9.09.2010 Prepping for meeting with reference

Burying the fed search interface doesn't help people use it. I feel like it should be almost invisible, presented in context with your other library database links. I dislike fancy names for things (Vera?) - call it something that makes sense, like Multisearch.

I really like how RIT says it's a "handful" of most popular databases. what you're searching:
http://library.rit.edu/databases/
and makes it easy to go to subject-specific databases.

But they aren't integrated into the subject guide pages: http://library.rit.edu/guides/chemistry/chemistry-including-biochemistry.html

I think integration is crucial - a common model for access across all representations of your resources. But how difficult is it actually to do these things?

But, MIT's Vera does give several "quicksets" for different users - this could be useful for our diverse users. (diff than something like a liberal arts college, where one set of databases might doa pretty good job for most first searches) (I wish MIT's would let you know up-front what's in the different sets, though. But maybe it's hard to do this without cluttering up the interface.)

Metalib is confusing in the native interface, I think. Xerxes makes it a lot better. Not sure how hard it is to set up.

Some libraries are going toward the mega-aggregate indexes (Primo Central (BYU), Serials Solutions Summon), but I'm not seeing that in our market basket institutions and I'm not sure how they integrate it with their other stuff.

I like Georgia Tech's second screen (crossearch) - would fit nicely into our existing subject guides.

From what is available to us as guest users, we usually can't see what the results pages look like or what options you have for dealing with results, or how easy it is to link to the source database and possibly rerun your search there.

Would be nice to see numbers on some of these things - Google Analytics baseline? Compare with changes?

========

Seems like the people w/ really good, intuitive interfaces are on their 3rd (or more!) iterations of fed search / mult database search / single box search options - RIT, BYU, Oregon State, Cal State? Check on how many versions they've had. Maybe gaining experience IS worth something in and of itself.

Xerxes http://arch.library.nyu.edu/
-front end for metalib?


Look at new LexisNexis interface - how they're addressing the problem of providing access to resources of many different types (they have separate entry points for each, but show them all - we don't have that luxury since we have even more types of resources.)


6.25.2010

Did a questionnaire after a Primo Central webinar - some of the things they asked about, we might want to look into, too. Note terminology "mega-aggregate service" for things like Summon, Primo Central, EBSCO Discovery Service.

If you are do not presently use the Primo discovery interface, what discovery interface do you use?
Aquabrowser
Blacklight
EBSCO Discovery Service
Encore
Endeca
Enterprise
eXtensible Catalog
VUFind
WorldCat Local
Other
Don’t know
I'd rather not say


Are you presently subscribing to any mega-aggregate service?
Yes, EBSCO Discovery Service
Yes, Primo Central
Yes, Summon
Other
Don't know
I'd rather not say
None at this time


=============
Terminology problem - what is "federated search"? What is "federated search" vs. "discovery layer"?
Possibly this?
  • federated search = when you make a query, it's sent out to all the various databases/resourcces, and then the results are sent back to you
    • Slow to return results, problems w/ how results are returned (order, deduping)
    • # of resources searched limited by time, computing power? speed tradeoff
  • discovery services = when you make a query, it returns results from a local combined index that was ALREADY compiled from various databases/resource.
    • Faster, more control over how results are returned (order, deduping)
    • But a lot of publishers (db vendors) are not going to want to give/sell their metadata to be included in a local index.
      • ebsco seems to have had some success in talking some companies into letting them index their data in their Discovery Service product, but companies w/ competing products are not going to let their databases be indexed this way by other companies (for example, ProQuest/SerialsSolutions has a competing product w/ EBSCO, so we wouldn't expect to see ProQuest databases to be available in the EBSCO discovery service interface)
  • Seems to be some blurring between these. Discovery services seem to often include faceted browsing.

Have to think as a library about what we'd want to USE this for? Of course we and our users want to be able to search everything at once, but you can't realistically search everything.
  • knee-jerk reaction: one box to search them all, YAY! I don't have to think about where to search or run my search a bunch of times. But
  • - discovery of which databases are better for your search? Use federated search to identify a good database, then continue your search in the native database?
    • how easy is it in fed search results to tell what the source database is and then get to that resource? (will vary between products)
  • Collections of general resources - good for undergrads / broad interdisciplinary use?
    • Grad students/faculty are more likely to search in the specific individual databases they know are useful for their field?
  • Collections of subject-related resources?
    • like a collection of databases that are good for biology, as in what's included in the research pages.


