Leveraging analytics for digital transformation in Indian General Trade market

India’s retail market is estimated to be worth USD1,407 billion in 2026 with general trade (or Kirana stores) holding around 80% of share in retail trade market. Consumer Packaged Goods (CPG) firms are at the forefront of adopting technology essential for their business. Leading e-retail firms including Amazon have capitalized on changing consumer behavior by leveraging Big Data insights. This has helped them streamline pricing across channels, discounting and offers, capture untapped markets and make critical decisions. In this blog, we will talk about the challenges faced by CPG firms in their inability to integrate with last mile retail stores for General Trade in India and how can they overcome the challenges by leveraging the analytical solutions.

Current Situation and Challenges for Indian General Trade Market

While CPG firms are adopting technology at rapid pace, General Trade (GT) market has remained a laggard in digital transformation as exposed during Covid-19 pandemic where most of the last mile stores were digitally absent. A research study shows that common challenges faced by last mile stores are 1/ competitive pricing 2/ product variety 3/ customer buying behavior. Lack of access to data from last mile store to understand buying patterns is a key reason these firms are not able to take advantage of technology. The data, when available, can help understand buying patterns, enhance product positioning and improve revenue to product ratio.

Typical situations encountered by CPG firms are categorized as follows:

1. Inaccurate demand forecasting

Inaccurate forecasting is a nightmare for CPG firms. It results in unmet demand and create ripples that disturbs the entire chain and can cause massive issues with supply levels. The key to this lies in sharing your forecasts with your suppliers and collaborating, so you can work together to meet demand. Few CPG firms have adopted auto replenishment mechanism that is based on enablement of advanced data analytics where firms are able to analyze inventory levels and supply situation and trigger an auto replenishment. This helps minimize human intervention and ensures that inventory levels are maintained for customer satisfaction.

2. Improve sales productivity by recommending products (SKU’s) and customer buying behavior

CPG firms need to have a view of Must-Sell SKU(s), Cross-Sell SKU(s) & Out-of-Stock SKU(s) for each kirana (retail) store in order for them to improve sales effectiveness. In order to bring agility, a leading CPG firm in India has built a recommendation engine to forecast sales and predict demand better. The data from sources such as Distributor Management System, market data from Nielsen and demographic data specific to the region was amalgamated for further analysis.

The recommendations on what to sell in a store are pushed to the handheld device that is used by sales agent. Before visiting store, agent can query through handheld device and receive recommendation on which products are likely to run out of stock and need to be replenished. It also informs about other products likely to sell at this outlet.

3. Lack of extensive intelligence for the market expansion strategies

The decision to expand into a new market is strategic and risky. It requires well planned strategy based on market insights. Many CPG firms struggle to compete in General Trade (GT) market and this is primarily due to lack of information from vast ecosystem of retail order history that will help them design strategy to enter GT market. A leading CPG firm have developed 100+ use cases using AI/ML based platform to support market entry decisions. They leveraged data generated from past order history through their retail network to understand buying patterns which helped them design right strategy for market entry into GT segment.

4. Reshaping of the distribution channels by integration of the last mile stores with e-commerce channel to provide seamless customer experience

CPG firms have started realizing the importance of integrating distribution channels to last mile stores. This allows them to have better insights into buying pattern in the last mile stores, be it products (SKU), promotional schemes, design/layout of products and competitive analysis. These insights help firms to improve customer experience as they can directly understand customer’s needs. A leading CPG company designed an ordering system for the last mile stores to place orders directly and these orders were serviced by the distributor or the CPG company itself.

The next section will explain how AWS native analytical solution can help CPG firms in bringing agility to their business operations.

Reference Architecture Overview

The big data analytical solution built on AWS can help address the challenges related to data visibility for primary sales at retail stores. For building use-case specific solution, one can organize the data architecture as a stack of five logical layers, where each layer is composed of multiple purpose-built components that addresses specific use-cases.

The layers in this architecture are as follows:

• Data Sources contains structured, semi-structured or un-structured data. Few examples can be secondary sales data from Distributor Management System.
• Ingestion Layer is used to ingest data into the data layer for storage and further processing
• Storage Layer is responsible for providing durable, scalable, and cost-effective components to store and manage vast quantities of data
• Catalog Layer stores the business and technical metadata about datasets hosted in the storage layer.
• Processing Layer is responsible for transforming data into a consumable state through data validation, cleanup, normalization, transformation, and enrichment.
• Consumption Layer democratizes data consumption across different personas by providing purpose-built AWS services that enable a variety of analytics use cases.

For details, refer to the following link.

AWS Partner Case Study focused on India GT market

Ganit is a data and analytics service company that provides end-end solutions across Data Engineering, Data Migration, Data Analytics and Business Intelligence. It partnered with a leading CPG company (ABC) in India to improve the efficiency and effectiveness of their “Beat Planning and Execution” process.

Situation:

Company ABC sells products in +100K outlets through its distributors. Sales representatives have a monthly beat plan which is designed based on criteria such as visiting certain outlet type at a pre-defined frequency, number of store visits per day/per month.

Sales reps were not able to meet targets, spend right in-store time at the specific stores, which eventually led to low coverage, missed targets and low revenue. The beat plans were static & modelled basis limited set of constrains. Sales reps were not mapped to specific stores and there was uneven effort balancing considering (distance, traffic and revenue coverage) with no contingency plan for ad-hoc holidays.

Solution:

Ganit’s solution created a beat plan for each sales rep along with the route level visualization to ‘maximize in-store interaction’ that led to 100% store coverage at the desired frequency and 20% improvement in the retail face time. The beat plan minimized time spent by the sales reps on the road/traffic. The dedicated territories were assigned for each sales rep to promote familiarity with the region & its stores. Also, the beat plan ensured balanced coverage by the sales reps.

Figure: High level solution flow for Beat Plan creation

To create an optimized sales plan, the solution considered constrains such as store classification, sales rep-store category/store mapping, no. of stores to visit per sales rep, no. of visits per day, total store revenue, travel distance, working days/hours, time to be spent at store, sales rep territory, store visiting window.

The custom solution utilized AWS services with meta-heuristics to ensure that all constraints are met, and the optimized routes can be generated. The solution starts with data engineering to ensure data availability, correctness, and completeness. Store segmentation is generated along with recommended optimized frequency for each store types. Sales rep and store performance is analyzed and ranked to define the right salesman-territory mapping. The optimizer model generated the optimized base plan & contingency plan for each sales rep based on various constraints. Refer to the architecture for this solution:

Figure: High level architecture diagram of solution implementation