Swiggy: Harnessing hyper-personalised approach for efficient food deliveries
Swiggy has 1.7 lakh delivery partners and more than 75000 restaurant partners, across 120 cities. The food delivery platform is harnessing the hyper-local and hyper-personalised experience, as it grows its customer base at an exponential rate
Headquartered in Bangalore, Indian foodtech startup, Swiggy, provides food ordering and delivery solution from nearby restaurants to its customers, across India. “Our vision is to transform Swiggy as an AI first company, which gives a unique way to rethink the business for a company which is powered by AI at its core,” says Dale Vaz, Head of Engineering, Swiggy.
Acknowledging that digitalisation is transforming the ways foodtech companies provide dining experience and do business, Vaz indicates that as the industry is undergoing a technology disruption in terms of embracing new age technologies, and online food ordering is becoming more eminent in India. Swiggy is harnessing the hyper-local and hyper-personalised experience, with its exponential customer growth.
“Swiggy is the unique combination of technology, ML and algorithms which is backed-up by extremely strong business and operation suite. We bring together very carefully artifacted/crafted symphony between technology and people, which is driving the experience powered by the people on the ground, our business, and operations team,” states Vaz, adding that they are matching up the demands in the three-way market place, i.e, customers, restaurant partners, and drivers; and meeting the customer’s demand within the promised delivery time.
The major action happens in a matter of a few minutes. Hence, Swiggy ensures that its customer finds a restaurant, communicates the order to the restaurant and ensures that the driver reaches to pick up the food and gets it delivered to the customer in the most efficient manner.
“With this intricate model, we have underlined various other complexities, i.e., traffic, weather conditions, different levels of challenges with drivers, and various sophistication levels that individual restaurants follow, etc. We assure/ensure to provide highly reliable customer experience within the given supply chain, for every single order” he explains.
Food and customer intelligence
Currently, Swiggy has 1.7 lakh delivery partners, more than 75000 restaurant partners, across 120 cities. The challenge with the hyper-local approach is two-fold, first, help the customers to discover the product and second getting it efficiently delivered.
The food has to be described differently for various regions in India. “Hyper-local discovery has catalogue intelligence which essentially enables us to understand and identify the products we sell on the platform. It collects intelligence about the product and identifies the similarities within the products and ways a customer would relate himself with a specific product,” says Vaz, adding that the second is customer intelligence, every customer has a very precise taste. They closely track the need of the customer and the trends within their buying behavior.
Churning out the data
Swiggy has terabytes of data and has built an in-house data warehouse, where they collect data signals from all types of interactions, i.e, consumers, drivers, the restaurant partners and their interactions.
One of the features on Swiggy’s mobile application is where they provide the customers with a highly accurate promise of delivery time. It is an example where they crunch billions of data points, received on a daily basis. These data points help them build predictions. Predicting which driver can reach the fastest, traffic and weather conditions.
“Technology is at the heart of what we do. Since 2017, we have doubled our restaurant partners and the number of drivers. Hence, with this rapid growth, technology is an existential need and we are using ML at the heart of Swiggy, have replaced human decision-making systems with ML systems. We have doubled the tech headcounts in the last 14 months and doubled our data science and machine learning capacities,” Vaz concludes.