Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Building the vehicle health data stack | Anup Patil @ Intangles
181 Plays1 year ago
Discover the journey behind building Intangles into a global SaaS business, focusing on mining valuable data from commercial vehicles. Much like an Apple Watch tracks and provides insights into one's body, Intangles helps vehicles in becoming "fitter" by identifying issues that, once resolved, result in substantial cost savings for fleet operators.
For more such interesting founder journeys, subscribe to our newsletter www.founderthesis.com
Read more about Intangles:-
Recommended
Transcript
Introduction and Company Overview
00:00:00
Speaker
Hi everyone, I am Manu Patlin, CEO and co-founder of Entangle Slab.
00:00:17
Speaker
You've surely heard the saying that data is the new oil. And this episode is a masterclass on how to profit from mining and processing of data. Intangles is a company that mines data from commercial vehicles and flags issues that once fixed can lead to significant cost savings for fleet operators. Just like an Apple Watch can show you data about your body and help you be fitter, Intangles helps vehicles to be fitter.
00:00:40
Speaker
In this episode of the Founder Thesis Podcast, your host Akshay Dutt interviews Anup Patil, the founder of Entangles, about his journey of building entangles as a global SaaS business out of India. Stay tuned for insights on how to make gold out of data.
Educational and Entrepreneurial Journey
00:01:03
Speaker
So take me through your career path, like a quick summary of what got you to the door of becoming an entrepreneur.
00:01:14
Speaker
So, I have done my engineering in polymer sciences and from MIT-Pune and then I went on to do master's in material science and engineering from Clemson University, South Carolina, and also applied for PhD in carbon nanotube based analytics in some of the top universities.
00:01:37
Speaker
But then I got the taste of ID world when I came back from US and started loving this entire field where you don't have to wait for work on for five or six months to build a sample and then run a spectroscopy and see your samples have failed. So that's what got me glued to this field and then learned all the tricks and traits of software development and then started my first company in product development.
00:02:00
Speaker
And now I'm here with a new fresh set of team members building IoT products with an interesting flavor of how we are solving some of the most pressing problems in the mobility industry. Prior to Entangles, I was the co-founder of two companies. One was Caspian Solutions that we had founded back in 2006.
00:02:22
Speaker
We had a company for one and a half years, but we had some fundamental structural issues in terms of the way we had set up that company and the kind of vision that we had set. We saw certain shortcomings in that company. What were you selling there? It was a services business where you were doing product services for a travel company based out of US.
Early Business Ventures
00:02:48
Speaker
We saw that there was no glory in it, because all of us were of a product mindset, the co-founders that we got to. So we took a conscious decision that we're not enjoying building the product for somebody else, or building something where things just come to you as a set of instructions, and there is no application of mind as to what you should be doing. So with that, we took a conscious decision. And this was software services.
00:03:19
Speaker
It was purely software services. We ran that for one and a half years. And then we shut the shop over there. And we started Teviska Solutions, primarily into product development, akin in travel domain, because that's what our expertise was. So that was purely into product development, where we build the entire end-to-end product right from search to book, and then the fulfillment of travel products. What kind of product? Product for which industry? I mean, travel is big, right?
00:03:49
Speaker
So the first set of customers that we went into was a multi-level marketing company which allowed customers to have their own branded website of travel bookings and anybody who did travel bookings on their brand website.
00:04:07
Speaker
would get a commission. So the way typically MLM product works, so there's always a downline and then there's a tree. So and the flow of commission is upstream. So that is how the oral structure was. Then the second customer that we acquired, a very large customer that we acquired. Just to understand this better, so this would be like an organization where once you become a member, then you can
00:04:33
Speaker
do transactions like most MLM companies sell something. So here what they were selling, one of those things was travel bookings. And when people buy any of these travel products, flights, airlines, hotels, whatever, then some commission goes to the up line.
00:04:51
Speaker
Yeah, so pretty much you will have a person, MLM, typically you've seen how aggressive that community is. So somebody will come and say, you know what, Akshay, you can have Akshay.travel.com as a website. And let's say if I come to your platform or your travel website and I do any booking, you get a commission as a part of your earning out of that travel product. OK. Yeah. Got it. OK. Very, very interesting. OK. But you're obviously not scalable.
00:05:20
Speaker
Yes, got it. It wasn't. So then we acquired our second customer, a large customer in loyalty redemption space. So you and I use credit cards. And every time we swipe a card, we get certain points. And you can go back onto the website and redeem those points. So the proposition was that you can redeem those points to do travel purchase. And that was our second large customer.
00:05:47
Speaker
So we are servicing one of the largest loyalty redemption companies in the world. And they were doing billions of dollars of booking on our platform. So that was a very interesting product. We scaled the entire thing all the way to the last one I had left Teviska from the management role. It was doing already north of to
00:06:07
Speaker
two to three billion dollars worth of booking that are happening on that platform. So that was the journey that we had at Taviska Solutions and then Entangles happened and that's a very interesting story as to how we landed up doing Entangles. So at Taviska we are partying outside of
00:06:30
Speaker
outside of Pune in one of the resorts and it was evening and I have two kids so that time my younger one was six years old and the younger one was just in the cradle so
Inspiration Behind Entangles
00:06:45
Speaker
It was getting dark and we all were busy chatting with our friends. And suddenly my wife realized that my son is going to be found. And everybody realized that he's nowhere around. So we all started running around looking for him. It took almost 30 minutes of searching all across the resort. One of the problems was that there was no network in that region. So if you had to coordinate, you had to come back to the center point and do the coordination and see if somebody found him.
00:07:15
Speaker
After 30 minutes when we are completely given up and to Frank, you know, those 30 minutes are really one of the worst time frame of our life. So all the bad thoughts that can come to mind actually came. We saw him coming, walking towards us along with one of his friends. They were completely unaware as to how the entire world had turned upside down for us. They're enjoying it at one of the rooms that were playing on one of the iPads and they're just walking towards us.
00:07:44
Speaker
So that's where it hit me. The idea is to start building something where I can start tracking my own kid. And I think that is the problem that a lot of parents would be facing in terms of knowing where the kids are and if they are where they should be at any given point of time.
00:08:02
Speaker
So, that's where I got hold of Jaishri, who I have been known for some time and she comes with a lot of experience in the embedded systems program. Like, I want to build a child tracker. She's like, that's a child's play for me. I can do that fairly easily. So, she started working on that. But as a philosophy that we always had, even in Tabiska and
00:08:26
Speaker
at entangles as well. We generally look aggressively in terms of if this is a market opportunity that we should be chasing. So we saw that there are a lot of regulations, and you're not allowed to carry gadgets in school, and kids are not always with the bag. So you can't put that in the bag as well. So there was no market opportunity. But Jai Shiving, Jai Shave extremely fast in terms of getting things done. She had already built a child tracker.
00:08:52
Speaker
So it got quite interesting. She's like, I'm ready. And I was like, there's no market for the product that you're building. What did it look like? Like something you can pin onto a child's clothes or something? Yeah, it was a tiny device. It had a footprint of or a form factor of roughly around 3 centimeter by 6 centimeter.
00:09:13
Speaker
So it was fairly large. So we're thinking of a size of an iCard that generally employs where to office. So it will be plugged inside, it will be put inside the iCard. That was the original idea. But then we realized kids while playing, they'll keep the iCard aside or they'll put it in the bag and the bag and the kid are not together. So there are so many constraints. So the original purpose was getting lost.
00:09:36
Speaker
So she's like, but I'm ready now. What do you do? So I said, let's do one thing. Let's start tracking vehicles if we can't do, if we can't track cars or if we can't track kids. So we went out in the market and we saw that our device was costing around 5,000 to 6,000 rupees to keep one device on the table. We did a 3D printing around as a casing for the device just to give it a very fine finished look.
00:10:03
Speaker
And we kept it on the table to one of the fleet operators. He's like, your device is costly. I have a device at 800 rupees. I was like, how do you beat that market? You are at like almost 7x or 8x costlier and you don't have a differentiation. So we did a thorough market analysis and we found that there is no market or a problem to be solved.
00:10:26
Speaker
And then I met Neel. So Neel was just moving out of Tabiska, my previous company, because his company was acquired and he was looking for new challenges. I told him, Neel that we have built a product, but I don't see a market for that product.
00:10:39
Speaker
So he's like, I have a junior from my college. He's from PICT Pune. He has done some interesting stuff of getting data from the vehicle's canvas. So I don't know what the data is. He's like, let's talk to him. So we got him on board. We started talking to him. He started showing us a dongle that he was carrying along with him, which connects to the vehicle and used to fire some commands and get data out of the
Technology and Data Collection
00:11:03
Speaker
vehicle. So he's like, if you can get this data, you can see. It connected wirelessly to the vehicle?
00:11:10
Speaker
No, it had an OBD port, so all modern vehicles have something called as a 16-pin serial port connector. Typically, when you take your vehicle to service station, the service guy will connect a laptop to that port to scan the vehicle. So we are also connecting to the same port and collecting some data. And what kind of data, like fuel consumption and what like?
00:11:33
Speaker
It's interesting. All modern vehicles have a lot of onboard sensors and typically they are closed loop systems, which essentially means that based on the feedback that is coming from the sensors, ECU of the vehicle takes a decision in terms of your top demand, power demand, fuel injection. There are all these comes as a part of a feedback to the ECU. What is ECU?
