Understanding how users search the University website

Stephen Evans
Thursday 17 November 2016

When someone uses the search box on any of our University webpages, the results are returned using our own Google Search Appliance (GSA). Example search box

The current search results pages are optimised for desktop. For those on a mobile device, the user experience is very poor as you have to scroll and zoom to see the information.

We want to improve the experience of searching the website, particulary for those using a mobile device. To do this, we need to understand how users currently search the University website.

Key features

The current results page has a number of features that clutter the interface. We wanted to know how many people use these features so we could understand the impact of removing them to make the interface simpler.

Filtering results

On the left hand side of the current results page is the option to filter the search results by different categories or by file type. We wanted to know how many users click on these links to filter search results.

Example search results page

Searching within a site collection

We currently have the option to search within predefined site collections (for example, just the computer and IT support pages) by choosing an option from a drop-down list. We wanted to know how many users search within a site collection.

Example search collection

Advanced search

On the top right hand side of the current search pages there is a link to advanced search. How many people use this feature?

Google Tag Manager and Google Analytics

It is possible to monitor whether any of these features are used by applying Google Tag Manager (GTM) event tracking (see introduction to Google Tag Manager for an overview). In this way we can easily monitor the behaviour of our users.

To monitor whether anyone filtered the search results by clicking on the left-hand navigation, we configured GTM to detect anyone clicking on a link with a class of dn-attr-v (this class was already used by our GSA).

To determine if anyone clicked on ‘advanced search’, we tracked anyone clicking on a link with a class of advanced-search (we had to add this class to the link).

Evaluating whether anyone searched within a site collection was more complicated. Two custom JavaScript variables needed to be added to pass the value of the site collection from the drop-down list and the search term. The JavaScript for the drop-down for site collection was:

function() {
var form = document.querySelector('#suggestion_form');
var selectList = form.querySelector('#site_collection');
return selectList ? selectList.options[selectList.selectedIndex].value : undefined;

While the code for the search term was:

function() {
var form = document.querySelector('#suggestion_form');
var field = form.querySelector('#search_term');
return field ? field.value : undefined;

These variables pass the site collection and search term values to Google Analytics.


At the time of writing, we have been monitoring events tracked on the search page for just under a month since 13 October 2016. During that time, GSA had a total of 49,868 page views (42,034 unique). The following interactions were measured.

Filtering search results

422 users used the left hand navigation to restrict the search results to particular categories. The most popular was to restrict results to ‘University’ with 105 clicks. Only 12 users wanted to search specifically within a PDF file.

Use of advanced search

44 users clicked on the link to ‘advanced search’.

Searching within a site collection

In total about 25 users decided to search within a specific site collection by selecting this option for the drop-down list. These were within the site collections for medicine, current staff, psychology, art history, international relations, modern languages, admissions and administration a to z. From the associated search query, it seems that users have very specific queries when searching through a site collection e.g. looking for a particular person.

A larger number (712) searched within the entire site collection (no selection from the drop-down needed).

There is insufficient data to determine what the most popular search term was.


Very few users choose to use advanced search or search within a particular category or site collection. Removing these features would make the interface simpler and not prove detrimental to the majority of users. The most popular task was to do a futher search within the results by using the search box at the top of the page. This feature should be retained.

While this is important and valuable data, we need to interview some users to understand how and why they use the University search pages.

Related topics

Share this story

Leave a reply

By using this form you agree with the storage and handling of your data by this website.