National Library of Australia - Gateways>

ISSN 1039-3498 | no. 51| June 2001


Testing the Usability of the NLA Web Site

In January this year the National Library of Australia employed a consultant from the Hiser Group to test the usability of the Library’s public web site. The consultant crafted the user scenarios, recruited the 13 testers (mainly through a marketing research agency), facilitated the tests and submitted a report of the outcomes. The Library’s role was to inform the consultant as to our business, our users, why they might visit our site and the outcomes their visit should achieve.

What is usability testing?

Usability testing differs from other forms of evaluation in that it does not tell you who is actually using your web-based services, whether their needs are being met or whether you are reaching your target clientele. Web site usage statistics, focus groups and surveys are better tools for measuring these things.

Usability testing is a quick way of detecting the weaknesses of a web site that may otherwise go unnoticed. The developers may be too close to the site to recognise terminology or workflow issues. Users may report a high level of satisfaction through surveys or feedback forms even when a site has significant usability problems. The quality of the content or service may be more important to them than how easy it is to use. They may not be consciously aware of usability issues or may not have particularly high expectations of the site.

It is also possible through usability testing to compare how a web site measures up compared to other sites targeting the same kind of user. This is achieved by taking the testers through the same scenarios on the other sites.

Large numbers of users are not needed. Jacob Nielsen argues that testing with as few as five users will identify over 75 per cent of usability problems [1]. The testers do not even need to be current users of the site. The hardest part is standing back and letting the testers work through the scenarios on their own.

What did we learn?

Through this round of testing, we confirmed that the strengths of the National Library web site are its content and the relative consistency of navigation aids across the site. However, we also found a number of weaknesses in the structure of the site and the graphic design that hindered the testers from completing their set tasks.

Problems that can easily be fixed are now being addressed. These include true bugs (broken or missing links, incorrect headers and footers and links to old versions of documents), navigation problems with specific pages and refinements to our web site design guidelines.

We are also having a close look at the graphic design, which testers quite liked at first but then found fairly conservative compared to our benchmark sites—the Library of Congress and the National Gallery of Australia. Additionally aspects of the current design are not aiding users to choose the most direct navigation paths.

Some of the problems testers found with the site had already been targeted for improvement through enhancement of the technical infrastructure. ‘What’s On’ pages are now automatically generated from an underlying events management system. A project is underway to merge Images 1 and the Pictures Catalogue to provide a single point of access to the Library’s Pictorial Collection.

A significant set of usability issues centring on web access to the catalogue will prove harder to address in the short to medium-term. Many libraries face similar problems: the need to provide multiple versions of the catalogue (Java, Z39.50 and Telnet); a limited capability to customise the user interface; and the OPAC service model itself, which may no longer be the best way of delivering access to the Library’s collections. The testers did not necessarily understand the distinction between a site search and the catalogue, or our collections and our merchandise.

In terms of our strategies for Internet service delivery, the testing highlighted the large gap that still exists between the services the Library currently delivers via its web site and its stated goal to ensure ‘all Australians, at their place of choice, have direct, seamless access to print and electronic sources of information.’ Achieving this goal will be dependent not only on enhancements to the technical infrastructure, but also on adoption of quite new service paradigms.

Cross-program working groups in the Library are currently documenting proposed strategies in relation to future services with an emphasis on e-commerce and electronic information resources. This work was in progress before the web site usability review but the review has given an edge to the business case for the proposed strategies. Initiatives arising from these reports will have a significant impact on future development of the Library’s web site.

An unanticipated outcome of the consultancy was the addition of usability testing to our repertoire of in-house testing and evaluation techniques. The methodology applies equally well to ‘micro-sites’ and specific services.
—Judith Pearce

[1] ‘Why You Only Need to Test With 5 Users’. Jakob Nielsen’s Alertbox, March 19, 2000. Online. Available: www.useit.com/alertbox/20000319.html. 2001-04-03.

| Contents list for this issue of Gateways |

Please address any comments and enquiries to The Editor, Gateways: dthomas@nla.gov.au

Gateways logo