00:11:57
Speaker
ECU is essentially the engine control unit that is making a decision. So when you press an ACPDL, the amount of power demand that is requested by the depression of that ACPDL, that decision is made inside the ECU by a small microchip.
00:12:12
Speaker
Interesting. So we tend to think of the engine responding directly to the pressure on the accelerator. But you're saying that's not really the case. The pressure on the accelerator is one of the inputs which decides how the engine responds.
00:12:29
Speaker
That's why they call it the fly-by-wire vehicles these days because they are not directly linked to the engine but there is a sensor that is sitting underneath your act pedal that is registering in terms of the act pedal depression and that gets translated and then there is additional parameter that comes in terms of in which particular gear you are driving.
00:12:48
Speaker
than the kind of load that you have. So there are a host of other parameters that comes as an input to the ECU. And on the basis of that, you get a certain feedback from the vehicle in terms of speed and acceleration of the vehicle. And what is fly by wire?
00:13:04
Speaker
So they are not directly connected, essentially. So these are all connected over sensors, and these sensors are essentially communicating with each other. So that's what essentially it means. So there's a direct wire that is taking you all the way to the engine. I remember reading this term, I'm into photography, and there's some lens where, you know,
00:13:24
Speaker
you have to adjust the lens focal length. And some of those had this term that the focal length adjustment happens through a fly-by-wire system, which means essentially that it's not responding to the pressure of your hand, but there's a sensor which is deciding about the focal length.
00:13:41
Speaker
Absolutely. Yeah. So pretty much all modern vehicles are operating in a similar way. So we thought that this is a wonderful idea. Let's build a solution where we will build monitoring of vehicles. Now that we have built a tracking, now we'll build a monitoring of the solution.
00:13:59
Speaker
We again started, and this is the core today in Tangles, we question ourselves a lot in terms of are we building the solution for the masses and the right solution. So when we did that, we started pulling out sensors from our own vehicle. That was a funny thing because there were no problems in the vehicle. So how do you create a problem? We started pulling the sensors from the vehicle.
00:14:20
Speaker
You should take it to the service station to fix the problem. And the service guys used to get completely bamboozled. They had never seen that kind of a problem coming from the vehicle. They're like, what are you guys doing with the car? We have never seen this kind of problem. It's like, take the money. Don't ask questions. Just fix the vehicle.
00:14:38
Speaker
It was quite funny. So every week our vehicle used to go to service station because we were pulling the sensors. Finally, we had a solution. But the obvious question was, there was no problem to be solved because we're creating the problem. It wasn't coming out. The problems were not happening by itself. So there is no market. And one other thing we realized is that we don't like to be tracked.
00:15:03
Speaker
You don't want somebody to tell you how to drive and where at what point of time or what your location is. Nobody likes to be tracked. Again, we have the same question. There is no market opportunity.
00:15:18
Speaker
Accidentally, we ended up plugging this device into one of a large 35 ton tipper truck, essentially, where we paid some money to the driver just to say that we have an interesting piece of device. We want to see how it works on a truck. He's like, yeah, here is my truck ticket and see what data you can get. And we found our first fault code without fiddling with the vehicle's instrumentation.
00:15:41
Speaker
and the fuel line was choked up. And it looks like there is some problem over here. But then, of course, by this time, your device would have been a device which can plug into that 16 pin port and transport data. Was that what it was?
00:16:01
Speaker
It was not really a very sophisticated tool. It was just a set of wires that were hanging out of our device. We used to literally plug in into each and every port manually. So there is always the power and ground that is there.
00:16:16
Speaker
And then there is a data that is coming from two of the ports of four wires were there. We knew which ports, which pins of the 16 pin were giving the data. So we used to actually, you know, if you have, and you would have done that in once in your life, you know, when you have a device which doesn't have an end plug, you took a, you take a pencil and then put two wires into the plug. Pretty much you can imagine similar kind of a thing that we have, we had done.
00:16:42
Speaker
Okay, so we got that data and it had some like a sim or something like that to communicate. No, no, we are just we're still collecting offline data, you know, so just firing certain 80 commands as it is called as in the world of collection data collection. So we filed some commands saw how the data looked like.
00:17:01
Speaker
copy that data, take it offline, then read it line by line, character by character. So there are so many alpha numeric characters that you get to read and decipher those byte by byte and then make a sense out of it as to what this data essentially means.
00:17:19
Speaker
then take the data to one of the service guys and say, you know, look, we found this fault. Can you tell us the significance of this fault as to how it is impacting the performance? And they're like, yeah, this vehicle should have a performance dip of at least 25%.
00:17:35
Speaker
And then it's like, how is anybody not acting upon it? And we then took it to the owner of the truck. And he's like, do you know that you're losing 40,000 rupees based on current diesel prices and the number of kilometers you're running? He's like, I have no clue what you're talking about.
Fleet Management and Market Fit
00:17:53
Speaker
And that's why I realized that this is a very interesting area. But we thought that we're lucky. So it was just one truck. And this may not be a larger problem to solve. We plugged into a second vehicle.
00:18:04
Speaker
And you saw that quick before you talked about the second record. So the reason why you were removing sensors from your car was to figure out which sensor is giving what string, like what characters are coming from which sensor, just to track and attribute. That was why you were...
00:18:22
Speaker
Exactly. So generally what happens is in the world of vehicle data, there are two kinds of things that are primarily that are there. There are sensor data, which is talking about how all the sensors are working. And then the second one is called as fault code data. If at all there is any fault code, then there are separate set of strings that are thrown by the vehicle's ECU. So we wanted to capture those fault codes. And hence, we are pulling the sensors out of the vehicle.
00:18:51
Speaker
So generate those fall codes. OK. OK. And how are we able to attribute which sensor is giving what string and stuff like that? Is there some documentation available online or you had to literally experiment and figure out which sensor gives what string?
00:19:09
Speaker
initial days, we actually had to guess because when you're just reading the set of data, the strings, they start repeating every character, they will essentially keep repeating. So you know that now this is how the performance is. But then when you pull the sensor, suddenly a new string appears in that and you now know that this is essentially a fault code string that has started coming up. So that's how we figured that out. But
00:19:35
Speaker
Eventually, when we got deeper into it, we figured out a sophisticated way of analyzing the entire set of data. But yeah, initially it is your right. It was by doing string comparison. And what was the sophisticated way that you figured out?
00:19:51
Speaker
Oh, then that's a close eye. That's where you record to us. In fact, we have reached a level where we have 2,000 power trains now. We have a way to decipher, and we can plug in any vehicle right now from North America to Japan. You give us a vehicle. And we have our own way to decipher all these structured and unstructured data that is there. We have our own mechanism that we have built.
00:20:17
Speaker
Even if it's completely new like if you've never encountered that vehicle before even then you will be able to decipher
00:20:23
Speaker
Absolutely. Every month, probably we are adding at any given point of time, at least 20 to 25 unique new models onto our platform every month. And something that you would have never heard of like Yutong, King Long, you know, like Dragon and all those kinds of, you know, South Asian brands that are there, they suddenly pop up
00:20:49
Speaker
and then you see that they're not there on your platform. In fact, with the regulations that are changing, it might happen that a platform is already there, but then there are regulatory changes that have come and suddenly your devices are no longer compatible and you have to make it compatible by understanding what are the changes that have happened at the OEM level and change the data collection strategy. Even that also happens with us.
00:21:15
Speaker
This is something I want to understand better about the regulation, but let's come back to the second truck. Yeah, so the second vehicle that we plugged in, the RPM sensor had gone bad.
00:21:27
Speaker
Again, the usual exercise, we took it to the service station. We asked him that, tell us how does it impact the performance. Pick up of the vehicle will be hampered. Again, to fix it, you just plug out the RPM sensor, clean it with a clean cloth, and just plug it back in, and then you're done.
00:21:45
Speaker
It's like, is that how simple it is? He's like, yeah, that's how simple it is, the problem. So, what we realized was that, you know, again, this was second truck, and it had some performance issues, although we couldn't quantify in terms of that, at that point of time, how it was impacting the mileage of the vehicle. But then, of course, that was one of the problems. Then we plugged into the third truck, which was an ABS fault.
00:22:09
Speaker
So it's an ABS fault. Anti-brake locking system. So all vehicles, modern vehicles have ABS for effective braking of the vehicle. So that is where we saw the failure of ABS. Now with that as a
00:22:24
Speaker
I think we realized there are a lot of problems. But again, as a part of challenging ourselves, we said, probably these are older trucks, and that's why they're facing problems. Let's plug it into a brand new truck and see how the performance numbers look like or how the overall data looks like. So we plugged into a vehicle. And I'll not name the OEM, but we plugged into the vehicle, which was parked in the yard of a dealership. And that also had a fault in it. And I was like, what's happening over here?
00:22:53
Speaker
So he went to the dealership and he said, there's a truck parked in your backyard, which is ready to be sold. And there's a fault code in it. He's like, yeah, that's fine. It doesn't matter. He's like, how does it don't matter? There's a fault in it. He's like, I'll connect a laptop to it, clean the fault code and sell the truck. But you're selling a faulty truck. He's like, yeah, this is how this industry works. So don't worry about it. It's a minor fault code.
00:23:16
Speaker
So all the co-found started looking at each other's faces, like, what is he talking about? So we came out of that dealership. I was like, look, guys, there is a problem in this industry. And we're doing a lot of reading. And it's a very well-known fact. In India, the cost of logistics is significantly high, to around 13% to 14% of GDP. We now develop countries at around 7% to 8%. And the guys, I now know why the cost is so high. Look at the quality of vehicles that are there on the road.
