The success of quick commerce brands in India is picking up, especially in the metro and tier-I cities. These services have paved their way into the daily functioning of the super-busy working class and are witnessing abundant demand. In an interview with the Business Today Magazine, Deepinder Goyal (Founder, Zomato) exclaimed, “Blinkit will be larger than Zomato in one-tenth the time. The scale it brings to the table is huge.” Blinkit is expected to start making a profit by mid-2024 and may even surpass Zomato’s food delivery profits by 2030.
For quick commerce brands to optimize their services, they need to plan their network of dark stores in high-demand areas ensuring 10 to 15-minute deliveries.
This case study highlights how we achieved this for another leading quick commerce brand.
Problem Statement
- Estimate demand for potential new store locations.
- Identifying potential markets/areas for new stores.
Methodology
Our solution for the given problem works on two aspects, i.e., identifying location variables/ characteristics that impact the success of a store; and second, building and training a machine learning model that identifies new lookalike locations that replicate success scenarios. Let’s dive deeper into the methodology.
Data Exploration
- We received the brand’s demand data for multiple cities, in the form of anonymized order data and addresses. For every existing dark store, we pulled data for over 3500+ GeoIQ variables across different catchments ranging from 100m to 5 and 10-minute drive times.
- Clustered order data based on value & frequency to identify hotspots of order density across cities.
- Conducted extensive EDA on store revenues and order data to identify patterns in sales data, study variations in demands across cities & markets
Machine Learning
- Basis the cluster of order data, a machine learning model was trained to identify characteristics of the high frequency and high-value locations within the city that were serviced well.
- Based on the model, identified look-alike high-demand areas that were currently either underserved or unserved.
- These potential high-demand areas were clustered and then nearby commercial locations were scanned to identify the best possible locations to serve these areas.
Insights
To ensure the success of new dark stores, we identified specific characteristics that define high-demand areas. These characteristics were determined through a comprehensive analysis of the existing high-demand areas and were used to replicate success in new locations. The analysis and feature selection led us to interesting insights from the demand viewpoint.
- As expected, these areas had a high density of households with income over 10 lacs.
- More than 45% of high-demand areas had a very dense presence of medium & large apartments.
- Another 20% of these high-demand areas had the highest density of hostels in their vicinity indicating high demand from students and the unmarried working population
- These high-demand areas also had a dense presence of dentists, salons & gyms within 500m catchments – indicative of spending happening on wellness & personal care & willingness to spend on convenience.
The presence of these characteristics in a catchment suggests higher demand. Therefore, by replicating the identified characteristics in new markets and cities, we provided a clearer picture of anticipated demand.
This approach ensured that the brand could strategically plan its network of dark stores, optimizing service delivery and meeting customer expectations for quick commerce.
Impact
GeoIQ’s data-backed ML model helped this quick commerce brand to enable faster, and successful expansion of dark stores to optimize delivery time within high-demand catchments. We ensured:
- 4X faster site selection
- 17% increase in avg. order value
- 74% delivery time reduction to the top 40% of most valuable customers
What is RetailIQ?
RetailIQ is a cutting-edge AI solution revolutionizing offline expansion for the retail sector. Built on top of extensive location data and brand data, it provides street-level answers to expansion problems.
It is a powerful tool providing site recommendations to maximize success at the store level and minimize the risk of closure. It also helps identify the total addressable market, conduct competition analysis, better understand your target audience traits and their presence, and perform detailed site analysis and reports, all adding up to a successful expansion strategy.