Full Transcript

Dhruv Sharma: Hello there listeners, happy Friday, we are streaming live. This is episode 96, today with Siv and I are chatting with Vedant, the very young and bright Vedant, who's building a company called Trak, which is Young India's investing app. Vedant, great to have you on the offline network. Absolutely. Thank you so much for having me.

Utsav Somani: Good to have you with us Vedant, thanks for bringing the young energy to the show. It's a Friday. So let's start off strong. How are Gen Z's investing? I mean, I thought Grow and Zerodha were catering for full stack, but you were 22 and you're building a young Gen Z led app, which is letting people invest. And how are they investing differently from somebody who's say 35 and using Zerodha, Grow and other brokers out there?

Vedant Gupte (Trackk): I think when we had started building Trak, I used to trade and invest in the markets when my 10th standard got over. And I remember chatting with my school friends, people that I know, and they were like, there are so many friction points that they were facing before taking the first trade. And I started speaking to them and started understanding that they want more context before entering an investment or a trade than the previous generation. The second thing that I realized was their investing pattern depends mostly on the goal that they are wanting to achieve. So it is more goal backward investing, where they firstly set up a goal that in the next six months, I want to, let's say, buy this certain shoe. And now what are the stocks that are helping me sort of buy that shoe? So this was a very big behavior that we picked up. And these goals are very different, right? Like for an 18 year old, this goal would be like, how do I save up for a concert ticket? And then they do not know what is actually helping them reach there, because they might have like 5,000 rupees, but the concert ticket that they want is for 15,000 or 10,000 rupees. So what is exactly that would help me sort of reach that goal? And that's when they start calculating key, what are the stocks that they should be looking at? Then they want the context for those stocks. That have they even interacted with this stock in their personal life? Like have they even bought something from that company? And then they think of making the investment decision. So whatever we built at track in terms of the experiences was around this sort of behavior.

Utsav Somani: So you're saying they basically want a personal connect to the company that they're investing in. And we had one of the guests recently on the show, and they said that now in movies as well, people are relating more to the character rather than the story in general. Yes. What are some of the, I mean, how are you branding these features? What are you calling them? How have you designed for them?

Vedant Gupte (Trackk): So the two behaviors that we honestly picked up. So we wanted to solve broking and we built it into three layers. The first layer was the discovery part where what we did was we understood that the user has to grow, go from nowhere to understanding or discovering the first talk or the first option. And the two behaviors that fabric behavior patterns that we picked up was the opinion behavior and the goal-based investing behavior. Under the opinion behavior, we realized that this generation have a lot of opinions about a lot of things. And most of opinions could directly relate to a certain stock that they should be tracking. We started asking these opinions on the platform itself. One of the example would be, let's say, would Durandar be a hit or flop? And on basis of what you answer, we end up showing you a stock with an insight. So let's say you say Durandar would be a hit. We would show you the stock, which is PBR with an insight saying like when Kabir Singh had hit the box office, PBR ended up going 5 to 10% up. Now this is your own opinion that is translated into, let's say, a stock. So these two patterns really, and we branded them as the opinion feature. Like what do you think? And the goal-based investing. So we let the user select their goal. And the goal can be as stupid as, let's say, buying a Shakira concert ticket. And then from there, we help them answer certain set of questions, risk profile the user, and by the end of it, they get to know the stocks that they should be tracking to, let's say, reach the goal.

Dhruv Sharma: Vidhan, do you think that it's important to your set of users to have context in app also because there's a deluge of information out there?

Vedant Gupte (Trackk): Context in terms of stocks?

Dhruv Sharma: Yeah.

