Making Freight ‘A Seamless Experience’
At BlackBuck, we are connecting Shippers (corporates, SMEs, traders) with Fleet Operators in real-time in more than 300 cities in India.
India has over 8 Million Trucks operating on the road and has one of the least efficient transportation networks where trucks travel less than half of the distance when compared to the trucks travelled in the developed world. Every truck owner/driver and every customer has different preferences and ways of operating their business. This presents unique engineering challenges. We design and develop highly available and scalable systems to solve these seemingly unsolvable problems. We don’t leverage/use technology and are not technology enabled — We are a technology company.
We have two sides of marketplace at BlackBuck — Shippers (who want a truck to take their shipments from city A to city B) and Fleet Operators (who have truck(s) and want to have a regular income from it). We strive for three core value propositions for both sides of our marketplace:
1. Availability — We build machine learning algorithms to ensure real-time availability for both shippers and fleet operators. Timing is a key variable here which makes or breaks the whole equation of the marketplace. Trucks need to be available when shippers need them. Orders need to be available when fleet operators need them. Even a few hours here and there would lead to inefficiencies in the system.
Instead of fleet operators tediously searching through a large set of orders, we show a targeted subset of orders that are most relevant to that fleet operator based on past executed orders, lane preferences, availability of trucks for that day etc. In addition, our matchmaking algorithm takes care of the overall health of the marketplace we are building by encouraging tenets such as fairness (to both shipper and fleet operator), incentivizing “good” behavior while discouraging “bad” behavior.
2. Fair Price — Charging more than a standard price can lead to short term profits but are not long term sustainable as they lead to bad customer experience. Charging lower than a standard price puts unnecessary pressure on company’s ability to scale and grow infinitely. Figuring out a fair price is one of the hardest problems in computer science and is a golden problem to solve for any marketplace. We strive to give a fair price on both sides of the marketplace. In the long run, fairness shown to shippers and fleet operators is the only thing that would matter when it comes to success or failure.
We treat pricing as regression problem and are modelling it around a variety of features such as cost, market, business, and external factors. We also personalize prices down to the level of an individual customer or truck owner on a particular day, for particular source and destination combination and for the material getting shipped. In addition to price calculation, we have a variety of experiments we keep running to fine tune the price delivery mechanism — reverse auction, dynamically varying price, bidding, etc. — based on a number of criteria. Pricing further gets segmented into two machine learning problem spaces — price for a day for a particular fleet operator/shipper and reactive pricing (intra-day price movements) to take care of demand-supply fluctuations throughout a day.
3. Experience — We are obsessed to provide a never-seen-before and seamless experience to our customers (shippers and fleet operators). We are taking on real-time logistics and optimization problems that are among the hardest tackled today by many academic and engineering disciplines, and the tools we build enable us to move more trucks to more places, more efficiently.
Route optimization is one of the many complex engineering problems we are solving to ensure we provide a seamless experience to both fleet operator (by providing a faster and better route) and shipper (by ensuring shorter delivery times).
Data that we get from Google Maps India is primarily collected from 2-wheel and 4-wheel vehicles. Hence, the routes suggested are for such vehicles and not always optimal for a truck. Hence, we are taking on the problem of providing better routing taking truck type into account.
We operate in micro-services model where all services are decoupled from each other and can be deployed independently on their own timelines. Each service owns a logically discrete, non-overlapping piece of functionality. Of course, with such flexibility comes a lot of responsibility in ensuring that we don’t have scenarios that can lead to cascading failures and are able to trace a request’s path through multiple services.
Stay tuned for exciting product and engineering updates on pricing, matchmaking, tracking, route optimization, supply engineering, fulfillment engineering, and growth hacking data engineering teams!
