Goutam Das December 11, 2019
Inside a store that stocks cement, Mandeep Rati sits across a table, partially in the dark. He is the owner of a truck fleet in Haryana's Bahadurgarh. His 10 trucks are in the business since 2011, when logistics used to be primarily a cash-based affair. To travel between Delhi and Gujarat and return the same route, his drivers carried Rs 35,000 in cash for toll payments, food breaks at dhabas, and other charges. Rati also couldn't tell his customers where the trucks were during the journey. Drivers, at times, did not pick up calls.
Two years ago, a Bangalore-based start-up - BlackBuck - approached Rati. They showed him a 'BOSS' app, through which he could track the truck's trip. His trucks were fitted with GPS devices and he could keep a watch on the number of kilometres the trucks ran in a day. He could also guide the drivers if they get on to a wrong route. Rati also uses the app for diesel payments and FASTag recharges. The tag is affixed to the truck's windscreen and enables automatic deduction of toll charges. It is linked to a prepaid account and employs radio-frequency identification (RFID) technology. Much like any consumer payments app, Rati uses various digital payments modes such as UPI, debit cards and Netbanking to pay through the app. This frees him from worrying about trusting drivers with cash.
Trucking is a low-trust business. Nearby, Jaypal Malik, another fleet owner, is resting at a fuelling station. He owns a bigger fleet of 40 trucks, but is grappling with a bigger problem as well. Some of his drivers steal fuel from his vehicles mid-way during a journey. "I cannot do anything about oil pilferage even when I know it is happening," he says. "There is a shortage of truck drivers and I can't shout at them," he adds.
Fleet owners as well as companies that ship goods grapple with many such inefficiencies that add up to higher costs. India's logistics cost, as a percentage of GDP, is nearly 14 per cent. Most advanced economies spend far lower. In the United States and Europe, for instance, it adds up to about 10 per cent of the GDP.
Gazal Kalra, Co-founder of logistics-tech start-up Rivigo, estimates the cost of inefficiency at about $100 billion. "This is inventory in transit. It is damages, pilferages, wastages on the road," she says. "In the cold chain sector such as pharmaceuticals, dairy, and vegetables, if a transporter switches off the refrigerator, the material can lose its value. It's a broken supply chain. This is why the cost of logistics in India is high," she says.
In answer to these problems, a range of technologies is now converging, which promises to change India's staid logistics sector for good. There is more usage of GPS devices, of course. But companies are also starting to leverage Internet of Things (IoT), cloud technologies, machine learning and Artificial Intelligence (AI) to cut down on inefficiencies that clutter every nook and corner of the supply chain. Using these technologies it is possible to resolve problems like middlemen, inability to track shipments and trucks, pilferage, quality or temperature monitoring and payments.
Innovative Indian start-ups are at the cutting edge of many technologies, building everything from dynamic routing and tracking algorithms to putting together predictive maintenance and AI-based vendor negotiation systems. Yes, in the near future, one wouldn't need humans to negotiate prices either. Algorithms, which can process historical and market data in seconds, can do it far better.
"For every Rs 100 that is spent, we may save Rs 28 on transportation by bringing in technology that helps in efficiency and optimisation," says Kushal Nahata, CEO, FarEye - a Noida-based logistics-tech start-up. No doubt, many start-ups are richly valued. Among India's 30-odd unicorns, or start-ups with a valuation of over a billion dollars, three are logistics-tech companies - Delhivery, BlackBuck, and Rivigo.
It is not difficult to fathom why India's logistics today is this inefficient. Transportation is an extremely fragmented market and doesn't have an entry barrier. Anybody with some capital can buy trucks and become a fleet owner. However, getting business isn't that easy. There is a whole host of intermediaries. Shippers usually contract third-party logistics providers, who connect with brokers for the trucks required. The broker, sometimes a sub-broker, interfaces with fleet owners. Since there are so many layers involved, inefficiencies creep in. Some start-ups are, therefore, focused on technology platforms that can marry demand with supply in more efficient ways. They cut out the middlemen, allowing shippers to directly contact the fleet owner through online marketplaces.
In these platforms, data science plays a big role. The demand peaks at certain times while the supply can peak at a different time. Historical data is typically used to anticipate the demand and the supply before optimisation algorithms can do the matchmaking. Depending on the quality and volume of the data available, the predictive power of the system increases and so does the precision of the algorithms.
The marketplaces are similar to those on the consumer side - over the past five years, Uber and Ola have made matchmaking of supply and demand more efficient for intra-city travel. In logistics, BlackBuck and GoBolt have created a business-to-business marketplace.
"Our goal is to create the world's largest network of trucks. We are creating a freight platform where on one side you have shippers and on the other side, there are several supply partners who get trucks on the platform," says Abhishek Verma, Chief Operating Officer of BlackBuck's Services Platform. "Shippers get better reliability and optimisation of costs. Supply partners get better utilisation. It is a function of reducing idle time by ensuring that there is a big enough stream of orders," he adds.
Shippers on BlackBuck's digital marketplace include both large corporates and small enterprises. The company was founded in 2015 and now claims to have aggregated over 400,000 trucks and over 15,000 shippers.
Gurgaon-based GoBolt has a hybrid model; it both owns trucks and operates a marketplace. "If you are completely asset-light, the serviceability gets impacted because you are dependent on the third party, which is not a very organised market," says Sumit Sharma, Co-founder, GoBolt. "In the hybrid model you keep a bare minimum inventory, but the scale comes from the marketplace," he adds.