Vedant Gupte (Trackk): I think we had tried a lot of these learning things on the platform where we were trying to build a lot of education modules like video-based content, then like textual-based content. But what we started realizing is that this sort of learning is something that they're already getting through YouTube, Instagram, and other different channels. So learning that has to happen on the platform has to be more conversational and has to happen in a much more practical way. Like this opinion thing that I spoke about, initially, what we did was we used to just show the stock on basis of the opinion that they answered. Slowly we started understanding for a user to actually learn out of this that insight is very important. Because of the insight, they are discovering 10 new stocks that they probably would have never heard of just because of the opinion. And new things that we get to know is like FIFA World Cup is happening, right? So would streamers in India start watching FIFA World Cup or would FIFA World Cup become a big thing in India? If you answer yes, you get to know about Zee Entertainment, which recently bought the rights of FIFA. And then the insight is actually helping you learn in a much better way.

Dhruv Sharma: Let's also talk about design as a differentiating factor, right? Because as they say, every generation has had its own and maybe even deserves its own investing app. Maybe you guys are going to be the defining investing app for this generation. So in terms of design, what are the users' expectations? What have they been informed by?

Vedant Gupte (Trackk): I think the users that we are going after is the user base that has got a mobile phone when they were 15-16, has grown up on the internet and very early in their life, they have used world-class products like they have been on Snapchat, they have been on Instagram. So design first thinking or building a very cool sort of a platform is very important for this user base because that is what they resonate to. So it should not feel intimidating. And certain things that we ended up changing on the platform, which also showed up in our metrics was when we were building the execution page, where a user could actually execute the order, we were following the similar theory that any other platforms in India follow. You get your intraday order, you could set your order type, you could set your target stop loss and stuff like that. But what we realized is that on that screen, the user is not trying to solve that. The problem on that screen is that how do I invest 50,000 in the stock? This is the only problem statement. So why don't you build that screen in the same way UPI page is built? Like when I'm G-Paying, I'll be momentarily lost, actually redefining and reimagining that execution page.

Utsav Somani: And how do you think about risk and nudges? Because you're doing AI-led discovery, but you're also tending close to advice as well if you're nudging people to buy a stock related to what you've disposed of as or given out as AI-led discovery sort of information to them. How do you think about risk for this young generation?

Vedant Gupte (Trackk): I think we have tried to solve for risk in a product experience way itself. We were building something for the option buyers, for somebody who wants to do their first option trade, because we felt that there is no platform that exists which is helping for a user who wants to do their first option trade. We built this AI native journey, where you just select the market, you tell the platform if it is going to go up or down and by when, and by the end of it, you get a contract that you should be trading. But what we ended up doing in this entire flow was, one of the questions actually asked you is that, how much do you want to make and how much do you want to lose? What we realized is that before a user enters a trade, if at that time you are asking the user to set up a stop loss and a target, they are high likely to sort of exit in a much better way and lose less capital.

Utsav Somani: I think we have lost Vedant.

Dhruv Sharma: We have. That is the risk you run when you do this live, but I suppose Vedant will be back shortly. You know, one of the questions I was maybe going to ask him himself is, do you have a real chance at employing like a Babel strategy? They do.

Utsav Somani: And I think it's very interesting because I saw the interface and I think the platform looks wonderful. I actually want to ask him the question, how was that first pitch meeting or that convincing meeting that converted Lightspeed and Infoedge to invest in this business? I think that'll be very interesting. But let's do a quick readout of the news today. Anthropic came out very, very strong. Two big announcements from them. The first is the new flagship model, Cloud Opus 4.8, which they're pitching less on raw IQ and more on honesty. I haven't personally used it, but it's four times less likely than the previous version to let its own coding mistakes slide. And the other bigger headline.

Dhruv Sharma: It's also more token hungry.

Utsav Somani: As with this thing. I read the news that some consulting client actually spent 500 million on Cloud because they didn't have a safe limit within the company, 500 million on Cloud in a month. But the bigger headline is they raised 65 billion in fresh funding at 965 billion. So that's just shy of a trillion dollars and pushes Anthropic past OpenAI, which is sitting at 852 to become the most valuable startup in the world.

