| Year | 2018–2021 |
| Role | Senior Product Designer |
| Scope | Design System, Frontend, Colour Management |
| Other Avito cases | Read Avito work case |
Design systems are only as strong as their foundations. At Avito, one of the largest classifieds platforms in Europe, we discovered a surprising weakness in ours: colour.
When I joined the Web Design System (WDS) team, Avito's official palette had about 40 colours. Yet when we looked inside the codebase, we found over 2,000 colour variations scattered across production. Roughly 8,000 individual instances of these colours existed — many of them obsolete, experimental, or inconsistent.
For developers, it meant wasted time trying to figure out "the right blue." For designers, it meant misaligned specs and endless corrections. For the company, it meant slower releases and higher costs.
Avito's design system was powering dozens of product teams, each responsible for parts of the platform. With multiple code styles, conventions, and legacy tech, colour consistency had quietly fallen apart.
Our primary metric was time to market: if we could save developers time and reduce bugs, we'd cut operational costs across the company.
Developers couldn't tell which colours were "official" and which were legacy. Designers struggled to enforce brand consistency. Duplicate or rogue colours slowed releases and introduced visual debt. The scale was massive: thousands of colours vs 40 in the official palette. This wasn't just a cosmetic issue. It was a systemic inefficiency.
I was the sole designer on the WDS team, working closely with developers. My background as a front-end engineer helped me bridge communication between design and engineering. My responsibilities included defining the problem scope and advocating for prioritisation, partnering with engineers to design a colour audit tool, creating guidelines and workshops for designers across Avito, and presenting results internally and externally (including at Avito Design Meetup, with 180+ attendees).
We crawled all repositories and ran a full colour audit. The results were clear: 2,000+ unique colour codes in use, most of them redundant.
We built a web app that analysed colour usage, flagged non-standard colours, and suggested the closest match from the official palette. In controversial cases, designers could manually choose.
The app even generated code patches on the fly, so developers could fix rogue colours semi-automatically.
We excluded intentional exceptions (e.g., illustrations, marketing microsites) and tested the tool across different product teams. Designers validated mappings; developers applied patches in real code.
Alongside the tool, I organised design workshops to align on colour usage. These sessions helped codify rules and reduce future inconsistencies.
We reduced active colours in production from 2,000+ to ~400 (mainly in illustrations/promos), cut developer confusion and rework (saving an estimated 1,200 hours annually), improved time to market across multiple teams, and enabled designers to communicate clearly with engineers through consistent guidelines. I also delivered a talk at the Avito Design Meetup with 180+ attendees, boosting knowledge-sharing culture.
Defining criteria for "closest colour" wasn't trivial — sometimes the algorithm's choice didn't match design intent; we solved this by allowing manual overrides. Not all colours should be unified; marketing assets needed freedom, and balancing standardisation with flexibility was key.
System debt compounds silently: colours seemed like a small problem until the audit revealed its true scale. Developer experience is as important as designer experience; solving their pain had a direct business impact. Fixing the codebase was half the job; the other half was educating teams to prevent relapse.
What looked like a minor "cosmetic" issue turned out to be a multi-thousand-hour inefficiency. By tackling colour management head-on, we not only improved Avito's design system but also reduced costs and improved release velocity. For me, this project reinforced a lesson: sometimes the least glamorous design system problems have the biggest business impact.