00:23:45
Speaker
And the infrastructure is improving, but the quality of trucks is not improving and fleet operators are not going to make money if the quality of trucks is going to be this way. So we have to build a solution where they can just monitor and they will come to know that if at all there is a problem in the vehicle and they're losing money.
00:24:06
Speaker
So that was the original premise of starting Entangles. So if you see the journey, so from trying to build a child tracker to vehicle tracking to monitoring system, so that's how we reached here as to what we are doing. And somewhere in the middle, we got our first customer.
00:24:28
Speaker
It is a large fleet operator, bus fleet operator, based out of Pune, called as purple persona mobility solution. And we plugged in our device over there, and we gave them for free of cost, because that's what typically happens when you have no customers, you end up giving devices for free. So we gave them one device for free. And they would be like office pickup drop, something like that, I guess.
00:24:48
Speaker
So they were operating for Pune Municipal Corporation city buses. So they had a fixed route and they had to complete certain number of trips at any given point of time in a given day. And they had to finish the trip also in certain duration and there were serious penalties for that.
00:25:11
Speaker
So we plugged into one vehicle and because we had nothing to do and we were looking at data manually at that point of time, just after 3-4 days of plugging the device to the bus, we saw that the vehicle's engine was overheating. Were you able to do real-time transmission of data by this stage?
00:25:32
Speaker
Oh, yeah, yeah. So that was happening. OK, so we added SIM because as a child tracker that we had built, it already had a SIM card in it. The only sophistication that didn't exist was just a plug and play device. So here also, we had multiple wires that were going inside the OBD port. And then we had taped it to the port so that it doesn't fall off. And the device was just sitting over there in a crude form and sending data.
00:26:00
Speaker
And we saw that one day it fired an engine overheating alert. It didn't fire an alert, but we saw that the graphs were showing certain anomaly. And we called the GM of maintenance, and we said that your bus is overheating. He was like, OK. And he called the driver. The driver said, there is no problem in the bus. The bus is running fine.
00:26:22
Speaker
He called us and he's like, your device is giving you all the wrong data. Go and fix your problem. Stop troubling me. And fair enough, just one device, the chances of you going wrong are significantly higher. So he got back to work. After one hour, he got a call from him. He's like, I want to talk to you guys. He's like, OK. So he went to his office. He's like, you need to tell me how did you find that the bus was overheating?
00:26:47
Speaker
We said, sir, here is the graph. It clearly shows that at a high engine load, your versus the temperatures are off the chart.
00:26:56
Speaker
They're like, look, after we spoke, after one hour after that, the bus broke down on the road. We created a huge traffic jam. And I had to pay a penalty of 5,000 rupee. And then there was additional towing charges. And on top of that, I had to arrange for another bus. But because all our buses were already on road, so I had to get a bus from somebody else. So I had to pay it to somebody else as well. So the cost of doing this was significantly high.
00:27:26
Speaker
But I think you guys got just lucky. So why don't you show me this on 10 vehicles, what your solution can do, and then we'll talk about it. Fair enough. We'll do that because there's no harm in doing it.
00:27:44
Speaker
While we were talking to him, there was a wide board on which there was an interesting organizational directive that was mentioned over there, which said that the entire organization had to improve the fuel efficiency by 2%. Now, if you look at purple person, it's a 2000 plus workforce. The entire organization is required to improve the fuel efficiency. We thought that that's a huge number, 2%. And that also has a significant meaning to the company.
00:28:14
Speaker
So we said, sir, why didn't you do one thing? 2000 headcount, how many buses? They had roughly 200 buses. Okay. So that were operating for UNAMN super corporation. So it was a PPP model under which they are operating.
00:28:30
Speaker
So we said, give us your 10 most or worst performing vehicle if you want to work with those. He was like, here are my 10 worst performing vehicles. And we deployed on those 10 vehicles. And worst performing on the fuel efficiency parameters. On any which, any parameter. It may be because those buses may be breaking down on frequently. Or they may be having mileage issues.
00:28:57
Speaker
or any other issues that might be there. So there may be a pickup loss or the transmission may not be working properly. So all those kinds of things. The driving experience may not be smooth. It may be jerky. So whatever may be there.
00:29:15
Speaker
We took those 10 vehicles. We found that the collective mileage of those vehicles was 2.4. They're all CNG operated vehicles. So we spent 15 days post that on their hub with all the service guys, service and maintenance guys. The funny days they were because
00:29:33
Speaker
You have to spend in the night at the hub. You have to put odomos all over your body because it is filled with mosquitoes. You don't want to get malaria or dengue. So we used to spend the entire night over there. In fact, our family started suspecting, where are these guys going in the middle of a night and coming back early morning? And that was the window.
00:29:55
Speaker
window of repair. All the buses come to hub at 11 PM in the night and 5 AM in the morning, they have to leave. So that was the only window where you used to get to see how these vehicles are being operated upon. So we used to spend all the time over there and see how they are fixing the vehicle if they get a certain set of data. What is the jargon they use for understanding a certain fault code? How did they go about fixing? And there are funny things that we found.
00:30:21
Speaker
You know, the interesting thing that we found was the driver used to come with a fault that my vehicle has lost the pickup. Can you fix it?
00:30:28
Speaker
And the next morning when he'll come to pick up the service guy would say, yeah, I have done something. Why don't you check and come back to me? And interestingly, he would have not done anything. Because he was under the impression that this driver doesn't want to drive. And that's why he's reporting a problem that doesn't exist. So he will only address the problem if he comes back second time, that you did something, but the problem is not fixed. That is only when he'll know that this problem is genuine.
00:30:57
Speaker
So that was their way of understanding if the problem was really genuine or not. And we found that very funny. It was like how this organization is operating. The triage system was that. Yeah. So we understood the way data is being consumed by the service guys. And on the basis of that, we worked with them, started defining how the platform should look like.
00:31:26
Speaker
15 days and post that we went to the GM maintenance and lecture, here is the outcome of the exercise that we did. Your aggregated mileage had jumped from 2.4 to 3.8. And that is way beyond 2% organization mandate that is written behind your chair. And this is what were the fault codes or the issues that we found. And this is how these vehicles are now operating on roads.
00:31:54
Speaker
So he's like, this is good. I want you to come to my head office and I want you to meet the owner of a personal.
00:32:04
Speaker
We went there. One question, how didn't you achieve this 50% jump in fuel efficiency, more than 50% jump in fuel efficiency? So it was fairly easy to do that. There were certain fault codes that were there that had impact on the mileage of the vehicle. There were certain fault codes that were having an impact in terms of the overall pickup of the vehicle.
00:32:26
Speaker
Now, and so we identified all those. And we just gave them the insight or the input to the service guys. And on the basis, they just fixed those issues. That's it. Nothing beyond that. So all they had to do was to know that there is a problem and just work upon it. Now, this is a very interesting question also from a perspective that if you look at it every night, 200 buses check into your hub.
00:32:52
Speaker
And you have limited attention span to go through each and every pass. So the philosophy that we had built was that we will, of course, we are working on 10 vehicles at that point of time. But the philosophy was fairly straightforward, that we have to identify the most problem
00:33:11
Speaker
matic vehicles in the system and bring it to the top. Because that's how you set the priority. So if you're going to set the priority for the service guys, it's like a dream outcome for them. Otherwise, they had to go through each and every bus and they couldn't do justice in the span of six hours. You can't scan 200 buses. Otherwise, you had to have a huge workforce again. So that's the kind of a problem we realized was there. So by giving basic input, they were able to fix these problems.
00:33:40
Speaker
So we went to the headquarters and it was interesting thing for next one hour, we were just sitting quietly over there. So the entire pitch was done by the head of maintenance and we got our first order. That's where we realized the power of predictive analytics rather than just building a remote monitoring system. And we have been building on top of that since then and reached where we are today.
00:34:06
Speaker
OK, amazing. So that was your product market fit moment when you onboarded that first customer. How did you figure out pricing? Oh, we have been experimenting. We did a lot of experimentation on that. And see, one of the initial problem was that we didn't knew that what is the impact that we are bringing
00:34:35
Speaker
Uh, to the table left per vehicle saving those data points were not available to us. So, so we experimented in terms of the overall business model, whether we should go pure OpEx or there should be a mix of CapEx plus SAS subscription. So we tried that with multiple players and, uh,
00:34:55
Speaker
One of the things would mean you don't charge for device, the device is free, you just charge a monthly subscription and capex means you charge a one-time device cost or an installation cost and then a monthly subscription. Yeah, monthly subscription but that is charged upfront on an annual basis.
00:35:17
Speaker
So we scrolled at both the models and we saw that no matter how much of a value you are bringing to the table, recovering money in an OPEX model was really difficult. So we just did that experiment with a couple of fleet operators and we saw that it
00:35:36
Speaker
There were issues in terms of paying it on a regular basis. And we were a capital intensive company because we had to manufacture the device, keep the inventory. And then if your cost of recovery of that is significantly long, then it was not a viable business. Because at some point of time, when you start growing, those were the number of devices you had to keep the inventory. So the rotation of capital wouldn't have been efficient enough.
00:36:03
Speaker
So we realized that we have to bring that down and recover the money upfront so that we can invest that money back into manufacturing more devices. So we quickly devoted towards CapEx plus SaaS subscription.