Dhruv Sharma: So trillion dollar private round like this is bigger than the largest IPO to date.

Utsav Somani: This is bigger than anything that Indian VCs have raised or deployed in the last four years combined. I think 50 billion in the last four years combined. So it's just ridiculous what the number looks like. Vidhan, good to have you back on the show. Absolutely. I don't know, where did we lose you?

Vedant Gupte (Trackk): I was just trying to explain like how we are building the experience where we are letting the user very early on in their sort of investing and trading journey, set up the target and stop loss so that as soon as they keep growing on the platform, it becomes a behavior that they have been asked to do since the very first day.

Utsav Somani: And I mean, so you were running AngelOne Rails before. I think you were tied up with them before. Now you've got your own brokerage license. What does it, I mean, does it change anything for you apart from economics?

Vedant Gupte (Trackk): Actually a lot of things. So for the first three years, we built the entire product on top of AngelOne's APIs. And it's also because a couple of things, right? Like we wanted to really spend time in learning about how broking like actually works like. And second reason was like, we couldn't apply for the broking license because we were in 21. But in terms of what changes is, we actually ended up.

Utsav Somani: I thought there must be some net worth criteria for the company that you were restricted by age.

Vedant Gupte (Trackk): Age, yeah. I mean, there is a net worth criteria. You require like half a million dollars as the net worth while applying for the license. But apart from that, the bigger criteria is like you have to be 21 to apply for one. In terms of what has changed is obviously the economics, like we could manage the funds in a much better way for the users. But the bigger thing is that we have more control over the tech and product. Like we were earlier directly connected to AngelOne's APIs. But now we have more control over how.

Utsav Somani: A very related question to this would be that, are you, I mean, as a platform, are you looking for people to become long term investors or are you looking for your users to become traders? I think you make more money when they trade more.

Vedant Gupte (Trackk): But yeah. So we have two cohorts forming on the platform. In at least at the moment, 95% people who come on the platform are discovering or the entry point is stocks like they're starting with investing first. But again, investing means very different for these users because it is goal based. The goals are usually three, six months long. And the number of orders that they do is not like any other investor's orders, like any other investor on a platform would hold, sit, buy and hold for, let's say a year. But this user is still doing two to three trades a month. Even the investor sort of cohort, because it's a swing term investor. The 5% people, the remaining 5% who come on the platform are starting with option trading directly. What we want to, want the users to actually do is we want them to start with investing first and then we want to keep building multiple asset classes for the same user base as in when they keep growing in their life cycle, because we have caught them very young on the platform. They are 21, 22. This is the first investing slash trading platform that they have seen. And as in when they keep getting older, we want to keep building multi asset classes for them.

Utsav Somani: So you'll grow with them because I'm guessing the ticket sizes are also very small right now.

Vedant Gupte (Trackk): Honestly, not. I mean, when we had started out, we felt that the ticket sizes would be very small. But this is a very interesting insight that we realized at least from tier one cities is that the people who are staying with their parents and who are earning between 5 to sort of 8 lakh rupee income in a year are still deploying around about 40, 45,000 rupees on the platform. And the reason is because you are staying with your parents. So your necessities like food, shelter, clothes and stuff like this is already managed. So you have a very higher disposable income. And if they were not investing, they were spending this money in, let's say, buying stuff, going for events and partying with their friends and stuff like that. But now because of the higher disposable income, they still end up putting in 40 to 45,000 on the platform.

Dhruv Sharma: What are the most promising user acquisition channels like at the moment Vedant?

Vedant Gupte (Trackk): When you're building track, we realized that building a good product is one thing, but we had to build a very funky and a very cool brand for the Gen Z's, right? And we did not want to go the same way like any other platforms are going like the likes of Grow and Upstox and all. So we, instead of doing influencer marketing, we started doing lifestyle marketing on Instagram and started creating content that really started resonating with this user base. So 50 to 60% of our users, we get through Instagram, remaining is all organic and word of mouth.