What is the typical workflow / thought process for someone using a federated search product? Do they understand it? Is it an improvement over searching dbs individually? How do students perceive things? I'd imagine how it's presented in the context of the library website and the actual design of the interface has a big effect on how users perceive/understand (or not) federated search

Google scholar? Comparative coverage?

Functionality questions?
  • How easy is it to see in different products what you're searching, and what is and what isn't available through the federated search?
    • BYU used to have a REALLY nice interface for this, butI think they've changed it. Le sigh.
  • Do different products make it possible for
    • librarians to choose different groups of databases to make searchable
    • USERS to combine what databases they want to search, on the fly?
      • BYU had checkboxes, not sure if they do anymore.

- how are results returned in the interface?
- options for narrowing?
- what kinds of things can be searched together? Dbs? which ones? catalog? Digital collections? what about our archives databases?

Presentation
  • Explain diff btwn discovery interfaces, fed search, metasearch
  • checkboxes of diff products.
  • Show some examples

Notes from Webinar:

The Success of Web-Scale Discovery in Bringing Net-Gen Users Back to the Library: The Summon Service in Academic Libraries (April 8, 2010)

Doug Way (Collection Development, Grand Valley State U)
Jennifer Duvernay (Marketing/Outreach, Arizona State U)
Moderator: John Law, Serials Solutions

Slide presentation is downloadable, presentation will be archived in ~4 days.

Intro (John Law)

Introducing Summon
  • Existing options for accessing library content
    • Library catalog
      • Physical Collection
      • Electronic resources (not in there at article level)
    • e-resources webpage / databases page
      • lists databases - hundreds of options - daunting to figure out where to begin
    • Compensatory behavior (when things are too hard, people bypass the library)
      • wikipedia
      • search engines
      • other open-web resources
    • Libraries have an edge - faculty (?) think of library as preferred source for quality, credible content for academic research
    • But, libraries are not the easiest way to start the research process
  • Summon provides single search box - more like google
    • unbiased content - no preference to any content provider
    • Build from scratch
    • pre-harvest the content
    • NO federated search
    • Available w/out authentication barriers
  • New era for discovery - distinction between these.
    • Federated search
      • affordable technology
      • query sent out to many databases, results returned.
    • Discovery layer
      • next generation catalog
    • Web-scale discovery (term coined by Breeding)
      • single index of all of the library's holdings (pre-indexed)
  • Simplified illustration of the summon unified index
    • harvested in advance
      • library catalog
      • e-journal articles
      • databases
      • newspaper articles
      • e-books
      • dissertations
      • institutional repository
      • etc.
    • works with entitlements info (does it work w/ SFX data?)
  • Published report: academic libraries and the struggle to remain relevant... (available from serials solutions site?)

Doug Way

Exploring the use of Summon at GVSU

  • launched summon late august w/ single search box on homepage
    • users had option to use classic catalog originally, but it's now their only search box
  • was it being used? no user stats for summon - looked at stats from other places
    • link resolver stats (serials solutions 360)
    • COUNTER stats
      • searches
    • google analytics
    • compared to prev. years, resources saw increased use overall.
  • database usage stats
    • outliers:
      • down for CSA, Wilson, OCLC. up for EBSCO? E
        • excluding ASP, CINAHL, then EBSCO was up even more?
      • asp was up 11%, 37%
      • Econlit, eric was down.
    • overall usage was down significantly for a lot of databases: 50%?
  • link resolver clickthroughs
    • all sources were down in spring semester (0-30%%), but then up 65, 92, 179% in following semester.
    • aggregator databases up a LOT.
    • journal packages down first semester, up second semester
    • individual journal titles
      • down jan-aug ~15%, up way more sept-dec (50-80%)
  • conclusion: drop in database searches & increase in full text downloads = summon having a major impact on use of library resources (circumstantial, though impressive)
  • Observations:
    • users like newspaper content - saw 500% increase in many newspaper databases in link resolver clickthroughs
      • People are getting articles from major newspapers
    • users may still get confused - book review titles went way up in accesses - do people know they're getting book reviews?
      • ref questions don't reflect demand for book reviews
      • book reviews can look enticing - like an article.
    • users like book % referrals to encore (catalog) from summon went up
  • conclusions
    • users were able to connect to content even when provider was not a summon parther (asp, elsevier, jstor) what is a summon partner?
    • we need to identify how to connect users to Summon and to core subject databses
      • make core dbs and summon stand out and stand apart
      • make it seamless

