The Impact of Deep Understanding, Indi Young, keynote, #InternetLibrarian

Presentation available at

Has a volume of letters by a relative who toured Europe in 1937 and documented everything in letters, photos, and movies.

Now, Facebook preserves our memories by date.

How will we make things for “digital natives” and their grandchildren?  Through “deep understanding.”

Matrix with quantitative/qualitative on one axis,  evaluative/generative on another axis.  Third axis: problem space/solution space.  Understand the person through “cognitive empathy.”

Solution-framed and purpose (problem)-framed.  Measure quality: does the solution work?  Is the purpose supported?

Daniel Goleman says cognitive empathy “leads to hustling.”

Support intents and purpose — more than tasks or goals.

Publisher has a solution: how do people see our brand? How are we doing?

Customers have purposes: how do I improve my skills?

We think, “If I were in their shoes, I would …”

Empathy is not you being in their shoes, but their being in their shoes.

Emotional empathy: understanding another’s feeling.  In work: to support another person through an emotional process.

Cognitive empathy: understanding another person’s inner voice, thinking when they are trying to achieve an intent or purpose.  It transcends time.  In work: supporting another person in achieving their purpose.

Walking in their shoes: act their character (not the way you would do it).  Need to listen to develop empathy.  Fast-moving orgs often skip that part.  They guess, they make it up.

Listening sessions, not “interviews.”  Like following a tour guide, you listen to what they want to talk about.  You don’t talk about the other building down the street.

Depth: reasoning, reactions, guiding principles, decision-making, work-arounds, source of reasoning, guiding principles.

Find out why, like a two-year-old, keep asking “Why?”  Not embarrassed about what you don’t know.

Start with scope: What went through your mind as you were trying to accomplish your purpose?  Then follow up from there.

Use an event: What about the last time you did that?  What about the first time?

Avoid judging, contempt.  Her pain is real, even if you don’t agree.  Don’t put yourself in their situation, as yourself.

Rapport: support the participant emotionally.

Active listening: turn off the noise in your head, no notes, no analysis, no problem-solving.

Attempt to banish assumptions.

Often businesses look at stats: gender, age, income, etc.  Not necessarily a correlation with anything.

Divorce rate in Maine correlates with consumption of margarine.  (Spurious Correlations web site)

Ask people on an airline why they check their bags: all kinds of reasons, nothing to do with demographic categories.

A user is a person with a relationship to your organization.

College segmented people by age and grade point, etc.  More useful categories: passionate about the subject, looking forward to college, a means to an end, exploring paths.  Purpose-based, behavioral segments, not “personas.”

Health advice for people who want to lose weight.  Different categories: the resigned, the sidetracked, the inconsistent.

University web users: matchseekers, pulse takers, active supporters, prideful belongers.

Public library users: confused by jargon, people sitting outside using the wi-fi (put in a bench).

Where does the funding come from?  “Budgeting for customer experience vs. advertising.”  A company had $25 million to drive people to web site and $20,000 to fix the experience once they get there.

Deep understanding is not something to squeeze into the production cycle.  It happens on a separate cycle: maybe once a year or once every several years.  The data lasts, it doesn’t go stale.

It’s not about how to build a service, it’s about what service to build.

One on one listening


Mental models (don’t use sticky notes)

Untangle common concepts

Grouping patterns, start to label things

Another blog post about this:

Updated to add links.


Inspiration architecture: the future of libraries #internetlibrarian #il2015 @morville

Peter Morville, Internet architecture pioneer


Told Library of Congress their web site was like the Winchester Mystery House. We need to grapple with culture and governance to make sustainable change.

Created portal and discovery tool for Harvard’s 75 libraries. Faceted results to filter results on the left. Helps people understand what they’ve found. But it’s not enough.

At the beginning, more features add more goodness, but at some point, it just gets more complex (and less good).

Agile is good, but planning is important too.

Planning, building, thinking, doing all part of the process.

We need to look outside our models to see external forces.

Categories can be dangerous but not bad. Cornerstone of cognition and culture. Don’t have to be a bounded set, but could be a fuzzy set or a centered set.

We use radio buttons when checkboxes or sliders would reveal the truth.

Info architects create paths to places, connections between categories, links between cause and effect.

Practice ethnography not just with users, but also stakeholders.

Learn about organization’s history.

People may change actions, but are reluctant to change beliefs.

Have we passed an inflection point? Are we making better decisions with more information?

We are living past our limits and beginning to feel the consequences.

River daylighting. (Yay, a water metaphor.) Making the invisible visible.

Redesigning Harvard business library’s 12-year-old web site. Did ethnographic study of research habits.

MBA student couldn’t do many tasks. She’s used the ref desk, but it’s hard because it’s a quiet room and she feels she’s distracting other users.

Homework: pick a context, map the system, then map the system outside the system. Then share it with someone.

Info architects use nodes and links to create environments for understanding.

Library is an act of inspiration architecture and a keystone of culture.

Learning from Medical Libraries #internetlibrarian #il2014

Renee de Gannes-Marshall, Canadian Medical Association

Web site for members:

Complete overhaul, new CMS, review of 30,000 pages. 46 people over more than a year. 400 members assisted.

Clinical tools are the top reason members use the site, but no way to search across resources.

Evaluated discovery services and went with Ebsco Discovery Service (EDS). Used API to build it directly into the site. More customization of search and more usage metrics, but it’s a whole lot more work.

Testing by librarians, then moderated usability testing with users. Hired a firm to do the user tests. 12 1-hour remote tests via WebEx.

Filters with library jargon like “document type” and “database” didn’t help users. Publication type is more meaningful to people. Hid some of the filters that only librarian use under “more filters.”

Final testing. Needed to do some additional tweaks after everything was moved over. Also, some adjustments for mobile access.

