Home / Success stories /Sales platform for seamless customer interactions

Sales platform for seamless customer interactions

Slogan

Assisting a fashion-tech company to create an innovative sales platform that revolutionizes retail and wholesale businesses.

About client

This is a Swedish fashion-tech company that is on a mission to connect brands, salespeople, and consumers through advanced technology. Using its solutions, retail stores and wholesalers can drive more profitable operations.

Industry

Fashion tech

Services provided

Front-end development, back-end development, software testing, staff augmentation

Tech stack

Front-end: React, Typescript, SWR, Zod, Tailwind, dnd-kit, Workbox

Back-end: Python, Node.js, Hono, Fastify, PostgreSQL, Cloudflare (Workers, KV), Google Cloud (Run, Buckets)

Service communication: via Cloudflare Queues

Dev & QA Team

Sweden

Country

Since 2021

Cooperation

About project

This is a cloud-based SaaS platform specifically designed for retail stores and wholesalers in the fashion industry.

It offers a wide range of convenient tools that simplify the clothing sales process, merge physical and online commerce, and enhance the overall shopping experience.

Thanks to an intelligent preloading and caching system, the platform boasts high performance while handling large volumes of media content.

Additionally, it supports digital showrooms, enabling fashion brands to transform their pre-order and retail processes. This minimizes the need for physical samples, improves brand perception, and boosts sales by allowing clients to interact with the full product range via interactive screens.

Our engagement

Our team had a goal to update two critical applications that were somewhat outdated and not as efficient as needed. As a result, the client faced multiple challenges that needed urgent resolution. Without intervention, they risked losing competitiveness, experiencing declining conversion rates and retention, and slowing growth in both B2B and B2C sectors.

To address these issues, we thoroughly analyzed the applications, identified key challenges, and developed effective solutions.

In-store challenges & solutions:

Limited search capabilities
The system supported only full-text and ID searches, lacking semantic search and inventory-aware filtering. This made product discovery difficult and led to lost sales opportunities.

AI-powered search
We implemented an intelligent, AI-powered search mechanism that enables descriptive phrase searches (e.g., "New Year," "Halloween") and ranked searches based on product popularity and inventory levels. As a result, search result speed was improved by ~30%.

Lack of recommendation mechanisms
There was no functionality for suggesting similar or complementary products or auto-generating outfit collections. This deprived customers of inspiration and reduced the chances of completing looks.

Recommendation system
Our team developed an advanced recommendation system with four key recommendation types: similar products, complementary products, pre-designed outfits, and AI-generated sets based on selected items.

Complex installation and updates
The old app needed to be installed on every device manually via the command line, leading to time losses, resource wastage, and a high risk of installation errors.

Web-based solution
It was decided to switch to a fully web-based solution to avoid manual installations and enable automatic updates on the server side.

Instability due to local storage
Vast volumes of media were cached locally on kiosks, leading to memory shortages and increased app loading times.

Central media service
Our developers built a central media service with internal caching and preloading, ensuring fast loading regardless of media quality or size.

Limited UI/UX on touchscreens
The app had an outdated UI and poor UX due to complex navigation and a multipage structure.

New app interface
The team created a single-paged app with infinite scroll, larger fonts, interactive elements, and intuitive navigation. Also, it embraced a new UI library, ensuring consistency across all platform apps. User satisfaction with the new interface increased by 25%.

Results

Greater flexibility and scalability

Faster feature rollout, seamless partner integrations, and 30% faster production changes.

Fewer errors and
faster updates

Transitioning to a web-based solution reduced critical deployment errors and shortened kiosk setup time by 70%.

Faster launch and improved stability

Cloud caching reduced local cache size by 95%, eliminating memory-related crashes, and cut app startup time by 40%.

Increased conversion
and average sales

Semantic search increased top products conversion by 5% while recommendations boosted purchases by ~12%.