Dhruv Sharma: Do you somewhere in the back of your mind wonder if the large incumbents also at some point will try this because Vlad Tenev of Robinhood famously, he, in fact, he keeps talking about this great $1 trillion intergenerational asset transfer that is going to take place between the 30s and 40s and how he's positioning the business to sort of, you know, be in the right place at the right time to do that.

Vedant Gupte (Trackk): Yeah. I think, so I feel broken businesses are generational businesses. You can get only two generations on your platform and then you keep growing with the user. So probably 10 years down the line, when we have 50, 20 to 30 million users on the platform from this gen Z user base at that time, it is very difficult for the platform to go backward and start building for the new user base that is coming into the market because your brand is at stake. Your larger user base that you have on the platform is at stake and they are requiring some different solutions. So I do not at the moment feel that a larger player will come into this, but a new player who's starting up can think about this market.

Utsav Somani: Give us some numbers. ET reported 2 lakh signups, 33,000 Demat accounts. What are the numbers that you can share publicly with us?

Vedant Gupte (Trackk): While we were with AngelOne, we had reached a 33,000 account base. We were doing around about a lakh trades per month. And right now we have rolled out the platform in the last month, like our own platform. And we have seen like great bottom funnel metrics from signup to Demat. So just a few numbers, but this is on a very smaller user base. We saw 63% people from signup to convert into a Demat account. From there we saw around about 50% people doing their first trade on the platform, which is 15 to 18x better than anyone in the industry. So we could see people actually coming on the platform, discovering the first stock, taking a decision on basis of that. And we feel by this year end, we would be at anywhere between half a million to a million Demat accounts.

Utsav Somani: Solid. As a final closing question, can you leave us with one Gen Z marketing tip?

Vedant Gupte (Trackk): A Gen Z marketing tip? I think it's very important for the founders to come in front of the camera and actually talk to the users. And I also feel that it's very important for any brand that is building to go out to colleges and do a lot of events over there because you end up also interacting with the users and getting a lot of brand love over there.

Utsav Somani: Alright, wishing you all the best on this journey. Thank you so much Vidant.

Vedant Gupte (Trackk): Thank you so much for having me. Thanks Utsav.

Utsav Somani: Alright listeners, moving on to our next guest. We've got Vinay. Vinay, welcome to the show. Hi Utsav. I have to ask you, how do I pronounce the name of the company? I don't want to get it wrong. See through. Yeah, see through. See through reality. Alright, so you've done a wordplay. Let's explain what your company does and what is Spatial Internet.

Vinay Agastya (Ctruh): Right. So thanks for having me. See through is basically a browser native XR platform. So just to sort of give context, creating 3D virtual augmented or mixed reality experiences today is very hard. The time, effort and money it takes to sort of curate and deploy 3D or XR content today is very, very high, right? Now, the reason being the infrastructure layer, which is the core software which people use today to sort of curate or deploy 3D content or XR content generally remains with the game engines like Unity and Unreal. A lot of IT services companies use these game engines, which are pretty GPU hungry to sort of curate 3D and XR content. Now it does not work on the browser. You need to have good gear in terms of like a gaming PC or a VR headset, pretty expensive ones to sort of consume content in the finest render quality, right? But what we have done is we have taken this to the browser, put it on the browser. All are typically all our 3D and XR experiences work on the browser irrespective of it could be an iOS, Android or Windows. It's sort of device agnostic and also browser agnostic. So SeeThrough basically enables people to create 3D content and XR content with super ease without understanding anything around design code, etc. It's a completely new platform enabling brands and users to sort of curate 3D and XR content on the browser and deploy it on the browser, right? That's what SeeThrough does.

Utsav Somani: And give us a live example, maybe I think that'll help us understand how does it impact the end customer? Suppose I'm a D2C brand, I have 500 SKUs, flat product pages, I come to you, SeeThrough, what happens after that?

