Jonar is a software company that sells applications to help businesses run smoothly. Jonar’s website is one of it’s primary marketing tools, but we weren’t converting enough leads to customers using free trials of the software.
Several UX issues on our website were found and addressed, resulting in more contacts and sales.
Scoping out the problem
Our sales team provided the marketing team with feedback that they often heard from current and potential customers. In their experience, potential customers often had trouble finding in-depth content about our software.
As a result, potential customers often assumed that our software didn’t offer certain functionality, or that it wouldn’t be a good fit for their business.
The anecdotal evidence was confirmed using Google Analytics. We noticed that almost all of the traffic on our website bounced between only a few pages.
The vast majority of our online content was being ignored – including many of pages designed specifically to increase sales. We suspected this was a main reason that visitors were failing to convert.
Guerilla usability testing
I conducted informal usability testing with friends, family, and new staff members who had little experience with our website. This consisted of asking people to navigate to certain sections of the website, as they narrated what they saw and what their thoughts were
Almost every participant would take extremely long, or fail to navigate to certain pages because of:
- Expecting, but not finding, direct links in our header.
- Not being able to search for features or keywords.
- Not opening the ‘popup’ footer (closed by default).
People managed to find content if it was directly linked in our header – which directly matched the pages where most of our web traffic was getting stuck.
Jonar.com homepage in August 2017
One person in a usability test remarked that “The whole website only has 4 pages” ( based on the links in the header), when it actually had almost 80 pages.
By reviewing UX literature, I found applicable evidence for some solutions which seemed promising:
- A redesign of the navigation bar, with more options (which help users spot keywords more quickly).
- Adding a search function.
- Moving our logo to the right, so it was more easily recognized as a ‘home button’.
- Having the footer always visible, with more direct links than the header.
Categorizing our content
Before planning our new navigation, I worked with the marketing team to lay out our current website structure and navigation. An unexpected benefit of this process was noticing both some gaps in our content, and some redundant pages.
Card sorting exercise
At this point, we knew that we needed dropdowns in our header, providing more links to visitors. To determine what dropdowns and groupings to use, we did a card sorting exercise.
Most of our content fit into a few clear categories. I made a few changes based on input received from both inside and outside the marketing team, and moved on to paper templates.
We used paper templates to test navigation designs with current page layouts. A key part of this exercise was that anybody could help us iterate on our designs – being able to see changes makes spotting issues easier, and testing improvements quicker.
When the design was finalized, we implement our changes online. The upfront investment in UX processes made our implementation much quicker than usual, especially considering it was a significant redesign.
This was largely because we did not make many last minute design changes – as they were caught in research or testing.
In the months following our changes to the website, we noticed:
- Number of visitors contacting Jonar via our website more than doubled.
- Decreases in the bounce rates of all pages, between 4% to 18%.
- Increased traffic to rarely visited pages (high variance).
- Decreased average time on page (about 10 seconds on average).
At first, some of our metrics seemed worrying – but these turned out to have good explanations:
- Decreased average time on pages – This mostly occurred on pages designed to generate sales or leads, which were increasing. In all likelihood, people were contacting us sooner.
- Some of our ‘introductory’ pages had lowered page views. By analyzing actual user journeys (behaviour flow in Google Analytics), we realized this was because many users were now skipping directly to more specific/relevant pages.
You are not the user
We tried to use a representative group of test subjects, based on our knowledge of the industry – but I expect if we had tested actual users, we would have more relevant insights into our UX.
We used a pretty varied blend of UX processes, analytics and testing. A lot of these processes were far from scientific – especially in terms of confounding variables. There are changes everyday at Jonar, so sometimes it can be hard to distinguish which strategies had which effects.
That’s part of the reason we used both qualitative and quantitative information – sometimes the numbers alone don’t tell the whole story.