Boardia: modernising an analytics platform for different clients
I led the design of the Boardia analytics app redesign: from UX and UI changes and making client-specific components reusable to hands-on implementation work and review. The goal was to prepare an established product for further development while keeping it understandable as data volumes and client requirements grew. Work began in early 2026. Clients started using the first redesigned version that spring.

The product and the initial problem
Managers at different levels of an organisation use Boardia to work with customer survey results, testing results and other related data. Each user can be assigned specific permissions that determine which data and features they can access. The application originally focused mainly on quantitative results in tables and charts.
Ongoing development brought more dashboards, client-specific changes and rebranding projects. There was also growing demand to work with qualitative data and move mystery shopping data analysis from an older application into Boardia. The original interface, navigation and parts of the frontend were no longer suited to this scope.
The main challenge was to create a shared foundation for different client needs while keeping the product clear for users without specialist data-analysis knowledge.
My role and decision-making
I set the redesign direction and was responsible for the design and its execution. My work covered UX, redesigning the UI across the application, modernising the frontend and making bespoke client components reusable. I also translated designs directly into frontend code. The product owner reviewed the proposed direction and outputs and held final approval authority.
Priorities drew on years of experience developing the application, internally collected client requests and problems, and known weaknesses in the original UX and UI. These inputs determined which changes mattered most for the product at that stage.
Selected design decisions
The redesign covered the entire application. The following examples highlight three decisions from the wider scope of changes.
1. Main navigation and progressive disclosure
The original solution was built around one main dashboard and a few supporting pages. The expanded product needed navigation across multiple dashboards, data types and levels of detail.
In the new design, each main page has its own navigation item. Users access details about responses and datasets through interactions within the relevant component or module. They can move from an overview to the source and specific content of the data.
We reviewed changes throughout the process with internal stakeholders who understood the product, engineering and client needs. We incorporated requests and feedback and assessed usability, clarity and fit with client needs. Stakeholders considered the revised navigation a significant improvement over the original.
A clear place for each overview
The main menu opens the relevant group of dashboards. More detailed information stays inside the selected module.
Select the image to enlarge
Interface illustration based on the demo application. English labels and illustrative demo data; presentation refined for this case study.
2. A comparison table for different client scenarios
The comparison table was originally built for a banking client. It started with a preset dataset containing the client's own data, against which users selected other brands and touchpoints for comparison. Their data appeared in an additional column, with differences highlighted.
I redesigned the configuration and data selection. When setting up a dashboard, it is now possible to choose a preset column with the client's data or leave the entire selection open. Users can build multiple comparison columns and work with individual brands and touchpoints or groups of them. The component is also designed for use outside banking.
We preserved the existing table workflow for current clients even though the application's UI changed. The extension did not require them to change their habits. We also began using the same component for other clients who had not previously had access to it.
Read the difference, not just the score
Compare datasets, track change over time and see where a result needs more careful interpretation.
| Metric | Demo brand | Brand − peers | Peer group | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Current | Previous | Change | N | Current | Previous | Change | N | ||
| NPS promoters | 72% | 68% | ↑ +4 pp | 420 | ↑ +8 pp | 64% | 67% | ↓ -3 pp | 510 |
| NPS detractors | 10% | 13% | ↓ -3 pp | 420 | ↓ -8 pp | 18% | 16% | ↑ +2 pp | 510 |
| Good-care cluster | 64% | 60% | ↑ +4 pp | 420 | ↑ +5 pp | 59% | 61% | ↓ -2 pp | 510 |
| Issue resolution | 84% ⚠ | 80% | ↑ +4 pp | 18 | ↑ +6 pp | 78% ⚠ | 76% | ↑ +2 pp | 26 |
NPS promoters
Demo brand
- Previous period
- 68%
- Responses (N)
- 420
Peer group
- Previous period
- 67%
- Responses (N)
- 510
Demo brand vs peer group: ↑ +8 pp
NPS detractors
Demo brand
- Previous period
- 13%
- Responses (N)
- 420
Peer group
- Previous period
- 16%
- Responses (N)
- 510
Demo brand vs peer group: ↓ -8 pp
Good-care cluster
Demo brand
- Previous period
- 60%
- Responses (N)
- 420
Peer group
- Previous period
- 61%
- Responses (N)
- 510
Demo brand vs peer group: ↑ +5 pp
Issue resolution
Demo brand
- Previous period
- 80%
- Responses (N)
- 18 · Small sample
Peer group
- Previous period
- 76%
- Responses (N)
- 26 · Small sample
Demo brand vs peer group: ↑ +6 pp
Why is a decrease in detractors shown in green?