Vinay Agastya (Ctruh): So imagine this like Shopify 3D on steroids, right? Once you integrate your catalog into the SeeThrough's ecosystem, we basically take a flat images and videos and start creating 3D assets of your products. It could be a shoe brand or a watch brand or a jewelry brand, right? So once it gets imported into SeeThrough's platform, we have something called WordSci. It's a foundational model of SeeThrough, which basically takes text, image and videos as input and converts them into 3D assets. Now, once the 3D asset is ready, that's the opening into the XR world. Now you can enable a virtual try-on, you can enable an AR placement if it's a physical, let's say, if it is a sneaker brand or if it is an internal, you know, interiors brand, right? All of that, once the 3D asset is created, now your PDP changes. The way people discover your products would be a 3D asset running smoothly on your browser within the product page. And then you can enable try-ons in a matter of minutes, like probably 30 to 60 minutes. Once you sort of do the onboarding onto the platform, it enables brands to sort of directly enable virtual try- ons or 3D asset classes, sort of change from a simple image video to a completely 3D assets, which people can try and make a conscious purchase decision.

Dhruv Sharma: Vinit, this world of like immersive technologies, and you can argue this in any way, has had a lot of false starts, right? Like, I mean, even Zuckerberg bet the farm on the metaverse and somehow it's always five years away. Along the way, what have you learned about how you actually, how can you make this technology work with practical real-world use cases?

Vinay Agastya (Ctruh): Perfect. Great question. So I've been intrigued with this tech for almost, let's say, since 2018, I was working at a company called Skapik. It got acquired by the Flipkart Walmart group. So pretty interesting folks got together. We started creating a 3D editor on the browser and enable AR easily. Just as a disclosure, I was an investor in Skapik. Oh, nice. I know the guys very well. Sai and Ajay, right? Yeah. So that's when I started sort of getting intrigued with this tech and started learning where you'll have seen the different phases of XR and 3D, right? One thing is very clear. What has not changed is the infra-level layer, which as I told you, right? Even today, creating 3D and XR content is hard because A, because you're heavily reliant on these game engines, and it has become a very service-heavy sort of industry, right? Now, if you see a lot of very large creative brands to IT services, James, all of them use these softwares. But the entire infra or the dev tool there has not seen innovation, right? That's where SeeThrough kicked in. We created a browser-based core 3D engine, which enables you to run, you know, a 900 MB file, a beautiful entire XR walkthrough directly on the browser, right? As I told you, the main problem in this industry was the time, effort, and money it takes to sort of curate 3D and XR content. Now, with SeeThrough, what changes is we have re-engineered the entire infra layer, where we have our own 3D engine and an AI tool like Versa AI, which people do not even have to know what's 3D design for that matter, right? You can simply plug in as a brand, vertical use cases in real estate, e-commerce, automotive, aviation, healthcare, right? Across all these industries, the core infra layer is SeeThrough's core proprietary 3D engine and Versa AI. So what changed? We have seen a lot of, and Meta did a lot of things around virtual reality. They said Metaverse is the next big thing. The core problem there was render quality was always funny, right? It looked cartoonish. It was not real enterprise-ready, or brands cannot simply pick and start using Metaverse because it was all gimmicky, and it was just dumb for marketing, to be honest, right? Now, if you look at the infra bed or the infra layer here, with the 3D engine and Versa AI coming together are two core pillars. Now I'm making it as simple as people can just start prompting Versa AI and get 3D assets out. And so that's the fundamental change through what has happened over the years.

Utsav Somani: And you're promising 4K to 16K 3D experiences all done within the browser. So I know you said patented technology is the key insight or the key powering engine behind this. But I mean, share a little bit more about this tech, because I think that seems to be the mode that you're bringing to the market.