Advanced Analytics and Target Market
00:36:21
Speaker
And since then we have stick to that model, it has worked quite well for us as well.
00:36:27
Speaker
What do you get in the base tier? You get a dashboard where you can log in. Do you get push notifications when something is going wrong? I mean, I wouldn't understand. Is it install and forget for a customer, or does this need a mature customer who's savvy enough to look at data points and graphs and make judgments?
00:36:52
Speaker
See, you don't need a fleet operator to interpret graph that is done by the system itself. But to answer your first question, they get access to the web dashboard and a mobile app where push notifications keep coming to you whenever there is a certain event that is triggered by the system in terms of either fault code or behavioral issue with the driver.
00:37:14
Speaker
or a certain event around fuel management. So those are the events which are captured by the system and pushed to the owner. There are graphs that are available, but we do not expect them to interpret those graph. There is a section which interprets that graph for them and sets and gives them an inference as to what it essentially means and that the action that they are expected to take on those components that are there as a part of our analysis.
00:37:44
Speaker
Give me some examples, what kind of recommendations would it be giving? So as an example, let's talk about engine temperature anomaly. And then this was a very funny thing that you'll find very interesting. So we saw that a lot of fleet operators, and because India transitioned from BS3 to BS4 and now in BS6, there are a lot of myths that had existed prior to BS4 era. So one of the myths was that if the engine is overheating, you just remove the temperature sensor and let them run.
00:38:12
Speaker
Okay. And it's like, it's a closed loop system. So how does that work? So typically what happens is, and it's a very interesting problem statement, because all engines are designed to operate in a certain performance band or a temperature band, there is a lower and higher level.
00:38:29
Speaker
Now, if suddenly the ECU stops getting the feedback that the temperature, the vehicle has not reached or the engine has not reached the right temperature, what it does is it starts pumping in more fuel to ensure that it is running at that particular temperature. So it keeps pumping in more and more and more fuel. And the mileage of the vehicle goes down because somebody has just plugged out the temperature sensor from the system.
00:38:51
Speaker
So we had to educate the fleet operator. And the initial days was he's like, look, I fixed the problem. I just removed the temperature sensor. And he's like, no, that's not how it works. You have to get the vehicle's cooling system fixed. You can't remove the sensor that is telling you the problem.
00:39:07
Speaker
So what our system did was it simplified the implication of that particular fault code. So the first thing that it did was to catch the attention of the fleet operator that your vehicle's mileage has gone down. That was the first thing that it would use to catch the attention. The moment they realized that just by removing the temperature of my mileage has gone down, it's like now I have to fix this problem. So that is what they did. The second one we used to tell them that now plugging the sensor back into the system
00:39:33
Speaker
And then third, if at all this is not happening, then you have a faulty temperature sensor, even if it is in the system, but it is not giving the right temperature, then you have to just check if the sensor is right. If it is not just replace the sensor, the cost is not that much. So we simplified this entire thing for the fleet operator to, we dumbed it down, essentially. And then we ensured that it can be easily interpreted by any fleet operator.
00:39:59
Speaker
And then we also had an impact of that as to how it is impacting the oral performance of the vehicle and how do you fix or the repair strategies should be. So that is how we bifurcated this entire problems that we are identifying into two categories. And that worked very well for us. So it simplified the entire data consumption from a fleet operator's perspective, although there will be a graph
00:40:23
Speaker
which will show a lot of dots and time series and everything. But we didn't expect it ever for a fleet operator to interpret that. But we did come across some very tech-savvy fleet operators who either had or maintenance guy who had worked with the likes of Cummins and Bosch, and they knew what that data point meant.
00:40:41
Speaker
So that is also something where they'd love to have their geek coming out and every day fixing the vehicle was getting bored. So they used to look at those and give a very interesting inputs to us. So that also worked in our favor to get a feedback from the market. Okay, amazing. I remember Windows used to have this
00:41:03
Speaker
troubleshooting wizard where let's say your speaker is not working and then it'll suggest to use a series of steps and after each one it'll ask you that did it fix the problem or not and based on that give you another solution. So you basically did something similar to that for
00:41:21
Speaker
Oh, it has become quite interesting. So pretty much on a similar line when the vehicle goes for repair, whether it is a regular maintenance or unscheduled repair, as we call it as. Now the guy sitting in his control room has a complete command in terms of if the right work has been done or not. He's able to do the troubleshooting sitting at his desk.
00:41:44
Speaker
And that is where they're having all the fun. They just pick up the phone and call the service guy that, look, my truck came in with five fault codes. It has come out with three fault codes still there. You have not fixed my problem. So I'm not going to pay your bill. First fix the problem. Amazing. Amazing. So that is how the world has changed. In fact, OEMs have taken it to an entirely new level. Now they're able to run real world simulations using the data that we are giving them.
00:42:12
Speaker
And the product development cycle and the feedback has become so fast. And the cost has come down so much for them is that they are absolutely enjoying. They're running their own MATLAB simulations based on the data, the engine data, the component level data. So everybody's geek has started coming out with the kind of data that is there. And now they're having more fun with the kind of data that we have.
00:42:35
Speaker
I'll come to the OEM business, but let me first go a little deeper into the fleet business. What's your customer acquisition strategy? Like how do you acquire customers? Fleet owners, so these are like traditional guys who do dhanda, old school kind of businessmen whom you're targeting. That's your ideal profile of customer.
00:43:00
Speaker
No, so our ideal customer profile is quite interesting. One, we have seen that in the fleet operation business, a lot of young generation or second generation entrepreneurs have started getting in who are tech savvy. They understand the power of data and technology. And they want to not run the business in the conventional way in which how their parents or grandparents were running. They want to do something different.
00:43:30
Speaker
And make it as a model use case in terms of showing off to everybody that, look, this is how I am sitting with you, but I have a full control of my fleet. So that is how our ideal customer profile looks like. Fleet operators who are large, they are tech savvy. And of course, they have a mixed set of fleet.
00:43:54
Speaker
So that is there. And of course, there are a lot of fleet operators who have stringent safety norms, or they have made mileage as one of the KPIs, which they have understood that for every kilometer, if you're not saving on fuel, this is a cutthroat market. So if you can't make money on per ton goods that you're carrying, then let's start saving money on the cost of taking that per ton of goods from a per kilometer.
00:44:24
Speaker
So, they have figured that out. So, those are the kind of ideal customer profile that we are going after. We don't go after customer profile where it's an aging fleet operator or an aging fleet for that matter, which is older fleets. We don't go after them or for that matter, government contracts, we don't go after those as well. So, that's a typical kind of a customer profile that we are going after in India.
00:44:51
Speaker
So, aging fleet means the trucks would not have enough sensors and data for you to add value. That's correct. That's correct. So, there are three classes. And why not government business?
00:45:07
Speaker
See, government business runs in a very different way. There is a lot of dependency in terms of cash flows here. And if the payments are not done on time by the government body to the fleet operators, pretty much your recovery also gets hampered. So that's where we don't typically like to get into that business. And then that's one of the primary reasons. We stay away from that.
00:45:34
Speaker
So this first customer, which was in running the Pune municipality buses, so that was an exception. You no longer chase such customers.
00:45:43
Speaker
We no longer chase such customers, but interestingly, they no longer have a government contract. So now they have become a private customer. Okay, interesting. And what numbers are you at currently? Like how many vehicles are on the platform?
00:46:06
Speaker
We have around 130,000 vehicles on our platform now. That's growing at a massive, massive pace. Every month we are adding at least 10 to 12,000 vehicles. So that is how fast we're growing. And this is just now our international numbers have started tickling in this year, this month. So that is also contributing significantly.
00:46:28
Speaker
We have around 20,000 fleet operators and amongst them this 120,000 to 130,000 vehicles spread is there. How many from India? What's this split? The majority of them is still in India because our international business has just kicked in this quarter and the numbers are still coming in. So literally like 98% of these numbers are from our India business only.
00:46:55
Speaker
Amazing, amazing. I didn't know India had so many fleet operators. Wow, 20,000. And what is the saving per truck that you're able to generate? I'm sure you must be tracking that obsessively because that's the one number which shows your customer what value you are doing.
00:47:14
Speaker
So there are interesting numbers, and that will tell you the value that we are bringing to the table. So one of the fleet operators, which is operating around 600 buses, based on the kind of data inside that we're giving them, they are saving roughly around 20 lakh per month, just behind the 600 vehicle. That is 2.4 karat per year. Wow.
00:47:39
Speaker
There's a very interesting thing that happened. One of the infrastructure companies, they had a fleet of 250 vehicles. Now, one of, again, an interesting problem is that whatever we talk about, it is too good to be true. So typically, whenever you go to suite operator, fleet operator says, you know, just show it to me on my fleet, and then I'll trust you.
00:47:57
Speaker
So we said, give us 50 vehicles and then we'll show you the outcome. So we deployed it on their 50 vehicles, gave them all the walkthrough and how to do the data consumption, trained their team. After two months,
00:48:10
Speaker
Of course, there are frequent touch points that we have created. But after two months, our account manager went for an in-person meeting just to tell them, give them a performance overview as to how things are looking like, how the performance has improved. The interesting thing is that he had kept the check ready for the rest of 200 vehicles. And he's like, what just happened over here? I have not come to pick up the order, but I just wanted to give you the overview of what those 50 vehicles have done.
