1 of 36

More Than a Map to Get the Treasure

Strategies for Easing Information Discovery in Digital Collections

🙢

In all great treasure hunting stories, our hero usually starts the adventure by obtaining a map that leads to where the coveted lost relic or magic book of antiquity lies buried beneath a desert, church, or even the streets of Venice. Almost with out exception, our heroes also need something more in order to find what they are looking for. ��When it comes to college students using digital collections, simply putting a collection online is not enough to generate usage. While “providing a map” to our digital collections by linking to the content management system can catch some, it really takes more to properly highlight the gems stored in your digital treasury. Easy information discovery is a crucial goal for libraries. ��This presentation will address three aspects of information discovery for digital collections as it relates to making searching easier for the user- Search Engine Optimization practices, Supplemental metadata schemes and library usability testing.

2 of 36

More Than a Map to Get the Treasure

Search Engine Optimization

🙢

3 of 36

The first and most important rule of SEO is to ensure that your content can be indexed by search engine crawlers

🙢

To understand SEO strategies, it helps to understand how search engines actually work. When a user performs a search in Google (or Yahoo, or Bing!) he/she is not actually searching the entire Internet. Instead, Google uses those terms to compare against its index, which represents only a fraction of the whole Internet. Google, and other search engines, build their indexes through the use of programs called “crawlers” or “bots” or “spiders.” These programs are automated to search through websites, gathering information that can then be added to the search engine index. These programs look for websites on their own, but most search engines also allow web designers to specifically request to be added to the index (which is done by the crawler).

Therefore, for SEO purposes, it is absolutely essential for web designers to structure their content in a way that allows crawlers to index it properly. Crawlers greatly prefer fixed and simply URLs for web pages. Extraneous elements, like session tokens, or variables in the URL, may actually prevent a crawler from making sense of a web page—resulting in a partial indexing, or a completely failed indexing.

4 of 36

Other important SEO techniques

Sitemap (Sitemap protocol)
Optimized <title> and <meta> tags
Clean HTML coding
Clean site structure
Quality content reflecting optimized language
Longevity

🙢

Good SEO should be seen as part of good web design. Most strategies are fairly common sense—make sure web page titles reflect optimized key words and phrases, make sure the coding is clean, and make sure the body of content is well written. It helps to consider websites as almost an advertisement for the content they house; make sure to hit the key points early and often. Most of all, it behooves web designers to create content with the end user in mind. This is especially true for libraries, as too often websites are clearly designed for librarians, not patrons.

There are a number of suggestions that search engines themselves make, as best practices, which should be followed if at all possible. One such suggestion is the use of the Sitemap protocol. This is slightly different than the sometimes provided user sitemap, in that it is actually a map for search engine crawlers. The sitemap protocol allows a website to give a crawler instructions on just how it wants its various pages to be indexed. Search engines also greatly discourage the use of “Black Hat” SEO techniques, practices considered unethical or illegal. Google has, in the past, outright banned websites from its index when caught using such practices.

There is a large SEO community where librarians and information professionals can find more detail about certain strategies and practices. The focus of this presentation is not so much on the strategies, as to why libraries need a raised awareness of SEO.

5 of 36

73%

Search engines return millions of results
The vast majority of users never look beyond the first 10

🙢

6 of 36

Search engine use

4/5 websites are accessed through search engines
Search engines are described as simple, speedy, and convenient
They are becoming less reliable, not more

🙢

However, the same studies also show that search engine users are using them very frequently, often multiple times daily. It is only natural that users have become very comfortable using search engines to answer questions—and their daily success, often for several years, begets a continuing demand for speed and convenience when searching for answers to questions.��Despite their success, search engines are actually becoming less reliable, not more so. Google has said they use hundreds of factors in determining page rank, but beyond that they are quite secretive (as are Yahoo and Bing!). Furthermore, these factors, and their relative weights in search engine algorithms, are constantly changing. This means that SEO efforts are aiming at a moving target.

Recently, search engines like Google have begun personalizing search results—meaning they take into account a user’s location, browser history, and other potential biases in order to filter results with that user in mind. Users are becoming accustomed to finding information in this way—despite the fact that search engines are increasingly hiding information from them.

