Tree testing is a way of testing your information architecture, or the way you’ve structured the navigation and flow of content on your website. We used this method of testing to better understand potential patients who might interact with our marketing platform, and we found out we definitely had some changes to make to smooth out patient experiences.
It’s not enough anymore to simply continue offering the same product or service you have before. Businesses must always improve to be able to cause disruption in their market, or, at the very least, fend it off.
With that in mind, we’ve been going over our processes and our services extensively in the past year.
Before I jump into some of our results from the tests we’ve run, I’d like to give you a quick bit of background about our company. Like a lot of other marketing providers, we have a few marketing challenges to balance. We provide marketing services for our clients to be able to attract new patients to their practices. We also do some marketing work for larger companies in the healthcare space. Again, the goal is to attract the right patients or customers to their services.
Then, we must also attract customers to buy our products and services. So, we market for others, and we market for ourselves. As a part of this marketing service for others, we have a marketing platform that we’ve built over the years and refined for our key audience. This audience seeks to serve patients all over the country.
That platform requires that we add new features on a regular basis, but we must also review what we have in place to keep up with changes in healthcare regulations, online marketing needs, and the patients themselves.
We work with clients in a variety of ways to measure the impact they have on patients and consumers. We use analytics and a variety of other tools in order to be able to make custom recommendations to each practice.
Last year, we used a new tool to better understand how patients seek out information.
Based on our research, a company called Optimal Workshop had the best pricing and solution for our needs with a service called Treejack. Here’s a quick overview of how the service works, and then we’ll get into how we actually used the service to improve our product.
First of all, you set up your navigation structure in the system. Your audience is only going to see the navigation itself, so you can’t rely on on-page content or large calls to action to get visitors where you want them to go. You’re only testing how well the navigation is structured (how intuitive it is).
Once you have your navigation set up, you create a set of tasks for your audience to complete. For instance, you may instruct your audience to find the address of the business or find out how to return a product.
When you have your set of tasks complete, you find an audience to test and invite them to complete the tasks. What I liked about Optimal Workshop is that they had a built-in way of recruiting and audience based on our demographic selections. This worked for us because the patients our platform interacts with most comes from a broad population base, mostly restricted to a certain age group. (We’ve kept in mind that this isn’t a perfect representation of the final patient, but it’s a good start for this test.)
If you have a more specific audience than what you can target with simple demographic selections, you’ll need to invite your users directly. For example, if you are targeting hospital purchasers, that’s not going to be an easy box to check in a tool like this.
In order for the test to really succeed, you will want to test out more than one navigation structure at a time. Only testing one will help you understand how easy it is to use the navigation you currently have, but testing two will allow you to get to real conclusions about the way people are engaging with your navigation much more quickly.
We actually tested three navigation structures on our first round of tests: one of the current nav, and two new nav structures. I should say before we go too far into this that we as a marketing team did not create the structures on our own. We worked with account managers and designers to create a few options.
I have to say I felt pretty confident that we would have a clear winner among our three options. Of course, that clear winner would be the navigation structure I liked the most.
I was wrong.
The navigation structure I preferred “won” by a slim margin, but it was not enough to be a conclusive win.
The single trend that surprised and helped me the most was how people selected varying answers that all made sense to specific questions that we had.
For instance, if we asked a potential patient to find where the insurance information should be on the website, we received a handful of logical answers. With the way that we had structured the tasks and their answers, there was only one right outcome. When we saw how many other options people had selected, it provided the insight that our patients could be mimicking the exact same behavior when they try to look for insurance information. By only mentioning it in one place on the website, patients were being forced to click through a multitude of pages in order to find what they needed.
When we created our next round of testing, we created new navigation structures, and we created more flexibility in our answers. The way we were able to create more flexibility is by allowing users to have more than one right answer for tasks where users had clearly indicated that there was more than one right answer. In the case of the insurance scenario, we provided information on a few different pages instead of forcing patients to hunt down the one page we’d written.
We still included our original navigation as the baseline, but we were sure to include the option for multiple answers on the related tasks, as well. Even still, the new navigation we developed as a result of our two suggested navigations far outperformed the navigation we had before.
Direct and Indirect
One of the fascinating aspects of this method of testing is a better understanding of how seamless (or not) the process is for your audience to complete a task. Optimal Workshop provides users with a variety of success or failure metrics for each task that the user provides.
For example, let’s say that we ask our audience to find directions on how to get to the business. One portion of the audience will select the exact right answer on the first try. This group is classified as having a “direct success.” Others may get to the right answer but may have to click around a bit before they get there. They qualify as having an “indirect success.” Imagine the inverse for both of these scenarios, and you’ll see how audience members could have a direct failure or indirect failure. Optimal Workshop also allows audience members to skip a task if they need to.
The directness of success is a hugely important factor. Think about how easy it is to get frustrated with trying to complete a task online. If you can get the work done but you’re upset by the time you get there, then you’ve still had a negative experience. If you’re putting this in terms of customer experience or patient experience, then your brand has had a setback even if you did complete a sale or book an appointment. You’ve started off the engagement on the wrong foot, and you will have to work hard to win over that person.
That said, keep in mind that you are never going to have a completely perfect navigation or experience online.
When a user finds that a task has resulted in a higher degree of indirect behavior, then you can dig deeper into the data to understand exactly where audience members went in a different direction than you desired. You can get really granular on understanding where the process went wrong. This deep dive into all of the possible answers was one of the ways that we understood how our audience members had some very logical ways of trying to answer the questions that we had provided.
One Interesting Note
One small thing that I found interesting… Optimal Workshop recommends that users do not include the Contact Us page as an option in the navigation structures that you are testing. Because so many people default to the Contact Us page as a way to stop hunting for information, the company has found that including the page can skew results and prevent users from developing a more helpful navigation.
Apparently, some of the big advertising services out there embrace this approach to providing support on their website by burying contact information, as well. But that’s another matter…
I don’t tell you all of this to convince you to use tree testing. As a marketer, it’s important that you know you have the option to use tree testing and how it would be helpful. We have such an increasing number of tools to help with marketing that it can overwhelm, and some marketers back away and just stick to the tools they know. It’s a tendency we all have to fight.
Written by Michael Roberts