00:48:36
Speaker
So let me tell you what I have done with your data. On those 50 vehicles, I found out the worst performing vehicles and the best performing vehicles. I have put those best performing vehicles on the tougher routes and the worst performing vehicles on the easier routes. Just on those 50 vehicles, I have already saved 12 lakh rupees per month. So here is a check for 200 vehicles. I want your device to be deployed on all the vehicles that are there.
00:49:00
Speaker
Then there's another fleet operator, which has more than 1000 plus vehicles on our platform. And based on the insight that we are giving them, they're paying more than three curves worth of incentives to drivers every year. So our take is the saving would be, this is just a 10% of the saving that they would be doing.
00:49:23
Speaker
connect the dots and see the kind of saving they're doing. And these are just the numbers on saving, but we put a lot of effort on safety also. And there are some interesting data insights that we give. So if you look at this industry, people have been talking about overspreading, heartbreaking, harsh acceleration. And we saw that there is more to it.
00:49:43
Speaker
in terms of driving behavior. So we started closely looking at the transmission utilization. And we saw when the drivers are driving down the hill in neutral to save on fuel and then take it out and sell in the market. And we realized that that was a myth that was there.
00:49:56
Speaker
So because when you're driving down the hill, it has a serious issue of the engine seizing, your vehicle losing control, and waiting with an accident. So we educated the drivers extensively in terms of why that is a bad driving practice. The outcome of that was that one of the fleet operators, which is based out of Mumbai, they have 600 vehicles.
00:50:17
Speaker
They have brought down these safety incidents by more than 90%. So every day, before the drivers go out, they get hold of the drivers. And the feedback is the way this entire code is, the objective feedback. So we just don't label a driver as a good driver or a bad driver. We tell them that you are failing on these areas. This is where the area of improvement is there. What are some examples of the areas?
00:50:41
Speaker
So typically we call these topics. So we are talking about, you know, the upshift and the downshift of the gear for how long the driver is staying within the peak torque ratio and going above or below the peak torque ratio, driving the vehicle in neutral or doing, you know, hard braking, harsh acceleration that those are also other parts are idling. So these are the different topics on which we give them a scoring and we do a peer to peer comparison.
00:51:05
Speaker
So we will typically compare a Tata Motors driver with Tata Motors and not with Volo as an example and then rationalize the entire score and give them a simple easy to understand scorecard.
00:51:17
Speaker
And we read them on A plus, A two all the way to D minus. And they know that a D minus driver has failed on a certain topic and he needs to fix his driving behavior. And these are the losses that we are incurring because of this particular area where he's not good. So it was fairly easy, simple to comprehend driver's scorecard that any guy could understood.
00:51:41
Speaker
So that is what we gave him and this fleet operator took that. And he just now has one guy whose job is just to get hold of a driver, telling him that, look, this is where your score is bad. You need to improve. And he created an entire driver incentive program around it and started incentivizing the drivers on best performance.
00:52:02
Speaker
Now, the beauty of that was behind 600 vehicles and roughly around 1,500 to 1,800 drivers, everybody is now striving to get those driver incentives. But the collective effort of that is that the safety score or incidents have come down by more than 95%.
00:52:20
Speaker
So, in fact, the graph looked so weird that we had all the problems that were there. They took that, they studied it for a week, and then suddenly the graph came down. It came down by more than 90%, 95%. We thought that there was some problem in our inferencing model that suddenly how did the graph see or saw a dip and then it was a flat line.
00:52:42
Speaker
So then we went to their hub just to see if they were fiddling with the driver. We saw the first thing. There was a driver training that was going on when he reached to their Mumbai hub, a new Mumbai hub. So that's how fleet operators are using the system to take the entire data to the next level. And they're tweaking it based on their operations. And that's the beauty of it.
00:53:09
Speaker
So, these are some of the interesting insights. This would not be in the base tier, right? You told me some of the driver behavior stuff is in the premium tiers. This is on the top tier plan that we have of 6000. That's where it is available. Okay, interesting. And you said it also talks of fuel management events. What does that mean?
00:53:32
Speaker
So we went to the fleet operator, and the fleet operator gave us a very interesting thing. It's like my driver takes off fuel from the vehicle. And can you help me with that? So we said, yeah, we'll try to figure that out.
00:53:48
Speaker
We saw that all OEM vehicles have onboard fuel sensor, which is essentially driving your fuel gauge on the instrument cluster. But it had a serious limitation in terms of the way it had been designed. And our drivers are very, very creative as we saw it. They have multiple ways of steering fuel. So one way of steering fuel will be, let's say, if I'm a fleet operator, I will give driver money to fill 100 liters.
00:54:15
Speaker
He will fill 100 liters, he'll create a chelan or a bill of 100 liters and take it to one of the nearest dhaba and sell that in the black market, let's say 50 liters in the black market. That is one way of making money. And the second way of making money is, given him money for 100 liters, he'll just fill 50 liters, but generate the bill for 100 liters, that is called underfilling.
00:54:38
Speaker
So we found a way to, and this is a patented algorithm that we have, now it's patented in US and India as well, where we figured out a way to take this entire fuel gauge resolution that was around 15 to 20 liters, bring it down all the way to half a liter to a liter by applying our machine learning data models. And on the basis of that, capture all the fuel filling and
00:55:05
Speaker
other instances of selling it in the market or if the fuel is being filled at unauthorized locations like one of the Dhaba because if a driver is selling the fuel to a Dhaba, then the Dhaba guy will adulterate that fuel and then sell it back into the market to some other driver who will fill that fuel.
00:55:21
Speaker
So we started figuring that out. Drivers are smart. They realized that we could figure both the instances out. They came up with a new way of taking fuel out. So they said that, see, these guys are figuring out a way to capture this when the vehicle comes to standstill. We'll not stop the vehicle. We'll just take out the fuel while the vehicle is moving. So that was even more interesting.
00:55:48
Speaker
So we saw that certain vehicles were not making sense to us. And our system was throwing an internal alert that typically the mileage of the vehicle should be, let's say, 4 kmpl. But this vehicle is showing a mileage less than 3 as an example. And we don't see any vehicle health-related problem. The driver is driving right. So there is something else that's wrong.
00:56:11
Speaker
So, we started looking at the fuel curve and we saw that it had a very interesting curvature that was showing up which is the excessive fuel consumption. So, we went back into the field and did a surprise check and we found that they had put a small motor underneath their seat, it was connected to the battery, a pipe was connected to the
00:56:32
Speaker
to the fuel tank. And the other end was connected to one of the, let's say 10 or 20 liters of, you know, plastic jar was there. These guys are so smart. Their calculation is so accurate that they knew that I had to start at point A because I have to sell that at point B. This is the amount of kilometer. This is the time it will take to fill the tank. So this is when I need to start doing that.
00:57:00
Speaker
So all the calculations are done. I was like, if you had done that kind of calculation during your school days, you probably would have been an engineer and not sitting and driving a truck. So that is how they were. So we designed a new algorithm that essentially could find this excessive fuel consumption and raise an alert. Now, this is not an unusual fuel pill fridge.
00:57:24
Speaker
So we raised a different kind of alert over there. That is, this vehicle is having excessive fuel consumption. Although everything else is fine, we need to talk to the driver and tell him that we have figured out that way also, the way you are taking out fuel. So those are the ways in which fuel pilfrage as a feature become a big, big hit. And now it has become a global phenomenon. Everybody loves that feature.
00:57:47
Speaker
Amazing, amazing, amazing. And the way you acquire your customers is through the like you have people who like account executives or business development guys who go out and meet these people and like that's how you're acquiring customers.
00:58:04
Speaker
So generally leads or prospects are generated through multiple channels. So one is, of course, we have an inside sales team that reaches out to these fleet operators. That is one way. And if at all there is an interested fleet operator, then we pass it on to the local regional sales manager who takes that forward. Then the second way in which a fleet operator gets onto our radar is through
00:58:28
Speaker
The vehicle would have been sold as a part of a standard factory fitment by the OEM. And when they see the value of this solution, then there is a moment that comes out of the fleet operator that if you can do this on this class of vehicle, then you can definitely do it on the other class of vehicles as well. So why don't you populate my entire fleet with your devices?
00:58:48
Speaker
and give me a single unified dashboard for monitoring all different makes and model of vehicle. Otherwise, the problem that they have is that if they have Tata Motors, they have to have a separate dashboard. Then there is, if they have Ashok Leland, then they have a Ashok Leland dashboard. And then there is World Wide Share. And then there is Mahindra. And then there is Daimler.
00:59:07
Speaker
So if they have a mixed fleet, then they have to have four different monitors, four different ways in which the data comes to them. And then you need to have people who are able to make sense of the data and bring it onto one single lab.
00:59:21
Speaker
the Excel spreadsheet or presentation to understand what is happening with each and every class of vehicle. We absolutely simplified that, and we said that you have one single dashboard that will do a comparison across the board with all the different make and model. So that was one of the pitch that we had.
00:59:43
Speaker
And the third is, of course, through the enterprise. A lot of enterprises have started partnering with us, and they're reselling the solution as a value-added services to their customers for whatever product they have been using. They are just saying that you have entangles. Now we'll take care of additional set of parameters as well for you while you are consuming our product as an example. What's an example of this? Like a company? So that's an oil and gas company. So that is an oil and gas company which is selling lubricants, fuel.