A lower share of detractors is favourable. The component can reverse the colour direction for metrics where a lower result is better.
What does the small-sample warning mean?
The component can mark results with a low base. The threshold is configurable. This illustration marks samples below 30; the values are examples, not client findings.
Difference between groups
The tinted column separates the gap between datasets from change over time.
Statistical significance
A configurable mode shows whether a difference is statistically significant, rather than treating every movement as meaningful.
Illustration of supported comparison-table features, using invented demo data. The significance indicators illustrate UI states and are not calculated from the example values. Presentation adapted for this portfolio; no client results are shown.
Choose the data for each column
The selection sits directly above the values it controls. Here, the first column offers the touchpoints available in the demo project.
Select the image to enlarge
A preset starting point
Start with the client’s data and select other datasets for comparison.
An open selection
Leave the selection open so users can build the comparison for their current question.
The selector shows the touchpoints available in the demo project. Depending on project configuration, columns can also contain individual brands or groups of brands and touchpoints.
3. A neutral foundation for client brands
The original interface was strongly tied to our own brand. Customising it for a client meant first removing that identity and then applying theirs. The changes went beyond colours.
I designed a neutral visual foundation for white-label use. We can now apply a client's brand without extensive rework or removing the original identity. The default appearance also works without individual rebranding.
Default appearance
A neutral blue foundation that can stand on its own.
Select the image to enlarge
Business theme
The same component and data with a different visual treatment.
Select the image to enlarge
Theme illustrations using colours from the application’s source styles. The Business theme is an available application theme, not a client-brand example. Values are demo data.
Adding qualitative data
Alongside quantitative summaries, we needed to make individual responses and their context accessible. The design addressed three levels: sorting and displaying individual responses, comparing them across participants in specific areas, and aggregate evaluation. This allows clients to work with individual responses as well as a broader summary view.
From individual answers to the overview
Start with individual responses
Read responses in a list, then open the detail for the context behind a particular answer.
Select the image to enlarge
Keep the answer in context
The detailed view brings the comment, rating and touchpoint together.
Select the image to enlarge
Step back to the broader picture
An aggregate view complements the individual responses. This example shows reasons for satisfaction.
Select the image to enlarge
Interface illustrations based on the demo response list, detail and summary views. Comments and values are invented demo data, not client findings or measured project outcomes.
Implementation, collaboration and trade-offs
I handled the move to a new version of the frontend framework and interface changes supporting the proposed UI and UX. I worked with a frontend developer on more complex tasks and changes to component logic. A backend developer handled system integration as needed.
For technically demanding changes, we weighed the expected benefit against the cost. For example, we proposed extending the ways charts and information could be combined beyond the existing model of one chart in one component with one set of information. We left this for later development: it was not essential to the current redesign, and we focused our time on other parts of the application.
I compared the implementation against the Figma designs, paying particular attention to navigation structure, individual items and microinteractions that provide feedback during use. The review covered both appearance and behaviour.
Implementation workflow
From design to working behaviour
- 01
Design direction
UX and UI proposals, reviewed with the product owner.
- 02
Implementation
Hands-on frontend work, with frontend and backend developers on more complex logic and integration.
- 03
Review in context
Compare the implementation with Figma: navigation, interface items and feedback during use.
Validation in use and results
After internal validation, we offered clients the new interface as a trial version while keeping the original available alongside it. We assessed adoption by comparing use of the two versions and monitoring ongoing client feedback. Questions about the new version mainly concerned getting familiar with the changed structure and where to find information.
The redesign is now part of the application used by clients. Concrete outcomes include wider use of the comparison component, the ability to apply client branding directly and broader support for working with data. For navigation, we have positive qualitative feedback from internal stakeholders.
Mystery shopping is already live in Boardia. Clients are gaining access to the first migrated projects and starting to use them. Further projects are being moved from the older application gradually.
Rollout approach
A gradual transition
- 01
Internal review
Review the changes with stakeholders familiar with the product and client needs.
- 02
Parallel trial
Offer the new interface while keeping the original available alongside it.
- 03
Use and feedback
Compare use of the two versions and follow ongoing client feedback.
Lessons from the project
In hindsight, we could have done better at breaking this large change into smaller work packages. We did not organise them so that they followed on smoothly, used team capacity efficiently and could be released independently. For similar projects, I would put more emphasis on dependencies and smaller, independently deployable increments from the planning stage.