A good website redesign challenges the status-quo. It questions a site’s navigation, layout, and plan and demands efficiencies without sacrificing user experience. Accordingly, a well-planned and researched design proposal will achieve improved user engagement, page views and time spent. The Engaging News Project, an initiative of the Moody College of Communication at the University of Texas, investigated the impact of design testing to provide news organizations with necessary audience feedback before going live with redesign.

To do this, the Engaging News Project partnered with a major Canadian news site and a major U.S. news site to conduct two experiments. The first experiment, the online test, showed consumers the old sites and the redesigned sites. They were then asked questions about their perceptions of the site. In the second experiment, the live test, the redesigned homepage was shown to a random audience (approximately 8,000 persons) in real-time. Key usage metrics were collected on these users to measure engagement.

The Findings

The Canadian sites redesign focused primarily on the number of photos and the amount of text featured on the homepage. The new site included more photos and less text compared to the existing version. The U.S. sites differed most notably in page width, emphasis on photos, and the amount of content. The new site featured a wider design and more vertical content. The new site had fewer photos at the top of the page, but larger photos after scrolling, compared to the old site.

• Live site tests consistent with experimental online test for most metrics
  The redesigned Canadian site in both the online experiment and live test, was favored. More page views per visitor, a lower bounce rate, and higher average time per visit were generated. Consumers appeared more engaged and spent more time on the site when they went to the site on their own (live experiment) versus those being direct to the site (experimental design).
  
  Interestingly, findings for the U.S. website, both online experiment and live test, favored the old site. The new site registered a higher bounce rate, lower time on page, and shorter scroll depth in comparison to the old site.

• Article recall differs
  Article recall for some articles was better on the new site and for others on the old site. Differences in overall recall corresponded with the presence of images and where the articles were placed on the page. For the Canadian website, those using the new site recalled more articles than those using the old site and for the U.S. website, there were no differences between overall recall on the old or new site. Importantly, overall recall did not differ based on people’s age or based on previous experience with the websites.The Engaging News Project further identified which specific articles on the sites were recalled and what might account for these differences. For the Canadian website, three articles were recalled more on the old site more than the new and six were recalled more on the new site than the old. For the U.S. two of the top 10 articles had better recall on the new site than old and six articles had better recall on the old site than new.

• Improving RecallIn researching individual article recall, the Engaging News Project concluded:

• Pictures impact recall
• Placement in the first column may also affect article recall
• The amount of scrolling to reach an article may impact recall and
• They found no differences in recall when articles are of equal focus in either old or new site.

When undertaking a website redesign, clear objectives and metrics must be defined and tested. Online site design testing ensures consumer feedback and quality assurance prior to going live. It should be included in all news site redesign plans.