Usability & Accessibility analysis, Performance & Satisfaction testing, Biometric data analysis, Website redesign, A/B testing
2 Weeks worth of work
Conducted during UK C-19 Lockdown
I was asked to analyse the Zara UK website by conducting a full usability review in order to create a website redesign focusing on the core website pages that linked to user's top tasks. This website redesign prototype was then A/B tested against the original live Zara site.
The Zara website has thousands of pages and a review of everything was out of scope of the project length. In order to correctly prioritise focus, 75 students from Loughborough University were surveyed online to identify their top tasks in fashion ecommerce.
Heuristic & Accessibility evaluations based around a 3-task user scenario of using the Zara website to purchase a new jacket were conducted. The scenario was formed from the top tasks.
Based on the methods developed by Nielsen & Molich, a heuristic evaluation resulted in a clear list of design recommendations, these were then assigned priority levels.
Using WCAG 2.1, a full level A through to AAA accessibility review of the site was conducted. The redesign recommendations were then sorted via the MoSCoW technique.
An aspirational competitor of Zara, AYR, was used to conduct a performance evaluation with 15 users in order to highlight areas requiring a systemic redesign.
A Mann-Whitney U test (one-tailed) revealed no statistically significant difference in Lostness between AYR and Zara.
Zara was compared against 4 competitor sites; AYR, Cotton Bureau, Pull & Bear, and Supreme in order to assess its satisfaction competitiveness in the market.
Using 75 participants, a remote survey was used to analyse multiple website's SUS score. Zara was conclusively middle of the pack and therefore presented a great opportunity to improve its user experience.
Several think aloud sessions using a 3 task scenario were conducted and transcribed. These transcripts were then processed through thematic analysis to further understanding of Zara's satisfaction.
Biometrics were the final key analysis area of the usability review and played an important part to create data with reduced user input bias.
Scan path data of using Zara was provided due to remote access limitations to lab equipment. The data was then analysed in terms of attention, direction, and regressive saccades.
Attention maps were generated using the AI EyeQuant tool, which allowed accurate replication of the initial 10 seconds of visual attention by a human.
A full psychological analysis was conducted against the 5 key psychology principles over the 3 levels of human psychology.
A website redesign of key pages relevant to the user top tasks was created in Adobe XD as an interactive prototype so that it could be feasibly A/B tested against the live Zara site.
The Adobe XD prototype was compared in A/B testing for both performance and satisfaction using 2 independent groups of 13 and 6 users. Due to the limited number of participants, only large or moderate effect sizes could be accurately diagnosed.
A Mann-Whitney U test (one-tailed) revealed no significant difference in lostness between the live site and redesign. However, reviewing the graph shows a clear improvement over the variability of lostness achieved in the redesign.
A Mann-Whitney U test (one-tailed) revealed a statistically significant SUS score between the redesign and the live site. This result allows me to see my redesign as a success in terms of improving the usability of Zara.com.