Alena KoptsikAlena KoptsikProduct Designer · B2B
Resume

All work

Case study

WELCOME 24 — Internal Platform for Real Estate Agents

From Excel and manual calculations to a transparent performance system

Role
Product Designer — UX architecture and UI
Domain
Real estate · internal platform
Period
2025 – March 2026
Platforms
Web · iOS · Android
+70%agent conversion
+42%agent engagement
WELCOME 24 platform overview
WELCOME 24 platform overview

Project Context

WELCOME 24 is an internal platform used daily by real estate agents and managers. Before the platform: operations were managed in Excel and external sheets, KPIs and revenue were calculated manually, analytics were scattered across sources, data quickly became outdated. The platform grew, but management tools didn't.

Problem

There was no single source of truth. That broke four things at once:

  • An agent's performance couldn't be seen without rebuilding it by hand in Excel.
  • Managers had no view of the network — only per-office spreadsheets that disagreed with each other.
  • Agents had no metric they could act on, so the numbers motivated nothing.
  • Every new office added manual reporting work for someone.

Decisions ran on intuition, not data.

Business Goal

Build one place to run agent performance from. It had to:

  • Hold KPIs, revenue, and daily operations in a single view.
  • Show an agent where they stand and what to do next — not just what already happened.
  • Replace manual reporting with numbers the system calculates itself.
  • Hold up at dozens of agents and at thousands, without a redesign.
  • Work the same on web, iOS, and Android.

My Role

As Product Designer, I had end-to-end responsibility for key product blocks: agent dashboard 'Performance', complete design system (tokens → components → guidelines), agent database, platform navigation hierarchy, UX support for Web, iOS, Android, solutions impacting motivation, revenue visibility, and network scalability.

MLM module was designed by another team member.

Approach & Process

1
Context & Process Analysis
How agents and management actually worked in real life: KPIs and income calculated manually or in Excel, data updated with delays, agents lacked a holistic view of their performance, management struggled to scale motivation and network growth.
2
Product & UX Hypotheses
An agent needs one main screen to quickly assess their current state; data should not just be displayed, but guide the next action; KPI transparency directly increases engagement and activity; motivation grows when progress and growth are visually clear; complex financial logic must be hidden behind a simple interface.
3
System Architecture Design
Defined key entities: agent, deals, income, KPIs, level; designed relationships between them; introduced time-based logic (day/week/month); created scenarios for agents and administrators; made sure the model holds as the network grows from dozens of agents to thousands.
4
Dashboard UX Logic → UI Solutions
Principle: “One screen — one story”: first key metrics (income, deals, KPIs), then dynamics and trends, only after that deeper details. Cards designed for fast scanning: large numeric values, color coding, mini charts, minimal visual noise.
The UI is designed to give the agent a clear answer in 10 seconds, reducing cognitive load in daily use.

Design System as a Product Foundation

The design system was treated not as a UI kit, but as an operational foundation for dashboards, tables, and mobile scenarios. The WELCOME 24 interface was designed strictly within the company's existing brand guidelines. For the business, it was critical that the product felt like a natural extension of the welcomehome24.ru ecosystem, not a standalone app.

WELCOME 24 brand guidelines
WELCOME 24 brand guidelines
This meant working under strict constraints: fixed brand colors, approved typography, predefined visual character.

Based on the official WELCOME 24 brand palette, I built a functional UI color system. Each base brand color was extended into a full scale from 50 to 900, enabling: consistent color usage across all UI states (hover, active, focus, disabled, background, border), a clear visual hierarchy, elimination of random accent usage, room to add new UI states later without inventing colors outside the brand.

WELCOME 24 UI color system
WELCOME 24 UI color system

The typography system is based on two brand fonts defined in the WELCOME 24 guidelines: Roboto Flex — primary UI font for long text, tables, and navigation. Montserrat — accent font for headings and structural blocks. I expanded the basic brand rules into a full product-ready typography system: 5 heading levels (H1–H5), 3 subtitle levels for cards/lists/data blocks, 4 body text levels for different reading scenarios, unified sizing/spacing rules for buttons (from Giant to Tiny), Caption and Label styles for system hints, statuses, and secondary information.

WELCOME 24 typography system
WELCOME 24 typography system

I expanded the brand guidelines into a complete design system: 12-column grid.

12-column grid system
12-column grid system

Click to read it up close.

Buttons, cards, tables, tags, form fields, states.

UI component library — buttons, cards, tables, tags, form fields, states
UI component library — buttons, cards, tables, tags, form fields, states
The system became a single source of truth for Web, iOS, and Android.

Dashboard “Personal Performance”

Dashboard Personal Performance — overview
Dashboard Personal Performance — overview

Dashboard Goal

Give the agent an answer in 10 seconds: “Where am I now? Am I growing? What should I do next?”

Key Decisions

KPI cards (deals, income, activity), time-based dynamics (day/week/month), mini charts, “Assets & Deals” block, levels and progress system (gamification).

Dashboard Personal Performance — detail
Dashboard Personal Performance — detail
The dashboard does not replace full CRM analytics. It follows the principle: One screen — one story. The agent immediately understands: current performance, growth trends, which deals affected results, how close they are to their goal. Deep analytics live in dedicated sections. The dashboard acts as a daily decision-making tool.

Agent Database

Agent Database
Agent Database

The Agent Database is used daily by: team leads, recruiters, agents themselves. They needed a tool with zero learning curve to: quickly find specialists, understand their expertise and geography, instantly contact the right person.

UX Solution

  • Unified agent cards (photo or initials, name, city, specialization, quick contacts)
  • Flexible filters and search (city, specialization, name — with context preservation)
  • Filter-aware pagination (fast navigation without losing orientation)
  • Incomplete profile handling (initials, simplified cards, preserved visual rhythm)
  • Full design system integration (cards, tags, icons, fields, states)
  • Scan-friendly layout (grid, compact tags, clear hierarchy)
As a result, the Agent Database: significantly sped up specialist search, reduced management workload, made the network structure transparent, improved internal communication and decision-making speed, held up as the network grew.

Mobile Experience

The mobile interface was optimized for fast, frequent scenarios: KPI checks, levels, deals, reports.

Click a screen to read it up close.

Mobile screen — 1
Mobile screen — 1
Mobile screen — 2
Mobile screen — 2
Mobile screen — 3
Mobile screen — 3

The product was successfully published in the App Store. At the moment, the iOS app is temporarily unavailable due to external regional restrictions.

App Store listing
App Store listing

For Android and mobile web users, an alternative solution was implemented: access via Web Application, ability to download the Android app directly from the web after authentication, full mobile functionality preserved.

This ensured uninterrupted mobile access and protected core user workflows.

Results

+70%
Agent conversion
+42%
Agent engagement
Automated reportingFaster decision-makingTransparent motivation systemReduced administrative load
WELCOME 24 became a single digital control center for managing the agent network.