). They also don't really care.
All they care about is whether or not it's showing them books they're
interested in. The biggest issue in usability testing of our service is
that they have trouble finding it on the full record in the catalog. (the
reason we are going to incorporate the actual browse into the full record
Associate Department Head
Information Technology, NCSU Libraries
[log in to unmask]