Hi Ed,
Sorry for the *very* delayed reply to your thoughtful message. I've been
traveling a lot this month.
Firstly, thank you for keeping us honest regarding the possible
alternate explanation for the short sessions. It had occurred to me that
this could just as easily point to failures and frustrations as to
successes, and regrettably neglected to include that in my initial posting.
Second, as to the possibility of different software having variable
session parameters, I can see your point there as well. However, both
the 2008 and 2009 numbers represented Primo (we went live in late August
2008), and the statistics themselves come out of Urchin (predecessor to
and locally hosted version of Google Analytics) rather than from Primo.
The comparable session-count numbers for our previous GEAC environment
(also via Urchin) also makes me think that, after a little over a year
of Primo, the shifts can be attributed to the "next-gen" tool with some
confidence. Unfortunately, we were not tracking session-length under GEAC.
Incidentally, it's the relationship of the 2008 to 2009 session length
numbers that lead me to suspect the reason is success rather than
failure: after a year of fine-tuning, improving, doing usability
testing, and running a user-education campaign, the session length seems
to have gone down significantly. The initial 2008 numbers were for the
first month of a new system, which makes some sense anecdotally. That
said, the search-fail hypothesis you suggest is also a distinct possibility.
Thanks again for your feedback,
-Corey
Edward M. Corrado wrote:
> Corey,
>
> Thanks for sharing your numbers but I wonder what more sessions mean. If
> you multiply the average # of sessions by the average length, you get
> almost the same number of minutes the catalog is being used. Different
> software has different time outs, session parameters.It is comparing
> apples and oranges in many respects. Since the total amount of time is
> similar, I would be interesting to see a more in depth analysis as to
> why the total time is close but session lengths are different.
>
> I'd also be interested in any evidence that supports "decreased length
> of session implies that they are finding what they need more quickly
> than they did when we first launched the new tool" --- couldn't it just
> as easily imply searchers are just giving up earlier and not finding
> what you need. In previous(?)-generation systems when I looked at search
> logs, often I found short sessions where failures, not signs of success.
> For example, some would search for a title like "The Truth About Cats
> and Dogs" and because of the initial article wouldn't get any results,
> and they wouldn't refine their search (or if they did they would
> narrow(!) it). The searcher was unable to find the movie even though the
> library had multiple copies.
>
> I guess what I am saying is while I really think these next-generation
> OPACs are useful and should provide better interfaces and make it
> simpler to find things, just looking at raw numbers only goes so far. I
> think someone earlier in this thread asked about circulation stats. Do
> you (or anyone else) have any evidence that show a change in circulation
> patterns since implementing a next generation system?
>
> Edward
>
>
>
> Corey Harper wrote:
>> Dear All,
>>
>> Sorry for the delayed foray into this thread.
>>
>> NYU is finding something similar to NLA's experience. We implemented
>> ExLibris' Primo in September 2008, and spent much of the following
>> year fine-tuning (and debugging) the service and providing instruction
>> in its use.
>>
>> Below are some statistics comparing September 2008 and 2009, the month
>> we rolled out a "next-gen" tool, and the first month of the new year.
>> The session information is very intriguing.
>>
>>> Average Number of Sessions Per Day
>>> Sept 2008 4000
>>> Sept 2009 10200
>>>
>>> Average Length of Session
>>> Sept 2008 15 minutes
>>> Sept 2009 6 minutes
>>
>> We didn't track the length of session before 2008, but the following
>> numbers for previous Septembers show that 2009's usage patterns differ
>> greatly from those of our old OPAC:
>>
>>> Average Number of Sessions Per Day
>>> Sept 2006 4500
>>> Sept 2007 3333
>>
>> Clearly, our system is getting *much* more use this semester than our
>> historic baselines for the start of the term, and the decreased length
>> of session implies that they are finding what they need more quickly
>> than they did when we first launched the new tool. It would be very
>> interesting to see that data for our old OPAC, though I don't believe
>> we collected it at that time.
>>
>> Best,
>> -Corey
>>
>>
>> Warwick Cathro wrote:
>>> The National Library of Australia implemented VuFind in May 2008 to
>>> replace our previous (Voyager) OPAC:
>>> http://catalogue.nla.gov.au/
>>> The usage of our catalogue in the year July 2008 to June 2009 was
>>> 140% higher than in the previous year, with 20.9M catalogue searches
>>> compared to 8.7M in the previous year.
>>>
>>> We are hoping for a similar response to the renovation of our
>>> national collection discovery service (which combines eight existing
>>> services) when we release it in November 2009:
>>> http://sbdsproto.nla.gov.au/
>>>
>>> Warwick Cathro
>>> National Library of Australia
>>>
>>> -----Original Message-----
>>> From: Next generation catalogs for libraries
>>> [mailto:NGC4LIB_at_LISTSERV.ND.EDU] On Behalf Of Kyle Banerjee
>>> Sent: Friday, 25 September 2009 7:49 AM
>>> To: NGC4LIB_at_LISTSERV.ND.EDU
>>> Subject: Re: [NGC4LIB] Stats showing impact of Next Gen Catalog ?
>>>
>>>> I seem to remember that some libraries using worldcat local reported
>>>> an increase in interlibrary loan activity. There may an argument here
>>>> that if there were a more transparent way across all libraries to see
>>>> holdings easily/centrally from within the catalog that circulation
>>>> would increase, a rising tide lifting all boats, or at least the
>>>> bigger boats with more holdings.
>>>
>>> Since implementing WorldCat Navigator -- a consortia level product
>>> that provides a discovery experience similiar to WorldCat Local, we
>>> have seen an increase in ILL across our member institutions. However,
>>> the impact is highly variable ranging from very little to over 200%.
>>> Some institutions also report increased use of electronic resources
>>> which could conceivably result from requests for nonreturnables that
>>> would normally go to ILL being routed through resolvers that identify
>>> electronic copy. Actual circ numbers are down.
>>>
>>> A number of logical explanations could be used to describe what is
>>> happening, but the reality is that associating cause and effect can
>>> be difficult.
>>>
>>> One thing that is often absent from discussion is the effect that
>>> service expectations have an impact on requesting. You can have the
>>> greatest discovery mechanism, but if you want reference libraries to
>>> plug your service and patrons to rely on your system, the delivery
>>> mechanism whether materials are physical or electronic needs to perform.
>>>
>>> kyle
>>>
>>> --
>>> ----------------------------------------------------------
>>> Kyle Banerjee
>>> Digital Services Program Manager
>>> Orbis Cascade Alliance
>>> banerjek_at_uoregon.edu / 503.999.9787
>>
--
Corey A Harper
Metadata Services Librarian
New York University Libraries
20 Cooper Square, 3rd Floor
New York, NY 10003-7112
212.998.2479
corey.harper_at_nyu.edu
Received on Tue Nov 03 2009 - 17:48:52 EST