01:00:13
Speaker
they may have fuel card. So then there may be a tire company and they will say, you know what, you can buy, you use my tires, but I'm going to give you the performance guarantee of the tires using integrals as a solution. So as an example, tire is a very interesting problem statement, right? You build long lasting tires, your mileage takes a dip. And if you build a better mileage tires, your tires life goes down. So you have to have a balancing act between these two.
01:00:39
Speaker
So entangles works very well over there when it gives you a better insight that the mileage of the vehicle has come down, not because of the tire, but because of either the health of the vehicle or the driver is taking out the fuel from the system. If you have time, I'll give a very interesting story that we had. We were working with one of the OEMs, tire OEMs, and they're R&D centered.
01:01:00
Speaker
They had an R&D budget of roughly around 20 karodes per year. And they were coming up with new kinds of tires which would give you better mileage. And then we went there and we deployed our device. And of course, the way they used to test is the truck will go first, go unladen, and then it will have full load, a partial load, then full load. And then they will look at the performance of these vehicles and then decide whether the vehicle is giving a mileage.
01:01:29
Speaker
When we deployed our devices, it was a funny thing. We found that the entire R&D was hinged on that driver who was taking out fuel and selling in the market. And all the R&D guys were just thinking that the tire is not giving the mileage. And they were just pulling their heads and they were trying to figure out what else do I need to do to get better performance.
Data Impact on Research and Development
01:01:51
Speaker
And after almost a year, it just took a month for them to figure out that the tire was not a problem, but the driver who was taking the vehicle for test drive was taking out fuel and selling in the market.
01:02:02
Speaker
So when we took that to the MD of the company, he was completely blown away as to this is how my R&D budget is being spent on trying to figure out the performance of the tire. But so those are the interesting insights that started coming out when the data that we had from the system. You were able to predict when a tire needs a replacement for a fleet owner? No, we are not into that.
01:02:30
Speaker
We haven't yet built a tire module so that we don't have any solution around that as of now. Is that a solvable problem? Is there a way in which you can look at proxy data and give some probability that maybe the tire needs to be looked at? We plan to build a tire module where this will come as an input. And now this will act as an addition to overall performance of a truck or a vehicle.
01:03:00
Speaker
where if the tire has gone under severe wear and tear, then how it has an impact on the overall? There are multiple things. Your braking distance goes down if your tire is bad, or goes up, sorry, it goes up. Then the second problem is your mileage has an impact. So all those aspects do have an impact. In fact, tire pressure is also an important thing. Overinflation of tire is one of the very well-known problems that this industry is facing.
01:03:27
Speaker
If you have an over-inflated tire, then you get better mileage. So drivers are known to inflate the tires, but the tire's life takes a hit. And that is the second single most biggest capex that fleet operators have after fuel. So that is also one of the important things. But we do plan to build a module around it to monitor the performance and the life of the tires. You'll need to add extra sensors, like maybe a pressure gauge on the tire or something like that.
01:03:57
Speaker
No, we don't plan to do that. This will be a little manual effort where the fleet operators are known to take a tire thread depths and the tire rotation and all those parameters. We plan to create a handheld tools that will help them ease the collection of this data. And that will act as an input. And on the basis of that, we'll plug into our current set of algos that we have and make them more smarter. So that is something that we plan to do.
01:04:26
Speaker
Okay, okay. You told me that these, like say a Daimler truck or a Tata truck already comes with a dashboard. This is like the same thing as what you are offering, like they have the same capability in their dashboard, like real-time data. So in theory, all these solutions are supposed to give real-time performance inside on the vehicle.
01:04:55
Speaker
But one of the problems that all these OEMs have is that to what extent they would want to open up their systems to fleet operators to know how serious or how many problems are occurring. And that's where we at Entangles acted as a neutral party.
01:05:14
Speaker
and giving all the data transparently. That is one and second is the unified dashboard. But fleet operators, yeah, all OEMs are working on that. And they're also trying to make the system smarter. But then there is a significant gap that exists at this point of time between what they're offering and versus what Entangles has.
01:05:31
Speaker
So I guess your biggest competitors would be the OEMs themselves, right? Absolutely. They would want to build something which is robust enough that a customer doesn't need to go to you. Although it may not happen, but in their minds, that would be what they would be trying to do, I guess. Yeah. So ideally, yes. So there are two reasons why they are actually a serious competition because they have access to all the data. In fact, they have access to more data than what we have.
01:06:01
Speaker
So, they are definitely that. And second, you're right, they would want to have access to every vehicle on which Entangles gets deployed. They're losing data on those vehicles. And they stop getting insight. So, that is definitely one of the biggest competitors. Why do they lose data?
01:06:22
Speaker
So they will lose data because fleet operator will typically not want to pay two different devices. They would want to have one single device that is solving the problem. This is like a paid service from a Tata or a Daimler. It's not bundled into the truck price. For initial years it is. But after that, they have to renew the subscription by paying some amount. OK. To Tata's or Daimler's.
01:06:51
Speaker
Okay, okay, okay. So there's a device onboard. So that device you replace with the integrals device. So therefore, the OEM doesn't get that data anymore. We don't replace the device because there is a separate can line that is there for the OEM.
01:07:09
Speaker
So we deploy in an aftermarket on an OBD port. So typically two devices can reside parallely. But if they don't renew the subscription for the OEM, then OEMs typically turn it off because otherwise they are incurring cost for a solution that is not being used.
01:07:27
Speaker
And there is a significant cost to keep the device running on a vehicle. There is a network cost. There is an infrastructure cost. So all those costs do exist as a part of the OEMs as well. So they typically don't keep it on. Sir, is there an opportunity for you to tell an OEM that why do you invest in creating your own dashboards? Use ours. We will widely label it for you.
01:07:52
Speaker
There is a significant value to OEMs, to latching on to what Entangles has developed. First of all, there are certain patents that we have filed. Of course, if somebody wants to build a similar kind of a solution, there will be some bit of patent infringement that will happen. The second piece is about
01:08:14
Speaker
The kind of analytics that we have developed, which absolutely makes it easy for a fleet operator to consume the data, we have simplified it. It's a tried and tested model. The third one is for the OEMs, the kind of value their product development and service stations or service team is able to derive.
01:08:33
Speaker
That is also another area which OEMs can build it because, of course, they have access to data, but it will take some time for them to build it. So there is a lot of value that OEMs can derive out of it. But do you see that happening? I mean, wouldn't that? There are conversations happening with the OEMs. And in fact, for the past three, four days, some very large OEMs have reached out to us.
01:09:01
Speaker
And they're getting quite eager to figure out a way to get this as a standard factory fitment. So those conversations are definitely happening. And they're happening across the board, not only with the ICE, but also in EV and other alternative fields as well. So across the board, those conversations have started happening with us. Amazing. That would be like a real game changer in terms of the sudden increase in number of acres on the platform and so on.
01:09:30
Speaker
That's correct. Tell me about your OEM business now. You spoke of what percentage of your revenue is from the OEM business. OEMs are using this for their R&D and so on.
01:09:45
Speaker
See OEMs have multiple lines of data consumptions that they have at this point of time. But to answer your other question in terms of revenue, prior to the start of this financial year, OEMs were contributing around 70% to our revenue. But as our other businesses started growing,
01:10:05
Speaker
That number, of course OEMs pie has increased in terms of revenue share or revenue contribution, but the percentage contribution has come down to around 40%. So our other businesses have taken off significantly, that is the direct market enterprises and our international growth has taken off quite a bit. So now OEMs are contributing around 40% of our overall revenue.
01:10:29
Speaker
I would have thought OEMs would be a much smaller part of your business because an OEM for R&D will not take it for thousands of vehicles, right? Maybe they'll take it for 100 vehicles at best. No, so OEMs are deploying this on each and every vehicle that is getting rolled out from their factory. So this has become a standard factory fitment now. Okay, so you already have these deals in place where OEMs deploy in a wide-labeled manner.
01:10:57
Speaker
Well, yeah, absolutely. In fact, they are not white labeling it. Some of the audience are not at all white labeling it. They're using entangles as is.
01:11:04
Speaker
But they see a lot of value right from get go where the device gets deployed at their end of line assembly. They're doing quality analysis of their vehicles. In fact, there is a mandate that OEMs have started where the vehicles, they can't leave their factory if it doesn't pass as the quality check done on all the components by entangles and they have zero fault codes.
01:11:30
Speaker
So that is the first level. The second is now they have absolute insight in terms of if the vehicle is leaving their factory and reaching a certain dealership network, they know how much amount of fuel is required for the vehicle to reach that particular destination. So they have an absolute good estimation of that. That is a significant saving on that. In fact, they are also now training their drivers to ensure that there is no rash driving that is happening when the vehicles are in transit.
01:12:01
Speaker
The third is now they're able to monitor the residence time of their inventory at each and every dealership. So if there is a demand that is coming from one of the locations and there are vehicles in other location, they can all, of course, swap those vehicles or transfer those vehicles from one dealership to the other dealership.
01:12:20
Speaker
They're also able to post delivery now. They are able to monitor the performance. So the warranty costs have come down significantly for the OEMs from multiple perspectives. One, they now know if the vehicle has been abused. They can see that clearly in the data. And they can decide whether the vehicle needs to be serviced in warranty or out of warranty.
01:12:43
Speaker
And the additional advantage that they have is let's say if I'll give you a classic example, there is an engine overrun condition that we call it as where the there's a RPM spike and that can damage the engine. So typically, these engines are expected to run at least for six years or six last kilometer as an example. But if the engine
01:13:01
Speaker
fails before that, then they can go back and look at the data and they can go and tell also the fleet operator that, look, we had fired an alert, which stated that your driver is causing an engine overrun condition and you need to tell him.
