Problem
When I arrived at Sapphire Digital, their main SaaS landing page for SmartShopper.com was underperforming. Senior leadership (and many clients) were becoming concerned about the low registration rates and how this was impacting the program's performance and ROI. Due to the following data, they came to the Product Team with the goal of increasing visits to the registration page by 5%.
64%
of users were exiting the site
directly from the landing page
directly from the landing page
14%
only 14% of all sessions included
a visit to the registration page
a visit to the registration page
3.6%
only 3.6% of sessions included
a click on instructional links
a click on instructional links
56%
of users never scrolled beyond
the topmost homepage layer
the topmost homepage layer
Team
I was paired with a Product Manager to tackle this work, reporting in to the VPs of both Product and User Experience.
Research
As we began discussing moving forward with various research methods, we found out that Sapphire Digital had recently hired an outside firm to perform a market research study, which at this point was complete. The research was sitting unused due to the design team being understaffed. The PM and I dove into the outside firm's findings to see if they had any hypotheses that might be worth exploring.
In the end, we decided to perform our own additional research, conducting interviews with our internal customer service team. We were eager to hear the nature of the comments and calls they were receiving from our customers.
30 minute interviews with 6-8 customer service reps
Our customer service team was more than happy to oblige. They enjoyed being included in the process and eagerly shared their service logs and scripts with us.
We broke our interviews down into the following 5 sections:
• Commonly Asked Questions
• SmartShopper Education
• Emotional Callers
• Tone of Voice
• Additional Comments/Questions
Hypotheses
After completing our interviews, compiling and synthesizing our findings, we were able to form the following hypotheses to inform our design decisions going forward:
• users don't fully understand the business model
• users don't understand exactly where the money is coming from
• users don't understand how the rewards values are determined
• users are unclear on the relationship between the program and their employer
• users fear this is a scam or that they won't receive quality care
• users feel that we are treating their healthcare as a commodity
Mapping
Armed with a set of hypotheses to guide us, we began mapping out the page breakdown, targeting each area to ensure we were addressing all the concerns our research uncovered. You can see this breakdown below:
Content Audit and Strategy
Once we reached consensus on the page architecture, we began working on our content. Our first step was an audit of our existing content to inform our strategy. Sapphire doesn't have a content strategist, so we devised our own approach, informed by the Nielsen Norman Group (NNG), and let their "Tone of Voice" be our guide:
"Your website is a tool that enables you to have a conversation with your users, so it's clear that a carefully considered tone of voice is critical."
NNG states further "a website's tone of voice communicates how we as an organization feel about our message, so it is more than just the words we choose. It's the way in which we communicate our personality and it will influence how users feel about our message."
The 3 "content" personas we landed upon were:
• The Expert (is informed, confident, and gives specifics)
• The Assistant (acts as a concierge and guides choices)
• The Buddy (is empathetic, emotional, human, and trustworthy)
Design
Armed with our new content, we entered the design phase, beginning with sketches, and then wireframes. Along the way, one research finding that kept coming up was that even though we had a live-chat feature, it was largely used by registered users. We thought it would be important to assure potential users that there is a real team of helpful humans standing behind this program. We felt that stock photography might not help us achieve this so we organized a photo shoot with our customer service reps.
User Testing
Once our designs were far enough along we launched a set of user tests on UserTesting.com. Due to HIPAA regulations we're often precluded from contacting or recruiting actual users, but since we are able to access anonymous data, we were able to construct our tests with similar target audiences.
When asked to use their own words to describe the program, the user featured in the video below received "high marks" for his accuracy and knowledge of the program's nuances.
Results
With last minute copy and design tweaks in place, we launched the new landing page to a select group of customers for A/B testing. To our delight, the tests were successful in increasing visits to the registration page by a median of 21%, far exceeding our goal of 5%!
Goal:
5%
Increase in Page Visits
Actual:
21%
Increase in Page Visits
Final Designs