Video about new features and tutorial. ( is available to the public.)

Elisabeth Marrapodi, Trinitas Regional Medical Center (N.J.)

Games and Second Life to improve health literacy.

Nalini Mahajan, Marianjoy Rehabilitation Hospital (Ill.)

People looking online for medical

Their library has done (or will do soon) web sites for persons with disabilities (and their potential employers), for parents of children with disabilities, and for those in rehabilitation. Links to Medline and other sources. Form to ask for more information.

Edited to add links and fix typos.

Unifying UX: Consistency with Content #internetlibrarian #il2014

Jordan Fields, Garfield County Libraries (Colo.)
Mark Noble, Marmot Library Network

III Sierra Opac
Vufind, open source discovery layer
Overdrive e-books

Patron tries to find e-books. Catalog is difficult to use. There’s a long help document. She goes to Amazon.

Asked users to do tasks and watched them. Also asked little questions, like what should this label say? Offered one “book buck” (good for paying fines, etc.)

It doesn’t make sense to users that catalog looks different from the larger web site. They tried to make it seemless. “Book scroller” (row of cover images), click on it and go straight into catalog.

Simplification: one result per title, not lots of different editions that confuse patrons. But you can still click on “show editions” to see all editions. Sorted so local copies show first, then available copies elsewhere.

Enrichment through NoveList: series, covers.

“More like this” based on metadata in catalog. Works with movies, music, etc. as well as books. Staff loves this stuff, instant reader’s advisory.

Web site is on Drupal. Navigation links across top is the same in both catalog and web site.

Before integration, it looked like four different web sites:
* desktop web site
* desktop catalog
* mobile web site
* mobile catalog

Now, they all look pretty much the same.

Gave staff access a month ahead of public and training. Staff felt listened to.

Took a few months to figure out best way to group records. If there’s a problem, a librarian can report it and library staff — rather than a programmer — can fix it.

Library is not part of county government, so their web site doesn’t have to look like the county’s. Some of the libraries in their consortium have to be consistent with a parent organization, but that can be done with iframes and CSS.

Other databases (Ebsco, Zinnio) allow a certain amount of customization by adding the MARC records. They have told vendors what they want from their APIs.

Edited to make things clearer.

Library mashups: what’s new?

Nicole Engard, ByWater Solutions, author

Presentation slides

Mashup mixes two or more functions in one.

Web service: a technology that allow information exchange between two applications

API: A set of functions, etc. that allow web services to work.

Get APIs from vendors, create RSS feeds for your content, sign up your library on social networking sites, use Yahoo Pipes and IFTTT.

If you’re depending on another site, remember that the service (or the site itself) may go away. Also, you may have speed issues.

Always read terms of service, which dictates how the API can be used.

Yahoo Pipes.

If this then that (IFTTT)

* Works with all kinds of web/mobile apps
* Works with WEMO, which controls home appliances
* Library example: If a new event is added to my Google Calendar, post it on Facebook.

ScribbleMaps: lots of icons to put on maps

OpenRefine: Cleans up messy data. For example, you could upload library catalog data and remove library-specific punctuation.

Google Fusion Tables: Load tabular data and change it to something else, such as a map.

Serendip-o-matic: Takes your content and compares it to sources such as Digital Public Library of America (DPLA), Europeana, and Flickr Commons to find similar items.

Pasco Library (Fla.) generates a pie chart of circulation.

Stanford Univ. has a mashed up catalog, using Blacklight.

Los Gatos Public Library (Calif.) pulls e-books from Overdrive.

BookMeUp (Montana State Univ.) uses Amazon, Open Library, WorldCat, etc. to make recommendations.

Houston Public Library has timelines generated from their collection data.

Career Log: Use IFTTT to grab calendar events, LinkedIn updates, SlideShare and put it all in a Google spreadsheet.

Tweet the weather.

Use Yahoo Pipes to mash together library feeds.

Regime uses several sources to analyze members of Congress.

Wellcome Library has a map of AIDS posters worldwide.

Edited to add links.

Polishing up your web site #internetlibrarian #il2014

Sonya Betz, Robyn Hall, MacEwan U., Canada

Microinteractions, macro results

Institutional repositories (have been called a”roach motels”): wanted to make a site faculty would be proud to use.

User testing = observation. Not about asking for opinions. Lets you observe problems.

* Representative users
* Representative tasks
* Observe what they do and where they have problems
* Revise and re-test

Microinteraction: a tiny piece of functionality.

Feedback to user: labels, error messages, etc.

Tested faculty on uploading documents to the repository.

Wanted to click on big “share” at bottom, so made that a link.

Nobody read “about” page or FAQ.

Nobody clicks on “Let us help,” because it sounds burdensome. Changed to “Let us do it.”

“Citation information” caused some confusion. They wondered what style to user, so changed to “Publication information.”

People wanted link to contribute another work. Also wanted link on their own names (“the academic ego”)

Getting ideas from ResearchGate and Dropbox.

Can make the difference between products we love and those we tolerate.

Weeding a website
Tabatha Farney, U. of Colorado

De-selecting “zombie content.”

New content management system led to more people creating more content. Library web site up to 300 pages.

Politics: people get nervous about weeding.

1. Identify low use pages with web analytics.
2. Get user feedback on those pages.
3. Decide what to weed.

Also looked at “bounce rate” (short time on page, no click on anything), last time it was viewed, last modified date.

Tried to put comment form of pages flagged for weeding. Flaw: already low use.

Met with content creators individually to understand why these pages exist.

Reasons like: it’s always been there, it makes the administrators happy, because some user may need it someday.

Consolidate pages like policies, staff directories, how-to info.

Working on a content strategy.

Marketing to web site users.

Archiving unused content. (Some are uncomfortable about deleting pages just yet.)