Jennifer Duvernay

Promoting Library One Search at Arizona State University

Research 1 university ~65,000 students at 4 campuses, online

Signed Aug 1 2009, beta rollout nov, full rollout January 2010
  • Students, admin, faculty asked for it (reduction of barriers between people, content). We got it for them. Now what?
  • Process (started 2 wks after signed - equal part of implementation plan)
    • Naming
    • Visual identity development
    • System Integration (in library, university)
    • Promotional Messaging
  • Naming something unique to ASU - local brand
    • Unique (Suma or Summon) vs. explanatory (Library One Search)
      • Unique names represent an ADDITIONAL barrier to users.
    • talked to students about what they thought of name.
      • "library one search" makes sense. Tells you everything you need."
      • "everything in one place - I get it"
      • "it needs to say what it is"
  • Visual identity development
    • identified criteria - logo and full version, local to environment, professional (willingness to hire graphic designer)
    • committee member was able to use university branding kit to come up with some nice-looking options.
      • went with the simple one - didn't want guy with a book to make them think that it was only books
      • so now, they have text logo, icon
  • System Integration: place your resources where users are going to be able to find them with the least amount of effort
    • it's prominent - front & center on library website
      • like GVSU, it's the ONLY search box (other than site search in header) that's available on library homepage
        • putting money where mouth is.
    • MyASU portal - when student clicks on library, onesearch box pops up (also has links to other stuff)
    • LibGuides: search widget to integrate into subject libguides
    • ASU libraries browser toolbar (how did they do this?)
    • Course management system - planning on this, pending on upgrade - will be integrated along with other library resources/links
    • Simple url: lib.asu.edu/one
  • Promotional messaging (what you'd think of as marketing)
    • being clever gets in the way of effective messaging
      • focus on clear and simple message
      • "finally a serious research engine"
      • "research made easy"
      • "researching has never been this easy"
    • Authentication - promotional messaging
    • myasu banner ads
    • digital signage
    • vinyl banners - one for each campus, 9 feet tall, placeed at entrance (the only marketing thing they spent money one)
    • other announcements
      • asu lib website
      • student myasu announcements
      • focused on ease of use
  • havent' had a whole semester yet, but people like it.


Q&A

  • GVSU: students only have to authenticate once they're trying to access
  • t does work with e-book platforms. e-b
  • ASU: good feedback, planning on assessment now.
    • in the future: spotlight video to highlight features - 2-3 min (I'm not that interested in this)
  • ASU is working on building an inst. rep.
  • GVSU has a contentdm and digital commons repositories - they are in summon.
  • How does this differ from fed. search:
    • searching resources individually - at the same time - relying on speed of things to come back, then it takes time to sort, dedupe, rank
    • summon: preindexed, so results come back faster, better relevancy ranking because data has been normalized
  • did circ of books increase? haven't looked at that yet.
    • usage of catalog has stayed consistent
  • Any feedback from fac on quality of research?
    • ASU: not yet (still early)
  • Doug:
    • why are you using both summon and encore?
      • encore serves the catalog/opac function
      • if they need a book, encore is still the easiest way to do it.
  • what kind of usability testing has been done?
    • lots, as part of development, and ongoing.
    • user studies being conducted
      • dartmouth - 2 studies (undergrads, grads) - published on dartmouth college website
      • manitoba - upcoming conference


Upcoming LJ webinar: thur, may 6, 2010, 12-1. Understanding the new discovery landscape: federated search, web-scale discovery, next-generation catalog, and the rest.