#2 Building the 3rd Party Order Management Platform in Swiggy

The story of building and scaling an API based product

My first role at Swiggy was that of an API Product manager responsible for integrating order management solutions with different 3rd party Point of Sale Solution providers for restaurants. This role at the product community at Swiggy was not considered as a Tier 1 role and was something nobody was willing to take at that point in time where most of the PM’s wanted to focus on building B2C products. I was not left with a choice as this was my first product job and hence had no options to get a different role and so I was fully committed to this.

Problem Statement 

Back in 2017, as the online food delivery business was growing in India there was an everlasting increase in the number of orders which the restaurants had to process in unit time. At peak times a restaurant will get close to 50+ orders in an hour from the combination of Swiggy, Zomato, Uber Eats and Food Panda. In this case the restaurant operator had to open all these 4 apps to see the order details and then punch the order details on their Point of Sale (POS) system which enabled the orders to be sent to the kitchen for preparation. This is a huge pain point for restaurants and if this is not solved it prevents scaling of restaurant food delivery business and affects consumer experience.

Why is solving this for Swiggy was important ?

1. Improving the consumer experience by reducing cases of order cancellations and wrong items sent to consumers

2. Improve the restaurant partner NPS by providing a scalable solution to this problem

Proposed Solution

Integrate swiggy order management API’s with 3rd Party POS systems so that restaurants can automatically receive the orders in their POS systems where they can accept the orders and automatically send it to the kitchen and generate a bill automatically.

Evaluating build vs buy

One way of quickly building this out was to acquire an online order aggregator company like limetray or urbanpiper and use their POS integrations to scale this to different restaurants quickly. We did not go for this option as the moment Swiggy would have bought/invested in any such company other food aggregators such as Uber Eats, Zomato or FoodPanda would have pulled out of the integration and hence the problem for the restaurants would never have been solved. We decided to build our own solution. 

Actual Solution 

Fig : Flow diagram of the 3rd Party Order Management API

How we built the Order Management API Platform ?

#1 Challenge - Building the engineering team to build the order management API Platform

Solution - We had to hire an external consultant who would help in doing the integrations and we built a layer over our existing order management API’s for internal apps to support 3rd Party POS integrations

#2 Challenge - Since we were also integrated with multiple national chains such as Dominos, Faasos etc bringing everyone on a common generic API was critical to scale our offering instead of using custom API’s for different partners

Solution - We had to come up with a generic API which solved most of the use cases of Swiggy and restaurant partners and then by leveraging business and sales team we had to convince the restaurant partners to come to the generic API instead of custom integrations

#3 Challenge - Scaling up the integrations with multiple 3rd party POS players

Solution - In order to scale integrations and meet the demand from the restaurants we published API documentation, test accounts on staging environment and technical support to improve the speed of integration

#4 Challenge - Improving time to launch new features for merchant on the API platform for accepting orders as each new feature launch would require changes on each of the POS player

Solution - The rise of online order aggregators such as Limetray and Urbanpiper which had integrations with POS and had a unified single screen to receive all orders from Swiggy/Zomato helped in solving this problem as these companies scaled massively and feature launches were very fast.

#5 Challenge - Onboarding restaurants on this platform was a huge problem as the Item ID’s of the restaurant had to be manually mapped to the Swiggy Item ID in order to send the order to the 3rd party integration system and was operationally a nightmare

Solution - We solved this via launching Cloud menu which was the API which synced the restaurant menu from the POS system to Swiggy in 2 seconds while earlier that required 2 days to change the menu and was operationally error prone.

Impact 

• This platform started with 2% orders of total Swiggy orders which scaled to 10%+ within a year and 4 years down the line since inception it is close to 35%+

• Faster acceptance of the order by the restaurants leading to 3-4% reduction in delivery time

• Reduction in poor consumer experience as the % of order edits and cancellations were drastically reduced.

• Emergence of an entire ecosystem of restaurant technology companies such as Limetray and urbanpiper

Hindsight view 

Sometimes a charter is not glamorous and is not something that is very valuable at a particular point in time. The point is to use adversity as an opportunity and solve real pain points for our consumers and thus by doing so one learns the different aspects of product management and in due time builds something which is truly valuable after sometime. 

Do Subscribe and comment if you liked the blog