01:13:16
Speaker
but you didn't listen to or you didn't adore to what we are firing. Now we are not going to service this engine in warranty, as an example. And there is going to be a cost associated with it. And typically, a 25 to 30-time engine is decently costly. It can range from 5 to 6 to 7 lakh, depending upon which class of vehicle you're driving. So fleet operators now, so service station guys can show that data and tell the owner or the driver of the vehicle that this is how it is.
01:13:44
Speaker
But let's say the vehicle was driven absolutely fine, then there was no driver behavior or any other kind of overloading condition that was there and the components still failed. Now OEMs have access to that piece of data where they know under what condition and after how many kilometers the component has failed. They give that feedback to the tier one supplier.
01:14:04
Speaker
as a part of their feedback. But they also pass on the warranty cost to the tier one supplier that your component didn't last for the committed drive cycles that it was supposed to last. And you need to own the cost of warranty. So suddenly OEM's warranty costs have come down. Their feedback to tier one suppliers are now in near real time. And they're able to bring about those
01:14:28
Speaker
You know, those improvements at a much rapid rate and the vehicle recalls have gone down. In fact, they have gone down almost to next to nil wherever they have onboard sensors for those components. So that's a huge value that OEMs have started seeing.
01:14:46
Speaker
Inside the OEMs, there are after sales service team, they are now able to monitor the dealer or the service station level performance. If the vehicle goes to a service station and comes out without a problem being fixed, then immediately the command center that they have set up, they know that this particular dealership or the service station is not doing the justice to the vehicle's maintenance, the regular maintenance itself.
01:15:09
Speaker
So those are the kind of insights and they're able to now train the service station guys in terms of the right way of fixing the vehicle and get the vehicle on time back on road for the for the fleet operator. So that's the kind of a value that all different involved parties are able to derive.
01:15:26
Speaker
In fact, we have taken it to the next level as the regulatory changes are coming on board and new regulations are coming on board. Fleet operators have now, OEMs have now have to do changes on the fly, which means they would want to upgrade the entire issue over there. That we have now started doing it on a single click of a button, which otherwise earlier was the owners was on the dealership, get the vehicle back onto the dealership.
01:15:49
Speaker
which meant that there was a downtime for the fleet operator, then the dealership had to deploy a dedicated person just with a laptop to upgrade the ECU. All those things have gone, and just on a click of a button now, the vehicles can be upgraded while the driver is having his lunch or dinner at one of the dabbas. Give me an example, a regulatory change causing a change in the ECU embedded software.
01:16:14
Speaker
So typically, India has gone into phase two of OBD2 norms that are this April, where the emission regulations. What is? Yeah, that's what I'm coming. So what typically it means is that the particulate matters in your exhaust have to be brought down significantly. And that data has to be published over the OBD port or the 16-pin port that we talked about at the beginning of this call. That is where this data needs to be published.
01:16:41
Speaker
So when those kind of changes are to be done, then the entire strategy from the OEM's perspective also has to be done. Now, this is extremely proprietary to the OEM, which is generally not available to the lives of entangles. So what we do is that we have created a secure channel for OEMs to deploy the strategy onto the ECU in terms of how they would want to manage their ancillary emission controllers and meet the regulatory requirements.
01:17:08
Speaker
So that is not impacting the performance of the vehicle. So we have created a secure channel on the basis of which they can upgrade the entire ECU using our data pipe through the device and do these changes. So that's what essentially we're talking about as a part of the entire ECU upgrade over there. Okay, and which OEMs have been using you?
01:17:31
Speaker
So there are three large OEMs at this point of time. They are using this solution as a part of our log. But you can't name them. I can't name them, yeah. OK, OK, OK. Like an Indian OEM, like these Mahindra Tata. Yeah, yeah, all three are, yeah, all three are Indian OEMs. OK, OK, OK. So for an OEM,
01:17:55
Speaker
Is this a value added service to the customer where they are upselling it? Or is it something which is an out of pocket expense for them to help them reduce their warranty costs, et cetera?
01:18:07
Speaker
So the way OEMs see it is that they generally sell the vehicle with a certain AMC commitment. So they see this as a way to bring down their warranty cost. That is how they see it. Post-AMC is over. Then they see this as a value added service feature for the fleet operator. So the initial two to three years for which the OEMs are paying to us
01:18:32
Speaker
They are selling this as a free offering. And that is where they let them have the taste of the entire set of features. And post that, when the fleet operator renews it, they will continue to get access to the data. And on the basis of that, they're able to do the service scheduling, the spare part management, and all the host of other factors as a part of their data, access to the data. OK, OK, OK.
01:19:01
Speaker
OK, interesting. So how does it happen if you go to a fleet operator and he has 100 trucks out of it? Let's say 30 trucks are already indirectly entangled customers because the OEM has provided that. And through the OEM, the entangled system is implemented. The remaining 70 are not. What happens then?
01:19:24
Speaker
So what typically happens, and this is a very interesting question because there are two consumers to this data. There is OEM deployment that has happened, which means OEM wants to have access to the make and model that they have supplied. But from, let's say they have a mix of model. So the moment we deploy,
01:19:44
Speaker
the solution, whenever fleet operator opens, he gets to see all the vehicles, whether it is Tata or Shoklin and Mahindra as an example in their fleet, they get to see all the vehicles on a single platform. But at the same time, in the same platform is opened by the OEMs, they get access to only their brand of vehicles, they don't get to see any other vehicles.
OEM Collaboration and Development Cycles
01:20:04
Speaker
So that's how the entire platform has been designed for all of them to derive value out of it.
01:20:10
Speaker
And your OEM customers get access to all their vehicle's data irrespective of whether that vehicle was sold with a bundled, entangled device or not.
01:20:22
Speaker
No, no, they don't get it. Only when it proceeds. Sold as a part of the bundle solution. Okay, got it. Interesting. You also said that you helped them with the new product development, reducing the time it takes to launch a new model for an OEM. What was that about?
01:20:41
Speaker
Typically, when you're doing a new product development, the most important things are that there are certain performance characteristics that are expected of the vehicle. Those performance characteristics are under different conditions, whether we are talking ambient conditions and the terrain and the driver and all those elements.
01:20:58
Speaker
So for you to have that access to that data, the typical process that was there prior to entangles kind of a company is that there will be an offline vehicle data recording system, or it is called the VDR. Which is like a black box in an airplane, something like that.
01:21:17
Speaker
Yeah, it's like a eight channel multi-can device that is listening to all kind of sensors, storing the data locally and the drivers will drive on the road. You would have seen those kind of test vehicles, those are camouflaged and has weird kind of camouflage strategy that they have. So those vehicles are typically fitted with that kind of a device that is collecting data.
01:21:42
Speaker
And they'll zoom pass through you, by the way. I've never seen them driving slow, for whatever reason.
01:21:48
Speaker
But then that data is collected for thousands and thousands of kilometers. It reaches their R&D center or product development center. They will download the data. And then they will run the simulation, and they'll find the faults or issues with those vehicles. And then that, again, the cycle will happen where they will make those changes. They'll give the feedback to the tier one suppliers. Those changes will happen. And this is a long cycle. So typically, because of that, the development cycle is more than a year for OEMs.
01:22:17
Speaker
Now with entangles in place, the added advantage is that they can change on the fly the entire data collection strategy and they can focus on a certain set of components and accelerate the entire data collection itself. That is the first advantage. The second is it is real time streaming of the data.
01:22:34
Speaker
Third is they can feed the data into their MATLAB system, which means they don't have to wait for the driver to come to office or to their development center. They can ask the driver to run the vehicle under certain conditions if they see certain issues. Let's say the vehicle is going up the hill and they want to repeat that performance characteristic as the driver to do that maneuvering again.
01:22:57
Speaker
So now after the improvements are done or any changes that are done, they can keep running those and keep getting additional data points. So the feedback mechanism is quite fast. And what they can do is they can open up a data pipe to one of the tier one suppliers for them to have access to just their set of pool of data and make a decision on the basis of what is exactly happening.
01:23:17
Speaker
So that's the kind of advantage that they get. That brings down or shrinks the overall development cycle by multifold. It's like when it is real time, you get to run the simulations quite faster. And that's the advantage that OEMs have started taking and bringing about the changes in their overall development cycle. What's MATLAB? You said they transfer the data. MATLAB is a simulation
01:23:41
Speaker
It's a simulation software that is used by product development team. And in that, they can run the permutation commission and understand how each and every component is performing under different conditions. Just a plain simulation software that is used by them. OK,
Funding and Investment Journey
01:23:59
Speaker
OK, OK. Tell me about your funding journey. I'm assuming initial days you must have funded it through the revenue that you were already generating at Tabiska.
01:24:13
Speaker
No, so the initial days was bootstrapped. When we got the initial R&D stages where I told you about the child tracker to the journey of building the monitoring system, it was completely bootstrapped. Then we acquired our first customer. And of course, it was our own personal money. Then we got attention of a few of the H&Is based out of Pune. And they liked the story. And they put initial pool of money. While that was happening,
01:24:41
Speaker
The co-founders of Daviska, they said that we love the story. We would also like to participate. So they also participated as a part of the Angel round that we raised. That was a pool of around 1.7 crores that we had raised as a part of our Angel funding. Then also there was another round of funding again by one of the HNI's. He's known as a quality guru and the only Deming prize winning winner outside of Japan as an individual.
