[Product roadmap] A network effect with food delivery led to a pivot for logistics optimisation startup Locus
Bengaluru-basedis an AI-powered decision making platform that automates human decisions in the supply chain, while moving anything from point A to Point B. This intelligence in logistics gives supply chain managers control on-ground operations and improve the experience of the end-user.
Locus offers users a suite that comprises a proprietary route deviation engine, order dispatch automation, field user app, route optimisation, scheduling, tracking for end customer, consulting services, and predictive analytics.
Nishith Rastogi and Geet Garg.
It optimises logistics operations for enterprises across sectors to improve efficiencies, resulting in higher profitability. The engine improves by comparing planning and execution every day, resulting in a strong competitive moat of real-world iterations and proprietary data.
The platform uses deep machine learning and proprietary algorithms to offer smart logistics solutions like route optimisation, real-time tracking, insights, analytics, beat optimisation, vehicle allocation, and utilisation.
But, what stands out in Locus' core system is its solution-based approach, instead of a product-based approach
Geet Garg, Cofounder and CTO of Locus, tells YourStory,
“The AI-based solutions are designed and deployed to solve complex challenges in the supply chain business space. Additionally, a solution-based approach instead of a product-based one allows us swift and robust scaling in deployment."
The drastic pivot
The CEO Nishith and Geet went on to co-found RideSafe, a route deviation detection app for women’s safety. The mobile app was built to make the daily commute safer and detect deviation from the intended travel route in real time. But they were pleasantly surprised when foodtech startups adopted RideSafe to track their native delivery fleets in early 2015.
Geet says, “This gave us material insight that while there were plenty of systems to answer the question of where a shipment was, there wasn’t any system that revealed where it should be. There was clearly a lack of a decision-making system in logistics.”
The epiphany was a pivotal point in building Locus, which aims to be the algorithmic supply chain officer for enterprises worldwide.
Nishith and Geet took RideSafe’s tracking feature, exposed it as the key API, and added a decision-making platform that could give insights to enterprises on which delivery guy should pick up orders from where.
The founders were using the app for the alerts when a deviation from the intended path occurred. Delving a little deeper, they realised a massive opportunity in unifying logistics with automation.
He said, “Although the solution was easy to use and deploy, we realised that there were a number of other underlying challenges in logistics and set out to develop a more advanced decision-making platform. For instance, one of our clients used to manually plan routes the night before dispatching orders. We automated the whole process, saving several man hours and time taken in manual route planning.”
Building the Fireworks Routing Engine
As Locus started working with enterprises to optimise their logistics operations, it was clear that the technology needed to be upgraded to suit diverse logistics demands in supply chains across a number of industries like retail, home services, 3PL, FMCG, and more.
The team built an intelligent order allocation and a dynamic routing solution, trademarked as Fireworks Routing. The solution could be used by its clients to automatically plan routes by factoring in business constraints and order variables like preferred time-slot, vehicle type and real-time factors like traffic conditions while minimizing time and distance on road.
The technology was built in-house, end to end. The need for a bigger team was clear, and Geet and Nishith started looking for data scientists.
“We later built a strong data science team with the right skills and passion, and developed the first version of Fireworks Routing Engine during Diwali break in October 2015,” they say.
While onboarding larger clients, more limitations came into the picture. The customer feedback loop turned out to be pivotal in the evolution of Locus’ products, and helped the company tackle numerous challenges like fairness in allocation, real-world constraints like route restrictions, and restrictions in heavy-vehicle travel.
A major Eureka moment was when Locus partnered with a leading fashion ecommerce player and noticed that it once used to do the sorting of large quantities of shipments manually. This led to the development of a new Locus product, IntelliSort, an AI-backed shipment sorting and rider allocation software.
Growth hacks and challenges
To understand the complexity behind the engineering that happens at Locus, one need to consider the startup's ambitious geocoding project, built on top of a combination of natural language processing (NLP) and machine learning (ML) algorithms, recurrent neural networks (RNN), and different statistical methods.
Using these methods, the team interprets human written addresses and converts them into accessible coordinates that can be understood by a computer.
For a country like India, addresses are error-prone and descriptive rather than formatted. Geocoding is imperative to automate decision-making in logistics. The location data is used to provide intelligence on rider reliability and changing business demands for several enterprise customers.
Geet says, “A growth hack that we unlocked in our tech journey was using managed services from Amazon Web Services (AWS). With the help of AWS, we were able to ensure that we had an automated code deployment pipeline. It also helped us continue our tech operations without putting much effort into an in-house DevOps team in the first few years of our operations.”
Scaling over time to meet the dynamic demands of bigger enterprises was a big challenge initially at Locus. Another key challenge was effective project management and timely delivery, amid complex cross-functional teams.
(Edited by Teja Lele Desai)