Small-Scale Usability Testing “ Evolution Not Revolution” Darlene Fichter [email protected] March 12, 2003 Computers in Libraries 2003.

Small-Scale Usability Testing “Evolution Not Revolution”

Darlene [email protected]

March 12, 2003

Computers in Libraries 2003

Outline

What is usability? What is user testing? Why conduct tests? 3 types of small scale tests

Cognitive walkthroughs Preference testing Task based testing

What is usability?

Ease of use

Ease of learning

Fitness for purpose

An effective product

What is user testing?

Involves actual users interacting with the web site

Typically users are asked to perform tasks while usability evaluators observe and take note of their actions

Why conduct user tests?

How do you know it works?

Why test?

As web managers we want to offer the best service that we can

Even the best designers are not representative of the users of their systems Poor judges of what users want and

need

Why test?

Web development is expensive, and supporting a poorly designed system is even more expensive

“pay now or pay later” Unhappy users who will leave; often

social and political costs

Reasons for NOT conducting tests

1. Too much money Usability testing can fit any budget

2. Not enough time This is not the case

3. Requires a expert consultant There are many tests you can do

yourself

Usability Evaluation Techniques

Less than an Hour Cognitive walkthrough

Less than Half a Day Preference testing

1 Days to 2 Weeks Task based testing

Cognitive walkthrough

Development team collectively walks through the site

They assess: Whether the user has the right

information on the current screen to make the next right action

Whether the user even notices the option

Whether the user will confidently make the right choice

Cognitive walkthrough works best when:

You clearly identify and personalize the user: Joyce, a 40 year old nursing student,

has to complete a paper on asthma education programs. She is working from home …

You designate a recorder to take notes of trouble areas

You have an open culture where you can question things, “would she”?

Typical user of a public library site

Jason, 12 years old, keen reader Likes Stephen King’s books He’s in the library, read all the

books at his local branch, and wants to request books from another location.

Old version

Old version of catalogue

Default is author

Stephen King

Results screen

New design

•Catalogue is more prominent

•But does it look clickable or like an ad?

•Does “liric” the brand name help or hurt?

New search screen

•Controls are placed BEFORE the box where you type

•Default is title keyword – most common?

•Will he think to change it?

Mockup Stage

Home page See if your “user” could find the correct

1st step on the path to completing the 5 most common tasks for your site

Workflow applications with forms Difficult to design without a cognitive

walkthough

Union Catalogue Search Page Mockup

When can you use a cognitive walkthrough?

Almost at any stage as long as you have a some ideas of the screens and wording

Best used early in development to prevent problems when live

Can be an effective technique to use on a live site before redesign

Preference testing

Also called “cookie” testing Zero in on troubling labels

Example: Interlibrary loan

Which label best describes asking your library to order a book from another library?1. Request book from another library2. Interlibrary Loan3. Request forms4. Books from other libraries

“Live Reference” on Health Sciences Library Site

Words in the label and the design were crucial to the use

How could we get the right words? What about the design?

Testing the Wording

Got a question?  Get an answer online NOW!

Click here

Got a question?  Chat with a librarian NOW!

Ask me!

Questions?  Ask online and we'll answer!

Ask me now!

Don't give up!  Ask us and we'll help!

Click to chat

Not finding what you're looking for?

Ask a librarian

Rounds 4-5: Small graphic and wording

Task testing with users

Users are given specific tasks Asked to verbalize their thoughts Observe, record,

and debrief

Guerilla testing

Guerilla style testing was popularized by Jakob Nielsen

Showed that simple “low tech” testing of five to seven users could yield excellent results

5 users will typically uncover 80% of site-level usability problems

Example tasks

1. Find a 5 scholarly articles on global warming

2. Find a book on patenting human genes3. Find out when your books are due4. Find the phone number for the library

closest to your house.

Measurements

Time on task Number of errors Difficulty in

NavigationUnderstanding site structure

User satisfaction

Edmonton Public Library

Took two tactics for testing: Full blown site wide testing on the

“current” site Small scale testing for iterations of

the home page Much quicker to carry out

Edmonton Public Library Home page

One week Had two designs

One more minimalist One “portal” style with many links

Tested for 3 afternoons Participants were asked where they’d

click to try to complete 7 tasks We wanted a very high success rate on

the main page

New Home Page – 6 Rounds of Tests

Task Based Testing

Excellent way to gather data If you can’t “formally” do tests,

consider “informally” was you watch someone try to find a book or journal on your web site. Train yourself to be an observer

Testing just 1 user is 100% better than testing no users at all

Can short tests be effective?

Absolutely! First testing gets done. You can focus on a particular aspect of the

site that is troubling and fix it anytime. Avoid the pitfalls of waiting for a big

study: Web team is overwhelmed by so many things

to fix that they are parallelized Big studies are timed just before the site

launches – too late to change anything substantive

Key Advantage of Short Tests

You can do more of them Often the “problem” Is easier to

spot than the “solution” Several rounds of testing help you

get a solution that works well

Iterative

In conclusion

Usability testing does not have to be: Expensive Time consuming Difficult

It’s a great technique for making your web site better for your users.

Thanks

Questions?

Darlene [email protected]