7 of 36

The unintended consequences of Google’s success, as explained by Dunning and Kruger

🙢

These search engine issues help explain why many library patrons have difficulty finding the resources they really need. Consider the 1999 Dunning and Kruger study of “meta-cognitive” ability, that is the ability of people to estimate their competence in a field. They found that roughly 2/3 testers, in every field tested, significantly overestimated their ability relative to the group, and were “ignorant of their ignorance.”

There is evidence that library users follow the same pattern that Dunning and Kruger found. True, some library users are quite capable, but many more appear partially or completely ignorant of how to use a library. They appear to be equivocating between the search engine model of research, and the library model—and that is what is causing their difficulties.

8 of 36

How users see search engines

A simple tool for a simple problem
Convenience first, then relevance
Users are unlikely use additional features, unless as a last resort

🙢

9 of 36

The research process is different

It is learning process, emphasis on the process

🙢

10 of 36

Making information discovery easier

One-stop search systems
Emphasize services that search engines cannot mimic
Don’t make promises that cannot be kept
User-oriented focus on quality

🙢

Research is not easy, but it can be made easier. Library patrons typically have to use multiple library systems, and look in more than one place to search through various library resources. This search can be simplified with “one-stop” systems like PRIMO or AquaBrowser, which are designed to search through multiple sources from a single user interface.

Libraries should also focus on emphasizing the services that search engines cannot mimic, particularly reference services and Interlibrary Loan. With reference, these services can now extend beyond face-to-face, to email, online chat, and even text messaging. Library web designers also need to be mindful of this, and remember to keep the patron in mind when creating content for library websites.

Libraries must realize what having digital collections really means.

11 of 36

Competing with everyone else

Digital collections want to be online
Resources are now competing with the whole world
However, SEO legitimizes the search engine model

🙢

12 of 36

Libraries can adapt

SEO is not unlike other library trends (mobile computing and digitization)
Technology allows patrons to use library without setting foot in a library, and without the need of a librarian
Librarians can look for new ways to teach information literacy skills
Search engines are just one tool in an arsenal

🙢

13 of 36

More Than a Map to Get the Treasure

Metadata

🙢

14 of 36

Dublin Core

Contributor
Coverage
Creator
Date
Description
Format
Identifier
Language
Publisher
Relation
Rights
Source
Subject
Title
Type

🙢

15 of 36

Our Dyal Ship Collection is great example of a collection that was improved by supplemental metadata. The collection is marked by having a photograph and a narrative of the ships’ history

A project sponsored by a local sailing ship enthusiast, which in the landlocked semi-arid town of Lubbock is a rare jewel indeed. �

This is the landing page for the collection in our previous content management system -�

In our previous content manager, searching this collection required the user to be pretty familiar with not only the system but also the ships themselves. We received feedback on the content of our collection (positive) but also how to search through it (negative).

Due to the system, we had organized the collection by items, which was good for the metadata librarian, but bad for the users.

16 of 36

17 of 36

18 of 36

19 of 36

20 of 36

21 of 36

22 of 36

Metadata

Digital Collection is not a Democracy

Google Analytics

Talk with your users

The Catalog does not equal the Internet

🙢

23 of 36

More Than a Map to Get the Treasure

Usability

🙢

24 of 36

What is Usability?

Systematic evaluation

Design

Why is it important?

🙢

25 of 36

Usability in Three Easy Steps

Step 1: Assess need.

Step 2: Watch.

Step 3: Diagnose and Treat.

🙢

26 of 36

Step 1: Assess Needs

Find out what users want/need

What are they trying to do?
What do you want them to try to do?

Develop 5 to 7 tasks

🙢

First step is to find out what people want from your resources. You can ask them directly, or you can guess based on what is available.

At TTU, we are planning on offering candy to have people stop by a booth and use crayons to draw out the perfect library website (paper prototyping en mass). We are also giving them a survey asking them what their top five reasons for using the library website are. With something like a digital collection, they may not know what they want from it, so you will have to guess based on your metadata.

Develop 5-7 tasks that represent things that the user needs to be able to do. “Find an image of Frank Loyed Wrights falling water”, “Find an article about management”. The first task should be “easy” and the last task should be similar to the first, to test memorability of how your site functions.

27 of 36

Step 2: Watch

Find 5 or 7 users

Watch them complete the tasks

Watch what they do
Listen to what they say
Measure how well they did

System Usability Scale (SUS)

🙢

Recent usability webinar said it best “Watch people use your stuff”

Person you watch should be one of your users, not a library person.

