By Luke Hardwick
November 10, 2020
Website navigation not only helps visitors get around your site, it also advertises the site’s offerings. However, because of the limited real estate on mobile devices, making the best use of the main navigation can be tricky. How much should you display? What should be persistent, and what should live behind a click? How are users interacting with your site? A large, UK based grocery chain used SiteSpect to answer these questions.
How to Optimize Navigation
The original navigation for this mobile site lived within the main menu, alongside items like home, search, and cart. When you clicked through, the full navigation then covered the main menu. Competitor research revealed that most other mobile grocery sites feature their navigation separate from the main menu, leaving both visible while in use.
To A/B test this, the brand created two variations. Both featured a new navigation that expanded underneath the main menu using the full width of the page so customers could browse all of the site’s offerings without covering those important top level options. Variant 1 of this new menu also retained the promotional space in the control, with room for personalized product recommendations.
Increasing Add to Carts
To measure the success of this A/B test, the brand looked primarily at Adds to Cart. The new navigation menus each saw over 5% increase in Overall Adds to Cart, suggesting the new menu structure helped customers find what they needed and drove better engagement with products on the site.
Variant 1 had increased performance of Adds to Cart across all taxonomy levels apart from the very highest level. However, this increase was outweighed by lower taxonomy performance in Variant 2, highlighting the customers desire for a distraction-free navigation.
Some lower level taxonomy navigation saw 600%+ increase in Adds to Cart reinforcing belief that customers preferred to shop from this level of the sites navigation.
To learn more about SiteSpect, visit our website.
Subscribe to our blog: