Our approach to location-tracking

How we tailor location-tracking features for our clients.

The average shipper spends 6 hours a day, 20 days a month, in coordinating deliveries. A lot of this time is spent on calls between the receiver, transport vendor, and the driver, to find out, and communicate location-based data for few important shipments.

Where is the shipment right now? How long will it take? Is it going to be late, and by how long? Is the vehicle making unnecessary stops?

Despite the ‘Information Age’ that we live in, this data is both cumbersome to acquire/communicate, and more often than not, inaccurate. As I observed in an earlier post, for information and assets to flow through a value chain seamlessly, people need to follow a shared workflow, on the same/compatible platforms. With the global logistics ecosystem being as fragmented and multi-tiered as it is, this has been hard to achieve, till now.

Pando solves this problem by powering real-time collaboration and information-flow between all delivery stakeholders through one digital, standardised workflow that can be monitored, controlled and optimised in real-time. This removes complexity, and reduces cost, time, effort and risk. One step in this direction is to get location-based services right.

Since 2014, we have experimented with hundreds of software and hardware products in the market to formulate the right combination for our clients. We have spoken to a few dozen supply-chain managers from businesses of different sizes, to understand what they needed. Through our research, we evolved a nuanced thesis on location tracking.

Pando’s vision is to digitise deliveries of all kinds, for shippers of all sizes. To us, technology in logistics is a democratising agent, and whatever solution we chose had to work for a wide variety of shippers.

First, let’s answer two basic questions:

1. What do different kinds of shippers care about, w.r.t. location-based data?
2. What metrics should a shipper consider before choosing between different competing products?

Warning: This is specific to our clients, but should apply to any traditional company (i.e not e-commerce) in India.

What do Shippers care about?

At a high-level, Shippers want to reduce effort and time spent on coordination, have data to make decisions with, and be notified of any anomalies. Consequently, they seem to care about the following, wrt location:

1. Real-time notifications: Most shippers did not have labour to spare (or care) to monitor all deliveries all the time. Periodic push notifications to everyone who cared would be helpful, and reduce time and effort spent in coordination. Exception management could be built into this, for those shipments that were priority.
2. Historic Location: Often, problems came up in retrospect. Since the way shippers currently operated was not data-based, giving them analysed historic data, on-demand would be useful. For example, a (large, public-listed) paper manufacturer in Tamil Nadu faces an interesting problem — their distributors in North India collude with the drivers delivering their shipments from TN, to sell the rolls of paper en route, making 2x the profits, and disrupting the local markets in those routes. Worse, this was predicted to increase in the GST-regime, where different states have differing prices for the same commodity (since transport costs progressively increase). Usually, this information surfaced months later, when someone reported it. Pando solves this problem by analysing historic location-data for specific routes, and highlighting anomalies. We go a step further to notify all parties involved of any anomaly in real time, playing on their fear — that information is now available. Such historic data is valuable in claims management, bill-settling, vendor rating and ranking etc. as well.
3. ETA: Loading and unloading happened at specific times at different locations — and delays meant demurrage borne by either party. Proactive ETA allowed all stakeholders to plan their activities, and reduce costs.

What metric should shippers consider before choosing a product?

1. Cost/Transaction: For the shipper, this metric is crucial, since it has a direct impact on the total logistics cost. Often, shippers ask for a per-month/per-year cost, which is misleading. The number of deliveries that a product can track per-unit-time, vs. Cost per-unit-time is what is representative.
2. Scalability: Most industries are cyclical between months, with Summer (end/beginning of the Financial year) or Winter (end/beginning of the Calendar year) being the peak/trough season. The product must work for fluctuating deliveries per-unit-time, equally effectively. This metric has a cost component as well — variable costs work better than fixed costs.
3. Intelligence + Accuracy: If the technology needed additional resources to manage, or was not proactive/intelligent/customisable, the purpose (of reducing coordination effort and time) is not achieved. Interestingly, accuracy is not very important in long-haul deliveries, with shippers being okay with margins of error of up to 6–6.5 hours in ETA.
4. Resilience: Long-haul deliveries are rugged environments. Warehouses, Factories that house heavy machinery, fluctuating internet levels on the highways, and very long transits without pitstops (no charging for batteries) is the reality. The product must be resilient to such vagaries.
5. Compatibility and User-experience: Whether a shipper dispatches one, or hundreds of consignments, it shouldn’t matter — the user experience should be seamless across different shipments, for different stakeholders.

Options, and our evaluation

Hardware Devices:

The market has been flooded with GPS-based hardware devices in the last 5–7 years. Most of these contain components imported from China, are programmed and cased in India, and retail at anywhere between INR 1500 to INR 3000 a pop. Several companies rent these out for INR 300–500/month. Wireless and wired options are available, but since most of our clients use market-trucks, wired devices are useless.

When return logistics costs are factored in, wireless devices work out to about INR 150/Transaction (assuming INR 300/M rental, and 5 trips/month/device + Return courier cost). This is prohibitive for smaller truckloads (avg transaction size = INR 5000–7000) but might make sense for higher per-transaction sizes. One device can at-most do about 5 trips of 3 days each (assuming 30 day months, and 3 day return SLAs for the device), significantly restricting scalability, and increasing costs. These devices are fairly accurate through, and intelligent, with proactive notifications, SMS-based tracking, visualisations etc. However, some of this intelligence depends on the quality of hardware used — which pushes costs up, further. Moderate resilience is built-in with most of these devices, that account for data-fluctuations, and extreme physical conditions. Some even come with IP67/68 grade casings — again, at an additional cost.