GoBolt today has aggregated about 10,000 trucks, but its platform is more of a managed marketplace where algorithms play a role in the matchmaking process. Registered vendors have vendor scorecards - an algorithm scores them on parameters such as capacity to serve, strength in terms of regional lanes (transport is a regional game), historical performance, cost efficiency, and professionalism. When there is a demand from a shipper, the GoBolt system picks five vendors based on their scores to roll out the offer. What happens next is even more interesting. An AI-based vendor negotiation system takes over. "There are no humans required to negotiate. When vendors put up a rate, the system has historical data and market data. It can auto-negotiate by throwing a target price. So, if the vendor says Rs 100, the system can propose Rs 95," Sharma says.
All these lead to cost optimisation for the shipper while the score card model ensures that fleet owners maintain high service levels.
Manufacturing companies in India have to maintain high inventories unlike their counterparts in advanced countries. While many industries in India maintain 20-30 days of inventory in warehouses, in the western world it could be of just five days. The problem is most manufacturing companies in India don't have good visibility into their supply chain. The road infrastructure is unpredictable and so are drivers. "The stakeholder doing majority of the work is the driver. Neither does he own the truck, nor the consignment. He earns Rs 12,000-25,000, depending on his experience and the route. He might stop at his village for two days on-route. These are true scenarios," says Nahata of FarEye. Manufacturing companies, therefore, aren't sure when the next batch of components required to make a product would reach the plant. Holding more inventory, however, is a costly affair.
Technology is solving the on-route visibility problem. Using GPS, IoT and cloud technologies, it is now possible to know the real time location of trucks, track movements, and predict the expected time of arrival (ETA). When trucks don't have GPS devices, the driver's SIM tower location is tracked. At the back end, data from GPS devices is "normalised" - there are different GPS providers, all having different formats. Companies such as FarEye, therefore, have to first break down this data into a fixed format. The cleaned data is next fed into a system with a predictive engine. "Our system has a predictive engine which, based on the source, destination, load, type of truck, number of drivers in a truck, predicts the ETA. We have a fair sense of the standard time between destinations A and B. We also get real time traffic and weather information. Based on this, we are able to arrive at the ETA," Nahata says.
Manufacturing companies apart, supply-chain visibility is even more critical to those operating cold chains.
Dr. Lal PathLabs is one of India's largest diagnostic companies with about 200 laboratories. It has 2,569 patient service centres and 6,426 pickup points from where patient blood and urine samples are dispatched to the labs. These samples are time-sensitive. The quality of the sample can deteriorate within a short span of time. "The entire supply chain has to ensure that every sample reaches the labs within a day. Any delay or loss is a big cost," says Amit Agarwal, GM-Logistics, Dr Lal PathLabs. The company needed better visibility of its supply chain and engaged FarEye. The start-up designed an optimal route that its 1,000 field executives can take to save on time. "The objective was to bring visibility into our systems so that we can track and trace our samples, how much time we are taking to pick a particular sample and hand it over to the labs, or monitor our field executives in a more effective manner," Agarwal says.
The company, he adds, has far better control on its logistics spending right now. "There has also been a significant increase in customer satisfaction. Earlier, the complaints related to on-time pickup for a sample were very high. These delays have reduced," he says.
Dr. Lal PathLabs is now piloting another project. FarEye is integrating temperature sensors in the bags field executives carry. Each sample has to be stored in a defined temperature zone and the sensors would enable effective remote monitoring.
Companies that own a fleet, such as Rivigo, have now converted their trucks into smart vehicles. Its cold chain customers can not only monitor the temperature in the truck real time, they can also set or re-set the temperature remotely. Rivigo trucks have about a dozen sensors. Besides a temperature sensor, there are fuel sensors and those that monitor tyre pressure. Alerts go off if there is fuel pilferage.
Rivigo currently has an asset-heavy model. It owns 3,000 trucks with a premise different from digital marketplaces. Since India lacks infrastructure and service quality is an issue, end-to-end control of the trucking cycle can ensure faster deliveries and higher utilisation.
The start-up banks on a 'relay' model. The journey from destination A to B is divided into four-five shorter hauls through a network of pitstops. Like in a relay race, a new driver takes over every five hours at the pitstops. This ensures less idle time for the truck. "The first driver drives five hours, gets off at the pitstop, and another driver who is already allocated, takes over. The first driver comes back to destination A in another truck that is matched," Kalra of Rivigo says. "We have a patented relay algorithm that allocates and maximises utilisation of both the truck and the driver."
Kalra says Rivigo's trucks take 50-70 per cent of the usual transit time. "If Delhi-Bangalore takes five-eight days, we can deliver in two-three days. We measure our internal turnaround time in hours. Ours is an asset-efficient model. We are working with high returns on capital employed. In a relay model, we are able to run the truck more," she adds.
The relay results in better service and lower cost. Transit times are lower as is the cost because the company sources drivers from non-metro areas. "Driver sourcing is 20-30 per cent cheaper than if I were to hire in cities. In the relay model, I don't need all 3,000 drivers to be in Delhi. I need them across many towns and the cost of living in these places is low," Kalra says.
The driver gets to stay at home and eat home-cooked food since in the relay model, driving is like a day job - the driver goes back home every day. His expenses on eating out in dhabas are much less. In the traditional model, a fatigued driver returns home after a week or 10 days. "In the US, there are strict laws around how many hours drivers can drive on the job. Relay solves this problem because it does not allocate duty before a driver has had the mandatory rest," Kalra stresses.
Algorithms are not just about efficiency and lower costs. They can lead to a better quality of life, too.