Higher user
engagement

Enhanced UI/UX increased kiosk usage time by 18%.

Innovative platform features

AI-powered background removal

Users can automatically remove backgrounds from product images, enhancing visual appeal and design flexibility.

Image-based search

Similar products can be found by uploading images, which significantly simplifies searching.

Automatic image classification

AI-powered classifiers automatically label and categorize product images, enhancing catalog navigation.

Image-based recommendations

The app analyzes model photos and suggests visually similar clothing items, helping users recreate the desired looks.

Showroom challenges & solutions:

Low performance
There was a need to maintain high performance, as the app processed large media files like images, videos, and 3D models. Without effective preloading and caching, rendering delays negatively affected the user experience.

Optimized app performance
We significantly optimized app performance by implementing several efficient solutions:
• Virtualized rendering reduced rendered elements by 70%, reducing load times.
• Custom media manager efficiently handles hundreds of requests, boosting load speed by 60%.
• Optimized preloading and caching cut wait times by 50%.
• Video previews reduced data usage by 80%.

Lack of convenient sharing options
Salespeople needed to share media boards via public links or PDFs, maintaining speed, reliability, and scalability with high-resolution media.

Robust sharing options
We implemented a Node.js-based backend that:
• Generates PDFs in under 3 seconds.
• Reduces the size of PDF files by 40%.
• Ensures scalable processing, handling multiple simultaneous requests efficiently.

Drag-and-drop functionality implementation
Seamless drag-and-drop functionality across desktop and touchscreen devices was needed to enable the creation of custom views.

Drag-and-drop feature
We integrated a specialized drag-and-drop library, ensuring:
• 100% compatibility with both touch and mouse interactions.
• 40% faster rendering with optimized drag events.
• Smooth performance even with large datasets.

Ensuring a fast-loading experience
There was a need to efficiently prioritize media requests for a fast-loading experience, focusing on visible content first.

Increased loading speed
Using the Async library, we achieved:
• Instant loading for visible media.
• 50% faster perceived performance by deferring non-essential downloads.
• Optimized bandwidth usage, reducing network load.

Layout adaptability and consistency
The layout needed to remain adaptive and consistent when users zoomed in via the browser.

Updated layout
To address this challenge, we implemented several solutions:
• CSS media queries to adjust elements dynamically.
• Dynamic font scaling, preserving layout integrity up to 200% zoom.
• Zero layout shifts, ensuring a smooth experience across all resolutions.

Results

Fast and smooth performance under heavy media loads

Our performance optimization initiatives significantly reduced data usage and shortened load times.

A more intuitive and efficient user cross-platform experience

Users can now leverage drag-and-drop functionality to create custom views while enjoying smooth performance.

Improved responsiveness for media-heavy pages

The use of the Async library ensured instant loading of visible media and a 50% faster perceived performance.

Better zooming experience

CSS media queries, dynamic font scaling, and zero layout shifts enabled layout adaptability and consistency.

Fast and user-friendly file sharing

A Node.js-based backend enabled seamless generation and sharing of small-sized PDF files with high-resolution media.

Client success

Our collaborative efforts resulted in implementing cutting-edge technologies, introducing innovations, ensuring high performance, and pushing client’s product offerings to new heights. By leveraging modern technological solutions and best coding practices, we've created intuitive, interactive, and highly performant user interfaces and built advanced features to deliver an unmatched user experience.

With our contribution to the development process, the platform has undergone significant technological advancements, resulting in enhanced product capabilities and a strengthened market position.

Today, the company has numerous clients across diverse European countries, and numerous fashion brands use its platform to make selling easy.

+11

Lviv

+24

Kharkiv

+15

Kyiv

+48

Poland

+12

Germany

+21

Lithuania

+19

Latvia

+12

Slovakia

Greece

Portugal

Netherlands

+15

Estonia

+21

Czech Republic

Andorra

With a pan European talent pool, Devico brings together the continents best talent and makes them available for you