Case Study·

Pricing Recommendation Tool - Panther ML/AI Pricing Platform

How we helped Panther Pricing migrate from Flask to FastAPI and build a new front-end, creating a powerful machine-learning-based web platform for optimizing pricing in brick-and-mortar stores.
Pricing Recommendation Tool - Panther ML/AI Pricing Platform

Small and medium shops are facing intense competition in the market. The post-pandemic era has caused all retail industries to have the need to use online tools with automatisation systems to compete and survive. Work with data in real-time has become one of the essential solutions for every company to make better choices. This time, Bravelab comes again with a Python background for Panther Pricing to help with this concept and provide a reliable result for all the Panther's clients in Germany.

The plan was to create a cloud software able to generate automated price recommendations for retailers in Germany. The mission was to allow bricks-and-mortar companies to have no more online sales with negative margin contributions. The platform should generate online benchmark prices for the clients.

About Panther Pricing

Panther Pricing is a German company founded in 2018 and based in Frankfurt dedicated to providing innovative online services to boost retailers in Germany.

Technologies Used

  • PostgreSQL - Database
  • Python - Backend development
  • JavaScript - Frontend development
  • FastAPI - Modern Python web framework
  • SQLAlchemy - Python SQL toolkit

Services & Deliverables

  • Frontend & Backend
  • IT Staff Augmentation

Challenge

Panther Pricing encountered three major challenges:

  1. Enhance System Performance and Scalability
  2. Augment Knowledge and Expertise Internally Lacking
  3. Time Sensitivity

Both needs were urgent, emphasizing the crucial role of time in addressing these challenges.

The client sought a reliable technology partner possessing specific expertise, ready to assist their internal development team. Given the urgency of the situation, the client had no time for a prolonged onboarding process and anticipated results at the earliest opportunity.

To establish collaboration, we initiated the client engagement process through the following stages:

  1. Brief and QA
  2. Introduction Call
  3. Discovery Workshop
  4. Ballpark Proposal
  5. Decision

Our approach prioritized gaining a comprehensive understanding of the client's challenges to formulate an optimal solution. This commenced with negotiations to define the project's scope. During this phase, a consensus was reached with the client that developing a new front-end from the ground up and seamlessly integrating it with the existing backend would yield the most favorable outcomes within the specified budget and timeline.

After careful consideration, we identified FastAPI as the ideal technology for the project. The subsequent step involved the meticulous design of a new front end, comprising reusable components. This design choice empowers the client to leverage these components as foundational elements for the incorporation of new features and functionalities, thereby maintaining a consistent and familiar appearance for end-users.

The client's primary objectives included enhancing the overall system performance and establishing a foundation for continuous scalability. In pursuit of these goals, the client sought external expertise in specific development areas. Additionally, a sense of urgency prompted the client to seek additional development resources externally, as the required skills were not readily available within their internal team.

Solution

The team prioritized establishing a transparent and efficient communication process with the client. Immersing ourselves in the project details, we sought to comprehend the scope thoroughly. Bravelab recommended migrating from the old technology (Flask) to the new FastAPI and refactoring the existing functionality. The objective was to optimize and diminish the technical debt of the application. Simultaneously, we adapted to the new front-end and expedited application performance. The solution is built on the latest front-end technologies.

Team Testimonial

Panther project has no middle management. Flat structure enables us to take the responsibility, make confident decisions, and contribute to the client's product development.

Mateusz Bryzik - Frontend Developer

Client Testimonial

The team consisted of three an experienced and highly motivated developers, we had frontend and backend developers, and we were working closely with UI/UX designers.

Nils Streitbürger - Managing Director, Panther Solutions GmbH

See full review on Clutch

Summary

Scope of Work

  • Concept idea and research
  • User interface and platform design
  • Platform development
  • Testing & launch

Insight Numbers

  • 1191 hours invested on app development

Developers Assigned to This Project

  • Mateusz - Frontend Developer
  • Michał - Backend Developer
  • Jakub - UX/UI Senior Designer

The Team

  • Mateusz - Vue.js Developer
  • Michał - Python Developer
  • Jakub - UX/UI Senior Designer
  • Mariusz - Delivery Manager

Technologies

  • Python 3
  • PostgreSQL
  • FastAPI
  • SQLAlchemy
  • Vue.js (in frontend)

What We Did

Bravelab undertook the migration of the codebase from the old technology (Flask) to the new FastAPI, coupled with a comprehensive refactoring of the existing functionality. The primary goal was to optimize and reduce the technical debt of the application. Simultaneously, we adapted to the new front end, significantly accelerating the application's performance. The solution was built on the latest front-end technologies, resulting in Panther Pricing boasting a consistent user interface and achieving high-performance levels.

Main Goal and Challenge

The main goal and the most significant challenge revolved around building a library of design patterns, rules, and UX guidelines to prevent inconsistencies when deploying products at scale.

UI/UX

Our UI/UX Designer focused on crafting a collection of reusable components, adhering to clear standards. This approach facilitated the assembly and construction of numerous applications within the platform while maintaining consistency throughout. The product transcended a mere compilation of reusable UI elements; it possessed structure and meaning, representing a system of concepts for sensible usage.

The Result

The comprehensive integration of research, design, and development led to the enhancement of a powerful platform capable of automating 100% real-time price recommendations using AI. The platform now features:

  • Fully integrated online price labeller
  • Always up-to-date online market prices for each item
  • Ideal online price proposals for optimal exploitation
  • Definable price minima via calculation/coverage margin requirements
  • Stationary price orientation to the current online price for sensitive product segments or strategic positioning of maximum prices (online +X%)
  • Inclusion of AI-based sales forecasts, price elasticities, online competitive pricing, item performance, current inventory levels, lifecycle, and order availability considerations
  • Optimization of price markdowns
  • Yield improvement
  • Liquidity improvement
  • Simplification of processes and cost reduction

By separating the back-end and rewriting the front-end from scratch, we significantly improved Panther Pricing's performance while maintaining a reasonable scope and cost for the project. The modular construction of the proposed front end allows Panther Pricing the flexibility to add new functionalities without compromising the visual consistency of the interface.