The Value of Navigation Order on Webpages

Author: Zach Richard

Much has been written by website designers and marketers on site navigation best practices and the importance of navigation positioning in delivering users a fluid experience. It's generally understood that users will scan a webpage from left to right and that the most important or most relevant navigation items should be placed on either the top or left of the navigation depending on the style. It's also been observed that users are more likely to engage with the final page they visit. Depending on your website this could mean clicking an application button, adding an item to a cart, or sharing the page on social media.

These behaviors are closely tied to the psychological phenomenon known as the serial position effect, which is the tendancy of a person to recall or assign significance to the first and last items in a series, known individually as the primacy and recency effects. In memory experiments, the primacy effect is illustrated by a U-shaped curve representing a persons ability to recall words in a list.

serial position

On websites this behavior can be harder to document, since designers and developers will typically position the most important pages at the beginning of the navigation as best practice. However, when each individual page on a site is of equal value, the importance of primacy and recency become apparent. Such a case recently arose when Notre Dame published a site designed to highlight and celebrate women whose scholarship and leadership are leaving an indelible imprint on the global community, womenlead.nd.edu.

Women Lead Navigation

The entirety of the Women Lead website is comprised of seven pages; a homepage and six pages profiling six female researchers at the University. Pages can be navigated to from the homepage either by clicking on an area near the researcher in the hero image, or in the more traditional navigation below the hero image. Subpage navigation is handled by a horizontal navigation which becomes vertical but retains it's order on mobile devices. For the purposes of this study, we've designated the homepage as having navigation position one(1) and each subpage with numbers 2-7 based on there order from left to right in the navigation.

Primacy

If we look at pageviews coming to each of the subpages from the homepage we can immediatly see the value of the first position in the homepage navigation (labeled 2 in the diagram). The subpage in the first position recieved by far the most click throughs from the homepage with more than 56% more pageviews than the subpage in position two (927 vs 594) and with a total well outside the standard deviation. These results also illustrate the perils of being at the end of the navigation list, with that subpage falling barely outside the standard deviation in the opposite direction.

Graph of Click Throughs From Homepage

It's also interesting to note that despite pages being of equal value and not related in any more granular ways outside of that defined by the site concept in general, users did not click between subpages at random. A typical user moved either from the homepage to the subpage, back to the hompage or from subpage to subpage in the order in which they are placed on the navigation. Meaning, users moved from page two to three, three to four, etc. 

Pageviews volume by nav position coming from previous page

User Flow

Recency

There appears to be a relationship between social media engagements for each page and the las page a user views. Engagements on the official University central outlets was excluded since those should have been equivelent for all subpages. In addition to the homepage, an average of 2.55 pages are viewed during each session. This average in between the pages in nav positions 3 and 4, which would be 2nd and 3rd subpages viewed if users follow a typical behavior. These are also the pages with the most social engagements, suggesting an importance of in recency.

social_media_shares

This calculation does not take into account the propensity to use social media of each of these researchers, which could have a significant effect on these results as well. A more accurate study of this effect could be performed by factoring in that variable or by adding social buttons/prompts to the individual subpages and measuring those results.