Introduction
In early 2022, there was a need from Nationale-Nederlanden to revamp their health insurance application process (NN Zorg). This is because the current application funnel is highly outdated. The product has been in existence for over 9 years and has seen minimal structural improvements during that time. Small UX adjustments, as a result, are time-consuming or sometimes even impossible.
As a consequence, it's challenging to create added value in a highly competitive market. This is evident from the data we have, which shows a significant drop-off. A whopping 30% of unique visitors drop out in step 1 (awareness & orientation stage), followed by approximately 95% of the remaining visitors in step 2 (decision stage). Users in user research indicate that this is partly due to them finding the user interface not user-friendly. Therefore, the honor is mine to come up with a better application funnel capable of achieving better conversion rates.
Key challenges
- How can I manage expectations from both NN and OHRA, considering their geographical distance and tight schedules?
- How can I seamlessly integrate two funnels, NN Zorg Pakketvinder and NN Zorg?
- How can I minimize cognitive overload for users?
- How can I ensure a seamless experience with NN products, considering OHRA doesn't have access to our design system?
High-level process
I began this project by working with stakeholders to define the briefing and success criteria. As a UX Designer, I believe in designing meaningful products, and that requires a deep understanding of the context and success criteria. To manage this effectively, I created and implemented the UX briefing canvas. As a result I could pin point some challenges that were coming my way.
You might be wondering, what are those challenges? So, let's explore the first one.
Challenge(s) and solution(s)
- How can I manage expectations from both NN and OHRA, considering their geographical distance and tight schedules?
To kick things off, I conducted a remote workshop with stakeholders from both organizations. To maximize efficiency, I combined elements of Discovery and Empathy workshops. I had several reasons for taking this approach:
- The decision-making authority for NN Zorg actually lies with OHRA, not NN, as many believe. This situation results in cross-dependencies, and I want to avoid potential issues and misunderstandings down the road.
- The product needs to be ready before the healthcare season starts (between November 12th and December 31st in 2022). This requires planning at least one month in advance, accounting for development, quality assurance checks, and unit testing.
- OHRA is located near Arnhem, while NN is situated in The Hague. That is a significant distance of 120 kilometers between them. Therefore, it's not practical to arrange an in-person meeting.
Given these factors, it's essential to align everyone swiftly. Fun fact: involving stakeholders in workshops promotes teamwork and reduces future problems because they had a say in the decisions.
After the workshop, I used the insights to outline three different concepts, each emphasizing key insights and offering a broader view of potential directions. It's important to note that I kept these concepts low-fidelity at this stage because there was no need to work out all the details just yet, which would have been a waste of effort and resources. The primary goal was to collaboratively decide what should or should not be included in the final concept since, often, some ideas may not make the cut. By creating different concepts, I could challenge them to focus on the long term, enabling us to plan for future scalability in our architecture. After all, you don't want to build a foundation that isn't scalable, as modifications can be quite expensive!
Next, I pitched these concepts in a remote Design workshop. After each pitch, stakeholders had the opportunity to provide feedback. This approach ensures that they don't forget the first concept by the time they reach the third. Finally, they expressed their preference for a concept through 'Dot voting.'
To prevent fraudulent voting, I explained that the concept with the most votes serves as the foundation, but they can choose to incorporate features from other concepts if desired. I opted for Dot voting because I wanted to quickly converge different ideas into one concept within a short timeframe. Dot voting allows users to easily and swiftly express their preferences.
Challenge(s) and solution(s)
After understanding the core concept and necessary features, my next step was to outline the user flow. This serves as the blueprint for the design, which then brought me to the next challenge.
Challenge(s) and solution(s)
- How can I seamlessly integrate two funnels, NN Zorg Pakketvinder and NN Zorg?
To achieve this, I worked closely with stakeholders to create a business/user flow. This process allowed me to map out use and edge cases, unveil hidden interactions and requirements, and establish the sequence of questions. As a result, it deepened my product knowledge, providing insights into more straightforward methods for handling specific functionalities, such as identifying the optimal placement for introducing NN Pakketvinder and eliminating redundant steps within both funnels.
After mapping out the business/user flow, I move toward creating a high-quality design. During this time, I had to face my other two challenges.
Challenge(s) and solution(s)
- How can I minimize cognitive overload for users?
Data highlighted that the drop-off occurred because there was an abundance of complex information on the screen (health insurance information is inherently complex). In the previous version, all the steps were displayed on the same page, causing users to experience cognitive overload (particularly on mobile devices, where screen space was limited). To tackle this issue, I chose to break down the entire process into manageable sections using a design pattern known as Progressive Disclosure. This method spreads information and actions across multiple screens, allowing users to focus on one step at a time.
Challenge(s) and solution(s)
- How can I ensure a seamless experience with NN products, considering OHRA doesn't have access to our design system?
I stuck to the plan of following NN's Brand Guidelines and using our Design System components for consistency. Although they can't directly use our design system, they can style their components following our design guidelines. The downside is that they'll need to update components manually when NN updates its design system. To make this process smoother, I connected the development team with NN's Design System team for better collaboration. If I can't find a suitable component, I create a new one, validate it with users, and propose it to the Design System team as part of their contribution process.
.png)
After completing the high-fidelity designs, I developed a prototype for user testing. It's worth mentioning that I only added interactivity to the necessary sections for testing and updated these parts with authentic content since content plays a crucial role in usability. The goal was to validate my assumptions rather than investing time and resources in building the entire product, which can be costly and time-consuming. After all, time is money. Making things interactive is crucial for users because it helps them assess intuitiveness by 'feeling' their way through the product.
Next, I proceeded to validate the designs with end-users in collaboration with the NN Research team. We conducted remote moderated sessions with users, allowing them to interact with the prototype as if they were applying for health insurance in a real situation. This was followed by a post-interview to gather their feedback about the experience. We opted for this scenario because we wanted to validate two short-term succes criteria (FYI: we had defined four for this project):
[GOAL] The objective is to enhance the product's usefulness and usability.
- [SHORT TERM METRIC] Our aim was to assess the number of positive responses during these tests, which we achieved by analyzing their feedback during the post-interviews and combining it with the error rate metric (as described below).
- [SHORT TERM METRIC] Users can complete their tasks without encountering errors. We gauged this by analyzing the number of errors, such as instances where respondents got stuck in the process, during the tests.
- [LONG TERM METRIC] The NPS score remains stable and doesn't decline due to the new NN Zorg product. We measured this by examining the NPS score and the feedback provided.
[GOAL] Customers perceive value in our product and are able to complete the process.
- [LONG TERM METRIC] The number of users who successfully complete the process and become NN customers. We assessed this by looking at the conversion rate.
By validating in advance, I can make improvements based on user needs and eliminate any accessibility issues before the developers start building the actual product. Making changes in a live build is much more complex and expensive. Fortunately, the results were highly favorable. Out of the seven people we tested, none encountered any issues during the process. They also responded positively to the sequential presentation of information and actions.
With the success, I proceeded to finalize the desktop designs and handed them over to the development team. Then, it was a matter of waiting to see how it would perform in the future. [Update 2023: Based on the data from Adobe Analytics, we observed an increase in leads and a significant decrease in dropouts. That's certainly a positive outcome!].
Challenge and solution
.png)