Vinay Agastya (Ctruh): Correct. So let's say running a simple experience, let's say, can we play PUBG on the browser? Not possible because it's a very heavy file and you need to install, need to have a great graphical power to sort of compute and then sort of render the beautiful graphics you see in the game, right? Now, flip side of it, why can't I get a 4K render on the browser? What happens is, Siru's 3D engine basically has a 3D asset caching mechanism patent. What it does is basically it's like an offline game once it loads within 5 to 15 seconds of the experience. So once the experience is loaded, this has a great level of detail. For example, there's a POV you're standing here and there's a real estate walkthrough, there's a bedroom, there's a couch in the hall, etc. Now, if you're going nearer and nearer to that particular asset, we sort of render that in 4K and 8K because these are textures being directly on the 3D engine. What it does is basically gives you that 4K real throughput, like the render quality, it looks photorealistic of the product or any other experience you see on the platform today. Now, we have another technique called SVD, Sequential Votering Texturing Technique, where the way we look at pixels in 2D, we have 3D meshes in the 3D world or the AR world. These meshes are computed and are sort of compressed, multiple levels of compression is taken and then we cover it with a texture like a notebook cover used to do. Now, that texture is finished with 4K and 8K and that gives you the ability to load it faster on the browser and sort of enjoy the render quality at the same time. So there's been a lot of engineering under the hood to build this 3D engine to sort of help brands put it on the browser because they have to put it on their website. Nobody's going to install apps and sort of enjoy any AR or VR experience of the brand, except, I mean, if it is on the browser, it works everywhere. That's the fundamental core technology which we have built for the 3D engine. And Versa comes in because we want people to create 3D content easily. As simple as the way you use plot today, you can use Versa to sort of generate 3D assets and content.

Dhruv Sharma: Vinay, the one interesting thing about your industry is that you guys have very specific grammar and vocabulary, but can you, like in the simplest terms possible, help us understand how a user can truly feel like they're immersed in an experience or in a frame? How is being immersed different from being present? For instance, if I call you on the phone, I'm certainly not immersed, but it'll still feel like you're present. I can hear you and so on and so forth. And so just, yeah.

Vinay Agastya (Ctruh): All right, so let's sort of break this down, right? Immersive is, let's say, if you're in a virtual environment, it's called virtual reality. And now let's say if you're in a mixed reality sort of environment, which is your Oculus Go or Quest, which you have tried, right? So mixed reality, basically, you have your entire surroundings the way they are, and then there's virtual content being put into it, right? Now, I'm talking about the first level, which is 360 and 3D, where from a photo, you're moving to a 3D asset. And now, being immersive comes when you're on a hardware device, which is letting you immerse in that particular experience, right? The degrees of freedom would be 3DOF, 6DOF, where you can interact with it in different, like just the way spatial internet is going to work in the years to come. You can actually use your bare hands to sort of interact in that particular, you know, in that environment. Now, being immersed is, I'm talking about the first year of creating 3D content. I mean, it has to be democratized. There's very less 3D or 360 content available on the internet today. The second part of it is, once you go into that virtual or mixed reality environment, you're actually moving away from the current chaos of 360 and 3D, and you're moving into a completely virtual environment. Now, that's being completely immersed. Mixed reality is going to be the new normal thing. You've seen Google releasing the, you know, spectacles, the XR glasses. Now, what happens is you just sit in front of your system, or probably you're trying to get immersed in the laptop. Now, the form factor of the device is going to move where you're just wearing some certain spectacles, and there's virtual elements coming in front of you, and that's 360 and 3D content.

Utsav Somani: So, the form factor- I think Meta did it with Meta, with Ray-Ban, right? I believe the last-gen model that came out, you can see stuff via XR.

Dhruv Sharma: And I think that technology seems to be working. Like, they also call it, I think, glanceable overlay. We've had it with heads-up displays for like pilots, even Formula One drivers, and that way your perception of depth, et cetera, doesn't change. So, that's one end of this, Vinay. The other end of this is, we're talking about augmented reality, and how would alternate reality, a complete fictional fantasy world, like, I don't know, the Dune landscapes.