Easiest and cheapest, bring someone in, sit them down, and look over their shoulder. Give them something for participating. Offer pizza, or dismiss their library fines. (joke, hopefully they laugh).

Look for where they mess up. Ask them to talk about it after the fact.

Triangulation: Look, listen, and measure. If you see them have problems, and they say they have a problem, and then you can show that they had difficulty performing a task, then you can be pretty sure there’s a problem, and the first step toward recovery is admitting you have a problem.

After they’ve completed the tasks, ask them to fill out a System Usability Scale. It’s 10 standard questions that will produce a number that represents the relative usability of your resources. You can use this number to identify if your site is so difficult to use, people will actually avoid it, or if it is at an acceptable level of usability. It’s a quick survey, and it can give you a crucial bit of data about your resource. Say your digital collection system gets a SUS score of 90 (highly usable), but the website that links to it gets a score of 60 (not so usable). Then you know where to focus future usability testing.

http://www.measuringusability.com/sus.php

You can just observe the participants, but there is always the option of recording them.

The simplest setup is just watching someone work. The next step up is a camera over the person’s shoulder capturing the screen and their vocalizations.

28 of 36

Step 3: Diagnose and Treat

Diagnose

Treat

🙢

Through watching people, listen to what they say, and evaluating their performance, you should be able to identify where the problem areas are. Now, you have to do the hard part. Figuring out why those problem areas are problems.

Identify Major problems. Look for 1-3 things that are the biggest barriers.

People have expectations based on previous experience. Most people are fairly comfortable with most websites because they have common ways of doing things. Whenever your site doesn’t do something the way someone else does something, you run into problems.

Be familiar with web standards. A tab should look like a tab. A button should look like a button. A link should look like a link.

Keep like things together. Example: We had a problem with our system where you could search for books from a search box, but you had to go to a tab to search articles, and you had to go to a link to search databases. Our users had huge problems finding anything but books because they expected to find the other things the same way.

29 of 36

Use the “F”

🙢

30 of 36

Keep it simple

Lots of options

Few simple options

What do I do?

This?

Then what?

Cool.

Or this?

Bingo.

What do I do?

This?

Or this?

Or This?

Or this?

🙢

31 of 36

Step 4: Influencing People

Focus on the main problems.

Focus on the nature of the problem

Is change impossible?

🙢

I said we’d do this in three steps, and we did. However, sometimes there’s a fourth step. Sometimes, your not the person in charge of the resources. Maybe a vender controls it. Maybe you have a separate webmaster who was not a part of the usability testing (for whatever reason). The point is, you will probably have to point out the problems to someone to get them fixed, and you will probably have to offer possible solutions.

Keep in mind that the important thing is that they realize there’s a problem. You can suggest different ways to fix it. Sometimes it’s simple, like changing a word. We had a case where students couldn’t search for articles in our new search tool. Big surprise, it’s because we labeled it “databases”- because in our mind that’s what they were. It turns out the problem was fixed when we changed the word to “Articles”. This was

Sometimes it’s simple color changes to make things stand out, or moving a link. The point is, there are many ways of solving a single problem.

Try to focus on the main issues. It’s easier to get one or two things fixed than to hand someone a list of 50 small changes. There’s diminishing returns for fixing the smaller stuff. People are generally able to adjust to small problems. They learn. It’s not necessary to fix EVERY problem.

If you are dealing with a vender, all you can do is suggest changes. What is possibly more important is that if you do usability testing on vender software, you are now more knowledgably about the problems people are likely to face, so you can provide better instruction on how to use the resource.

32 of 36

Getting Serious about Usability

Software: Morae ($1500)

CUEP Certification from TTU ($1000)

EyeGuide eye tracking software ($1500)

Professional Usability lab (About $30,000)

🙢

33 of 36

Stuff to remember

Continuous testing

A few people, a few problems, a few times a year

Webmasters get defensive
Pizza is always good cheap compensation
Recruit many different ways
It’s better to see than guess at problems
It’s cheaper to do it yourself

🙢

34 of 36

How Usability Fits in

Usability of all levels of access

Metadata usability
System usability
Portal/website usability
Search engine context for each page

🙢

35 of 36

The End

Questions?

🙢

36 of 36

Contact

Ian Barba ian.barba@ttu.edu

Shelley Barba shelley.barba@ttu.edu

Joy Perrin joy.m.perrin@ttu.edu

🙢