Let’s be clear, though — hardware needs software i.e. while the hardware has gotten commoditised, the software that runs it (on the cloud) is the differentiator, and the intelligence.

All in all, a good GPS hardware device along with the cloud-based intelligence should cost anywhere between INR 150–300/T. Pando partners with a whole host of GPS hardware providers in India and in the Valley, to provide our clients with hardware-based tracking services.

While some of these devices are very useful for fleet owners, and transporters, our focus was on serving the Shipper — to digitise deliveries of all kinds, for shippers of all sizes. With its relatively high costs/transaction, and low scalability, only a specific kind of (larger) shippers were able to afford these devices: hardware just didn’t cut it deep enough. We had to explore other options that worked for smaller and medium sized shippers as well.

Software-based tracking:

The case for software was very strong. Low per-transaction fixed costs, very high scalability (zero cost of download, ready to plug and play) and accuracy, and high and increasing intelligence. Software Applications are also super resilient — one had to consider only battery and data consumption, and not the resilience of the device itself, since that belonged to the user. Apps also worked with most android phones, ensuring compatibility and a seamless user-experience. Since scalability and resilience were high, at a low per-transaction cost, this seemed like a definite winner!

Around the time when we began creating our own location stack, we met

Kashyap

who was building Hypertrack — a plug-and-play location service for apps like Pando to quickly and cheaply go-to-market. Hypertrack is a great product, built by a great team — both factors equally crucial for startups as well as larger companies. Our satisfaction with the product is not just because it works well for our use cases, and can be quickly deployed; but also because of the inclusive growth mindset of the team.

Today, Pando partners with Hypertrack for the app-based location services for multiple clients.

Reality check

Let’s be clear, though — software needs hardware i.e. smartphones to function. And at 300 Mil, smartphone penetration is far from adequate, in India. Despite this, our view is that app-based tracking is the future, because:

1. Penetration is high, and growing rapidly:

Statistica, 2016

While these ^ numbers paint a decent picture, I think geographical differences also play to our advantage, in the trucking market. Whenever I pitch to a prospective Client who is sceptical, I love conducting an experiment with them: I suggest that we step outside the air-conditioned meeting-room, on to the streets, and check how many truck drivers have smartphones — this number has never been less than 8/10, till today (6/10 with internet connectivity). India doesn’t disappoint.

(Add to this, two of my educated forecasts)

2. As Internet penetration increases (read: Jio), and devices get cheaper (read: Micromax et al), the pace of smartphone penetration will quicken

3. Infrastructure and proactive government policy (read: Digital India, and Modi) will further increase the pace of smartphone penetration

Having said that, our research shows some interesting hurdles to software adoption (read this bit with caution, and don’t extrapolate — the sample size might not be representative):

1. Differences in geography: The southern peninsula (including Maharashtra, Chhattisgarh and Orissa) has significantly higher smartphone penetration than the Northern states (even when you compare two towns of similar economic sizes);
2. Internet connectivity in the North-East is sketchy on average, non-existent at worst;
3. Friction in user behaviour: There is friction in long-haul drivers downloading the application; they have to be incentivised to use their phones for tracking.

Our approach to Location Tracking

To tailor location-based services for our clients, we ask them four questions:

1. What is the minimum % of shipments that you NEED location data for? (i.e. does coordinating a few, specific shipments take up a majority of your time?)
2. How much are you willing to spend/transaction?
3. What degree of accuracy is important?
4. What degree of control do you have on your supply? (i.e. do you own the vehicles, do your vendors own their vehicles, or do your vendors/you source from the open market?)

Hardware makes sense when you need accurate location data for all shipments, and have a sizeable budget. The investment in hardware will pay off if shippers own their own trucks, or have contracted owners who operate only for that shipper.

Software works best when you are comfortable with (not all, but) a growing number of shipments being tracked accurately, are not willing to spend a lot, and can control/influence your supply. Software can also supplement hardware — so your per-transaction costs are reduced, while maximising number of shipments tracked.

Interestingly, there is a third bucket of shippers — who need location data for all shipments (but accuracy is not crucial), don’t have a sizeable budget, can influence their supply, but cannot pass on costs to them. Here, hardware doesn’t work because it’s too expensive, and software doesn’t make sense because it leaves out a significant number of transactions (the effort spent on tracking is equally distributed, and so tracking a few shipments doesn’t reduce the time spent on coordination significantly).

For these shippers, the Pando-Port application periodically urges transporters to find out the location of the truck (offline) and update it via the in-app chat-bot, notifying the shippers and the receiver proactively. You’d be surprised at how well a little reminder works :)

Consequently, our approach to Location Tracking is a combination of three sources, converging to one user-experience:

Watch this space for more thoughts on how we are solving problems at scale, for a complex industry that is low on tech!

Check us out at: pando.in | Write to me at: nitin.jayakrishnan@pando.in