Utsav Somani: Do you think if we don't have a reference point- Dhruv is living in the AI world, that's why his T- shirt says AI native.

Dhruv Sharma: I know.

Utsav Somani: Elon Musk was saying that you can actually have full generative worlds as well now, in a few years, maybe.

Vinay Agastya (Ctruh): So, now, there are a lot of companies in the US, I think World Labs is one of them, who's basically where you just upload an image, and they create an entire walkthrough of that particular facility, right? Now, these are real worlds, then you can create a digital twin or a Fiji twin of the same thing, of your facility, or then alternate reality, for sure, Dhruv. I mean, Meta tried it with Metaverse, didn't work properly, but then there could be shoppable worlds where brands can interact in that particular layer, and sort of immerse themselves into a great shopping experience, or it could be a fair, it could be a virtual mall. I mean, a lot of people are already thinking in these lines.

Dhruv Sharma: So, Sitru sits at the- Yeah, because your approach seems much better than all of the others, right? It's like, you're futuristic, but just enough, you know, you're still kind of overlaying on the flat rectangle and not asking people to put these clumsy headsets on.

Utsav Somani: But maybe talk to us about the impact on commerce as well. I think that will help us understand the impact of this. Like, I think what Dhruv was saying, with near term, very near future. So, maybe some brands that you work with, what is the uplift in commerce sales or anything? I mean, customer conversions, basically, for them.

Vinay Agastya (Ctruh): Right. So, let me give you a quick, while I cannot take a lot of brand names, we have close to 32 enterprise clients in India, US, and the Middle East, right? Now, there's a shoe brand, which we work with, a couple of them. There's a jewelry brand, which we work with, right? Now, what people have seen, the virtual try- on or the 3D asset. First, there's an AI sizing guide where the user opens up the camera and sort of sizes on their bare feet. And then the second class is, the entire catalog is 3D enabled. They can simply try on the shoe and then sort of make a conscious purchase decision. So, we did a small workaround for this. We said, why don't we target abandoned carts, right? The lot of abandoned carts for all these brands. Then we do a WhatsApp nudge with a link to try on that particular things, which are in their cart. Now, the conversion is more than 70%, 70 to 72% for most of the brands, especially the abandoned cart campaign. And then slowly now, they're navigating to a phase where they do away with the physical photos, videos, et cetera, which is mostly AI generated these days, right? And then sort of let them take it for a spin to do it in 3D and even try on their particular product directly on their website. So, the first stage is enable their catalog, make it 3D, and then sort of run certain campaigns like abandoned cart to show them the value, right? The user, let's say if it is a lipstick brand, we work with a couple of them, right? People are able to simply apply lipstick, shade, foundation, and all of that, and then directly purchase the particular product. It could be a fashion brand we're working with where AI sizing guide tailors the entire experience. And then, you know, for example, I am a 42 wearing, I'm a male 42. The entire catalog basically navigates to a place where everything 42 is displayed and you can try all of them virtually through a phone or a laptop or an iPad. So, not asking people to do away with their existing devices, but just changing the form of content from very, very simple photos, videos, which is not seen innovation over the last two decades, right? Now, moving into a 3D and then sort of giving them a try on sort of facility at the comfort of their home. So, brands have seen crazy amount of improvement in their conversion, add to cart to purchase conversion, but then there's also a lot of traffic because they're launching campaigns on their Instagram, showing those 3D renders in different, there's a shoe brand, runs all their campaigns with all the 3D assets which are generated by see-through, right? So, they sort of take it onto their stories, then Instagram reels, and there's a CTA to sort of go try it. So, the different things which are working out for different categories and different brands, but then one thing is pretty clear, we are going to do away with photos and videos in the next couple of months or years.

