Revolutionizing Distributors' Sales Performance With Analytics Platform

LeverX developed a comprehensive sales analytics platform with numerous modules and integrated it with SAP to streamline distributors' operations and help them optimize their performance.

Customer

CPG Data is a US-based company that empowers the beverage industry with real-time, total market visibility. CPG Data provides only the necessary reports and analytics for tracking the major retail metrics.

Challenge

Our team was tasked to deliver a B2B platform, including web and mobile applications, for efficient uploading, storing, and managing sales data collected by sales reps from multiple channels (stores, restaurants, etc.) in a centralized location. The collected data contained information about represented or “displayed” brands and volumes at each point of sale, including tracking menus, shelves, and store displays. 

The solution aimed to provide on-the-go working opportunities and enable managers to access real-time analytical insights.

Over the years, CPG Data collected tremendous amounts of data. In terms of numbers, it has:

• 1M+ physical locations tracked
• 1B+ bottles tracked
• 350K+ cases in stores only
• Thousands of DAU

Due to the data volume, functionality, and client base growth, we have overcome a number of challenges since the project started in 2013:  

• Moving from PoC to MVP and then — to the fully functional B2B platform;
• Data integration with existing customer databases to merge brand portfolio and physical locations’ data;
• A requirement to work offline;
• The requirement to provide the most convenient and fastest solution for sales representatives in the market for report creation;
• A need for prompt answers to data-related questions.

Solution

We developed and extended a sales analytics platform on AWS Cloud in the form of native Android and iOS mobile applications. The platform provides centralized data management, online data sharing capabilities, real-time updates, data analytics and visualization, and other tools to streamline distributors' work. 

Product Key Features

• Incentives Live

CPG Data brings real-time incentive results straight to your phone. Sales reps see the relevant data, the amount they earn on each incentive, and a leaderboard with their peers.

• Ad Tracker Live

The real-time mobile gap-tracking feature Increases the efficiency of the distributors’ ad campaigns to 100%. It's simple; sales representatives just take photos of shop displays, upload ad features, and get instant alerts on gaps (if any).

• Supplier Access

Real-time statistics on how distributors run ads and execute incentives are provided in an aggregated view to suppliers’ mobile devices.

• Offline Support

The feature allows working completely offline while still using features to check in when entering the location, create reports, take photos, and synchronize them with the server when possible.

AI Features

• Smooth Onboarding

Seamless integration between customer selling points and brand portfolio records with CPG Data using AI tools for data cleanup.

• Brands Recognition

Industry-leading AI and Image Recognition perform with 91% accuracy, saving up to 80% of time for sales representatives' report creation.

• Next-best Brand Recommendations

By analyzing historical sales data across similar nearby accounts, the solution suggests brands that can be the next best candidates to add to the existing account portfolio.

• ChatCPG

Emerging Generative AI technology enables a chat-based interface for large data sources, providing quick answers to ad-hoc queries. By leveraging a comprehensive AI pipeline built on Amazon Bedrock, ChatCPG derives robust insights from customer data. 

Technical Landscape

The major stages of the product development included:

• The project was initially launched on Python 2.7, Django 1.6, and the MySQL database. However, with the release of more advanced technologies, these were replaced with Python 3.8, Django 3.0, the PostgreSQL database, Celery for scheduled jobs and task concurrency, and Websockets for progress bar updates.
• ReactJS was chosen as the main frontend library.
• Establishing integrations with multiple data providers, including SAP, 3rd party APIs, and  SFTP servers.
• Setting up DevOps with Kubernetes cluster + Gitlab CI/CD process + Grafana monitoring visualization.
• Using native Android and iOS mobile clients with custom AI models for brand recognition. Built with PyTorch, optimized for Core ML.
• Full offline support for reports’ creation.
• Relying on AWS Bedrock as the LLM provider for ChatCPG

Other Technologies Used

• CloudFront, VPC, EC2, ALB
• AWS ElastiCache
• AWS System Manager Parameter Store
• Amazon Bedrock, Apple Core ML, PyTorch

Result