01:25:10
Speaker
So he loved the entire tech stack and he understood the value of all the data inferencing we are doing. So he's like, I would like to invest in Entangles. So his name is Janak Mehta. So he is a veteran in the world of quality. So he invested in Entangles. So that's how we got two rounds of investments done. Then by then, we had a healthy exit out of Tabiska. So I had a lot of personal capital that I had invested.
01:25:41
Speaker
into entangles. Then we also went into the market to institutional investors, but the problem that we faced initially was first we are small and second was it was a little difficult for anyone to understand as to why we are different than any other telematics players that have been there in the industry.
01:26:03
Speaker
And when we start talking about digital twain, the value, it was a lot of times we saw that the analysts were lost as to what is this guy talking about. And the other problem that I also learned as a part of our pitching was that there was too much of technology in our pitch.
01:26:22
Speaker
And I'm sure the question that while the analyst was listening to our pitch was that, how am I going to put this entire thing into my information memorandum that I have to go back and convince my partners that we have to invest? Because this is too much of tech that I will have to write. So we got rejected at the analyst level itself in a lot of these pitches. So we ended up talking to a lot of people. So we didn't have enough success during our initial days.
01:26:51
Speaker
We had a good amount of internal capital, so we were able to survive through that. And then, of course, the recent round of funding happened where we figured out a way to talk about how much of an impact that we are bringing from a fleet of perspectives and why they would latch onto the technology rather than going too much deep into talking about what the technologies are about. So we kept that superficial and we focused more on the value of the entire tech stack.
01:27:19
Speaker
Because you had enough case studies by that time of customers who were actually using and you had enough traction and like you had shown that this is beyond product market fit and there is already customer love. That was one of the things and of course
01:27:40
Speaker
What also had happened is that since India had moved on to BS6 or Euro 6 equivalent, we found that the problems that we are seeing in India is not just specific to India, but that is also something that will be there in the developed economies as well. So that was one of the advantages we had in our pictures when we said, look,
01:28:03
Speaker
We know that US and Europe for that matter has moved on to Euro 6 and those are far sophisticated Indians, but India also has now sophisticated power trains and we are seeing that the complexity of our problem has gone up multi-fold. So this is going to be a problem that you are going to face on a global level. So there is a much bigger dam that is there compared to what we had anticipated earlier.
01:28:28
Speaker
And that was one of the game changers in terms of the way we pitched our product.
Global Expansion and Competition
01:28:35
Speaker
Okay, okay, got it, got it. By the end of this year, what revenue will you close this at? Like, what do you think? So, we will be definitely growing more than 2x this year. So, we would be crossing a very interesting milestone of doing more than 100 crores of revenue this year. And that is again, 2x of what we did last year. So, that's the kind of revenue that we are tracking at this point of time. Which countries have you opened up, past or in India?
01:29:05
Speaker
So we are live in now Canada, where we have already onboarded around four to five customers. Then we are live in US, where we have onboarded one of the very large city operations now. And interestingly, our new customer profile has changed in US. There we have gone live with a fleet which has a government contract, like India.
01:29:26
Speaker
So that's there. Then we are going live in Argentina. That is through our global partnership with the oil and gas company. We are also going live in Turkey and Saudi. We are already live in UAE and we are also live in Malaysia and Philippines.
01:29:46
Speaker
And the other countries that we have seen immense potential, but we have put that on back burner or we are not focusing on this quarter is Australia and New Zealand, Spain and Vietnam. So these are the countries that we will be focusing on the fourth quarter. We want to stabilize these countries that you're talking about at this point of time. Do you have a direct competitor who is also doing this kind of a digital twin?
01:30:14
Speaker
Not digital twin as such, but there are companies which have narrative which is very similar to what we are doing at Entangles in terms of relative analytics. But what we have seen is the level of data accuracy that we have is far higher
01:30:33
Speaker
in case of entangles. So technically, nobody has been able to come close to the kind of insights that we are able to provide. And that is true across the globe, not just in India, but we have seen that outside of India as well. And that's the advantage that we see when we have gone in US or Canadian market. We have displaced some of these top layers already, we saw the fleet operators.
01:31:00
Speaker
So like the traction website mentions Samsara and Lokonav as your competitors. Are these the companies you're talking about which have a similar pitch? These are largely not doing like engine maintenance and stuff. They're more of like fleet management and location and stuff like that, right?
01:31:22
Speaker
See, if you look at the narrative of Samsara, it is, again, same. There is another company called Uptake that is also doing a similar kind of stuff. LocoNav is not, they're focusing primarily on track and trace and on the ancillary services around track and trace. They are not into vehicle health, per se, or fuel management directly. They do, of course, do fuel management by getting an additional sensor onto the vehicle's
01:31:51
Speaker
a vehicle's fuel tank, which we don't do. So that saves not only on cost and time, but also improves on the accuracy, because we are using a robust OE fitted onboard sensor. So that has a far more accuracy. And you're not disturbing the integrity of the fuel tank, which is extremely important for you to have an accuracy on the data.
Operations and Key Learnings
01:32:12
Speaker
But these are the kind of players that we're talking about.
01:32:16
Speaker
There are companies like one of the companies, the turbocharger manufacturing company, that applies digital to him. But their analytics is only limited to the turbochargers that they develop. What's the turbocharger used for?
01:32:36
Speaker
So turbochargers are, they are typically used to boost the pressure of the vehicle. So that is effectively essentially used to improve the overall performance of the vehicle. And they are used in generally diesel powered vehicles to get the best out of the vehicle.
01:32:57
Speaker
What does your supply chain look like? Because you are also selling a physical product. So how have you solved the supply chain problem? Do you manufacture or is it third-party manufacturing?
01:33:11
Speaker
We have an in-house design and development team that designs this entire hardware, which sits onto the vehicle. We source our own components. We are sourced from multiple channels and countries. And during COVID time, we saw how the supply chain can get impacted. And we have seen how a lot of companies that chip
01:33:34
Speaker
Shortage was one of the most buzzing news at that point of time and we were also significantly impacted at that point of time. So the new strategy that we have
01:33:45
Speaker
deployed and we know that you know there's a trade war that is currently going on and there are some kind of embargo, some country would block supply chain from some country. So those can have a significant impact on a company like Entangle. So what we have done is we have multiple designs that are there for a single device and the critical components that are sourced through multiple channels. So if let's say there is a communication modem that we have
01:34:13
Speaker
that is being manufactured in Taiwan, then we also have a design where that particular modem is designed in Italy as an example or source from Europe or it might be sourced from one of the South American countries. So we spend significant amount of time in terms of deciding our design strategy and the kind of component that can sit on to the vehicle. And that is essentially one of the reasons why we are not impacted by the supply chain.
01:34:40
Speaker
And of course, we are still a small player, even with the kind of a volume that we're talking about. It's fairly easy for us to even change our supplier for that matter, if at all we see any constraint that is coming from a country or a particular supplier for a certain component.
01:34:59
Speaker
And it's like the assembly and whatever is done by a third-party manufacturer. Yeah. So the mounting of these components on a PCB is done on a job work by one of the EMS players. And once that is done, those semi-finished goods come to our premises where we have a secret facility, which is where our proprietary code, the one that you have been talking about for quite some time as to what we are doing in the certificate collection, that's a very well-guarded secret.
01:35:30
Speaker
That is pushed onto the device and it is encrypted so that you can't put a sniffer device or try to reverse engineer the firmware. So that is something that is done. So the folks who are loading this firmware don't get access to even that piece of code. All they have is just two buttons. One is load the code and clear the code. That's it. That's how well protected it is.
01:35:59
Speaker
Okay, so my last question to you, what have been your top three learnings in this journey of building entangles as a founder, as a leader?
01:36:10
Speaker
The first and foremost, and that is absolutely important now that this is my third startup. The team that goes behind developing the product has to be absolutely passionate. They should be aligned to the vision. And then you just, when they are aligned, you just empower them and let them run the show and believe in your team. That is one of the things.
01:36:37
Speaker
The second one is you have to be really brutal in terms of understanding the product market fit and understand the value proposition that you're bringing to the table. And you have to keep questioning yourself that you're not starting a company just to be that cool entrepreneur out there, but you're questioning yourself that are you really doing justice, not just to yourself, but to a lot of people who are going to spend time with you and they're going to believe in your story.
01:37:07
Speaker
So that is one of the most important things. And the third is you have to ensure that you are cash flow sensitive. You're not just out there to burn money, but you're always striving to make that entity a profitable entity.
01:37:27
Speaker
whether you become a unicorn or not, and whether you become large enough, that is a problem that will get solved along the way. But these three are things where you are always conscious. And that is how you're developing the entire and the whole whole culture of the company has to be developed in that particular way. And then you'll see that, you know, things do fall in place automatically, you build end up you end up building a product,
01:37:54
Speaker
That is not just specific to a particular region, but you end up building a global product.
Conclusion and Call to Action
01:38:01
Speaker
And that brings us to the end of this conversation. I want to ask you for a favor now. Did you like listening to the show? I'd love to hear your feedback about it. Do you have your own startup ideas? I'd love to hear them. Do you have questions for any of the guests that you heard about in the show? I'd love to get your questions and pass them on to the guests. Write to me at adatthepodium.in. That's adatthepodium.in.