Rivet (formerly Like|Minded) is a software platform using proprietary algorithms to connect colleagues within organizations, fostering lasting relationships. Workplace connections are crucial for fostering a sense of belonging, leading to improved team performance, inclusive culture, and enhanced onboarding experiences. Officially launched in 2021, the team has been working to add features and functionality as new customers join the platform. For admins, filtering and segmenting was something missing from the launch of the platform. Users take a short assessment that provides matches based on the proprietary algorithms.

How might we enhance the admin dashboard experience within the Like|Minded platform to increase connections, engagements, customer referrals and retention? By identifying the major pain points in the dashboard experience, I added the ability to filter and segment data.

Note: Wireframes and mockups shown reflect the brand鈥檚 user interface before rebranding to Rivet. However, all references will be Rivet throughout the case study.

My contribution

User research UX design

The team

1 脳 design lead 1 脳 ux designer 1 脳 engineer




Hearing and learning from customers

With an already launched platform, we heard from customers about features they鈥檇 like to see. Taking note of the highly requested features, the ability to filter and segment data became quickly popular.聽

The filters would be based on user data (anything asked in the initial assessment) and can help admins focus on specific subsets of their users. Admins could also ideally save filter configuration segments. This allows them to get quick and repeated access to their most relevant segments by saving their preferred filtering criteria for regular use.

Examining other SaaS dashboard filters

With a lot of data stored in a product, dashboards are very complex. They need to present lots of numbers and metrics in an organized yet digestible way that won鈥檛 cause information overload. I looked at mostly indirect competitors with comparable dashboards. I found patterns for filters and segments that I used as a reference for the Rivet dashboard.

Key Takeaways

  • Endless ways to visualize filters: I found there were infinite possibilities on how to present filters in a dashboard. It really comes down to the amount of filters and the complexity of them. For our assessment filters, we'll need to reduce cognitive load with our amount, especially the specific activities listed in the assessment.
  • Chips are needed to show selections: In order for users to keep track of selected filters, the usage of chips is necessary. Most dashboards analyzed used chips to show filter selections, especially ones with many filters to select.
  • Dropdowns and lightboxes: The use of dropdowns and lightboxes for the filter selection was commonly used throughout viewed dashboards. This is a great way to hide this information out of view and illuminate it only when needed to keep the overall data in the dashboard easy to understand and simple as possible.
No items found.


The HR Professional mindset

With our current customer user base, I visualized a persona of a busy HR professional who is looking for the most efficient ways to do her job due to her workload. Say hi to Lisa! Lisa loves the Like|Minded platform and praises it for its matching abilities, but wants more.

A simple flow

While the act of filtering data is usually a simple click of a button, it was still important to lay out what it might look like so I knew how many screens to make into wireframes. The data from the assessment was also extensive (i.e. specific hobbies).


Designing dashboard filters

Despite a rather simple design like filters, I still sketched out some potential options for the filters. With the amount of selections users could apply as filters for their data, there are many solutions that could be explored.

Since the dashboard is already live, the focus was mostly on the filter modal design, which didn鈥檛 require mid-fidelity wireframes. Instead, I went straight into high-fidelity wireframes.


Taking it out for a test drive

With the high-fidelity wireframes complete, I configured them into a prototype using Figma. With a current customer base, we recruited some current power users to conduct usability testing. I wanted to test the overall quality and ease of use regarding the navigation and flow of the design through the filter modal. I also wanted to assess how integrated it felt within the overall dashboard. Lastly, I wanted to test the ease of applying filters to segment the data.

Perfecting the filter

After we conducted usability testing, we compiled insights from feedback into a deck to distinguish the pain points and successes of the designs in action.

Key Insights

  • 6/10 users initially had trouble navigating to the feature from the Explore screen. Two of those 6 were unable to successfully complete this initial task.
    • Revision: Adding a keyword search for the specific interests users select during the assessment.
  • 7/10 users had expressed confusion and had difficulty interacting with the budget section of screens.
    • Revision: Removing the ability to save the segment, which will be in a future release.
  • 6/10 users expressed difficulty in finding the 鈥淎dd to Itinerary鈥 text link and were further confused by other CTAs and buttons on the Experience detail screen. Admittedly, this is one screen where Airbnb could improve.
    • Revision: Added the ability to export the data segment once a single filter is applied.

Redesigned modal

Reworked the design slightly with a simple UI and changed the selection checkboxes for clearer selecting. The filter selection was initially going to be a dropdown, but ultimately decided on a modal due to the amount and complexity of filters.

Replacing filters

Rivet's assessment has many data points that could be filtered. However, we wanted to launch this new feature with the ones that would be most beneficial to our customers. That's why we decided to replace Hexaco traits with relationship status. I also decided to omit demographics at this time. These could be added later in a future sprint.

Searchable interests

Within each interest category, there are multiple sub-interests. In order to reduce cognitive load, Imade a searchable sub-interest field.

Filter chips

To track what filters have been selected or are selected for a segment, we added chips to the top of the dashboard.

No items found.


So, what鈥檚 next?

Upon my handoff, this feature was eventually implemented and launched to the customers. During Q4 2022/Q1 2023, the startup gained some investors who supported rebranding and redesigning. At this point, I was no longer doing contract work with them. While the feature likely doesn鈥檛 exist today, it was still one of my first real features shipped collaborating with a software developer.

What did I learn?

  • No feature addition is as simple as it seems: I went into this project of adding this functionality thinking it would be simple. When I鈥檝e used search filters on websites, they didn鈥檛 seem super complex to me. However, I quickly learned there are many solutions and ways of going about this design. Nevertheless, it was interesting to hone in on this single feature versus designing out the full dashboard.
  • Understanding development constraints: With a small team consisting of a sole developer, there鈥檚 definitely constraints that exist. I collaborated with our developer, understanding current constraints and designed with these in mind
  • Testing uncovers more opportunities to delight users: During our usability testing sessions, we were able to uncover pain points in relation to this particular feature. However, it also opened up the conversation to speak with current customers on general feedback they may have. As an early stage startup, many of these customers have signed on knowing there would be kinks to work out as the software and business continues to grow. We're now able to take all this feedback and use it to inform future design decisions for sprints to come.

Next project