TUI

Optimising a clunky design system to achieve +81 points on the NPS of devs and designers

2025
Design systems, UI

When I joined the team, the foundations of the design system were already in place, and I was contributing sparingly to it and using it for my day to day work.

After some time, using it felt like navigating a labyrinth. Contributions were not structured, there was no clear strategy and no agreement on how to build components. Even experienced team members struggled to locate and trust components.

This was affecting productivity and quality across the board, for both designers and developers. So, I set out to rebuild.

After the rework, I managed to raise the NPS score across developers & designers by +81 points and improve the guardrail metrics by +24%.

Problem

After years of unstructured contributions and unclear work flows, we found ourselves in a bit of a situation that was affecting speed, morale, ways of working, and ultimately quality.

this was the dev trying to find the components in the DS

this was the dev trying to find the components in the DS

this was the new hire designer who needed to start using the DS

this was the new hire designer who needed to start using the DS

this was the long term designer who needed to maintain the DS

this was the long term designer who needed to maintain the DS

I decided to interviewed 6 designers and 5 front-end developers across teams, to understand their perceptions and pain points. I chose the Net Promoter Score as the North Star metric, and added three guardrail metrics.

Navigation

3.5/5

Clarity

2.9/5

Documentation quality

3.1/5

NPS ⭐️

-45/100

Pain points

  • Hard-to-navigate file structure

  • Bloated number of variants and components

  • Minimal documentation (if any)

  • Clunky, unintuitive workflows

The results were quite clear: a change was needed.

the dev who’s trying to find the components in the DS

the new hire designer who needs to start using the DS

the long term designer who needs to maintain the DS

Goals

Based on the results, I decided to focus on 3 objectives

  1. increase flexibility and ease of use

  2. improve designers’ workflows

  3. improve satisfaction for designers and developers

To achieve this, I planned to focus on 5 main activities

  1. make components more flexible

  2. reduce number of components & variants

  3. improve navigation and structure

  4. improve clarity

  5. improve documentation

Jump to the results

Activities

  1. Make components more flexible

Before the rework, a new component was created for each different case. To improve this, I introduced slot components in the composite components, to act as placeholders that could be swapped to match the designer's needs.

  1. Reduce number of components & variants

In the previous version, a new variant was created for each new case that was needed. To improve this, besides the slots components, I added boolean, text, instance swap and nested properties, to enable the creation of less variants.

  1. Improve navigation and structure

The old version had macro sections, which contained many artboards with different components. This made it very difficult to find what you were looking for. For this reason, I reorganised the sections to host a component (or a set of components) each, to make it easier to locate specific elements.

  1. Improve clarity

Before the rework, very different components like selectors, dropdowns and text fields were all in the same page, and items constituting components were scattered around the artboards. I reorganised the structure in a way that clearly divided items (which are not supposed to be used on their own) from components. This helps having a clear understanding of what to use when designing, while keeping flexibility.

  1. Improve documentation

In the previous version, the documentation was scattered around the artboards (if it was even there to begin with). I allocated a specific section for documentation in each of the sections, to define guidelines.

Results

After the rework, I ran again the same test that I had previously done to measure the impact of the work done. The results were very encouraging.

Navigation

Navigation

Navigation

Before

After

Navigation

3.5/5

4.5/5
+24%

Clarity

2.9/5

4.2/5
+26%

Documentation quality

3.1/5

4.2/5
+22%

NPS ⭐️

-45/100

+36/100
+81 points

Guardrail metrics
+
24% overall
+19% for developers (+12% | +28% | +16%)
+28% for designers (+32% | +26% | +22%)

Guardrail metrics
+24% overall

+19% for developers
(+12% | +28% | +16%)

+28% for designers
(+32% | +26% | +22%)

NPS ⭐️
+55 for developers (from -80 to -25)
+54 for designers (from -17 to +71)

The rework drastically improved system usability and trust. Designers felt empowered to build faster, and developers had a clearer foundation to work from. But more than anything, we re-established the design system as a trusted tool, not a source of frustration. This was a foundational project, and the beginning of building something scalable and delightful for the whole team. Other teams adopted the same approach and best practices.

Drop me a line!

I'm always happy to chat about design and DIY!

I speak…

🇮🇹 Italian native

🇬🇧🇺🇸 English C2

🇫🇷 French B2

🇩🇪 German B2

🇪🇸 Spanish B2

I've been featured on Path to Design!!

Take a look

Drop me a line!

I'm always happy to chat about design and DIY!

I speak…

🇮🇹 Italian native

🇬🇧🇺🇸 English C2

🇫🇷 French B2

🇩🇪 German B2

🇪🇸 Spanish B2

I've been featured on Path to Design!!

Take a look

Drop me a line!

I'm always happy to chat about design and DIY!

I speak…

🇮🇹 Italian native

🇬🇧🇺🇸 English C2

🇫🇷 French B2

🇩🇪 German B2

🇪🇸 Spanish B2

I've been featured on Path to Design!!

Take a look