Dhruv Sharma: Dhruv, any final closing question? Oh, I've actually run through all the ones I had. This is a very interesting business. I'm still learning more about it, yeah. No, wishing you all the best. Thank you for coming on our show, Vinay.

Utsav Somani: Thanks, thanks, thanks. It was great chatting with you, man. Thank you. Thank you. All right, listeners, moving on to our last segment, we've got Ajay of Amiza. Ajay, welcome to the show. I think you're on mute. Sorry.

Ajay Rao (Emiza): Hi, nice to meet you both.

Utsav Somani: Good to have you on the show and thanks for being a friend of offline. So, 2015, you jumped ship from logistics and decided to enter the Indian warehousing industry. What was the challenge or the white space that you saw back in the day that you really wanted to solve with Amiza?

Ajay Rao (Emiza): Yeah, so I was, before I left and before I joined Amiza, I was in a 3PL and there we were basically doing warehousing for large enterprise, auto brands, chemical. And so it was a very, very traditional type of a business and e-commerce was emerging at that point in time. And the white space we saw was that there were a lot of these, with e-commerce coming in, distribution was getting democratized, right? You could now, a small brand could reach an individual customer, which was never the case earlier, right? You had to use traditional channels. And that type of a business required a very, very different sort of a fulfillment platform to enable them, right? So there were two, three things. These brands being small, they couldn't afford any upfront costs of setting up a warehouse in, you know, upfront costs, number one. Number two is they needed multiple stocking points across the country to be able to reach their consumer faster. Number three was, they were selling on all of these different channels like Amazon, their website and all of these, and which were all fairly complex operations. But these are all individual, single units going to an end customer, huge amount of customization, returns handling, none of the existing 3PLs even had a clue about this kind of an operation, right? So these are the three gaps that we saw back then. And we felt that there was a 3PL, there was clearly a space for the 3PL that could democratize warehousing, give access to world-class infrastructure on a pay-per-use basis, give distributed inventory across the country, and be able to solve for all of the problems that come with e- commerce, like I mentioned, including huge amount of variability, right? All on a variabilized model. So that was the gap we saw. And that was sort of, we were one of the first people that actually were to sort of launch the solution for these smaller D2C brands in the country.

Utsav Somani: And I'm sure a lot has changed because of quick commerce as well. I believe you're working with 180 plus brands. So some of them like fashion and beauty, and I mean, they might have different timelines versus say something else in different categories as well. Do you have separate software which helps you in prediction of what needs to be stored where? Or do you borrow the software from somebody else?

Ajay Rao (Emiza): So we primarily are an infra and an operating platform. So the brand takes the onus of planning what inventory to keep where, because they know what sales they're going to run. They understand the demographics behind their products, right? So we basically an enablement platform. So we typically have a WMS or a warehouse operating system, which connects with the brand systems or like there are two or three softwares which are fairly prevalent in our ecosystem. So we're connected with all of those softwares. So the brand is either operating on that software or on our system. And we receive the inventory, connect our software to those available systems that are already being in use and the orders then flow into our system, right? From there we fulfill. So we don't get involved in the projection or the planning that is entirely in the brand scope.

Dhruv Sharma: Arjun, can we do a quick then versus now of, you know, when you first started the business, what the warehouses looked like on the inside? How was the software stack? And what's it like now?

Ajay Rao (Emiza): So I think back then the warehouses were definitely, one was only in the three major cities. So one big difference is the footprint with Tier 2 and Tier 3 sort of gaining prominence. We're seeing the network expanding from three or four cities to maybe about 12 to 15 cities. That's one difference. Second difference is the size of the warehouses. Back then, maybe 20,000, 30,000 was considered par for the course. Now, most of our buildings are one lakh to one and a half lakh. Third is the level of capital expenditure that goes into the warehouses. Earlier warehouses were fairly simple. Now, all of our facilities have literally buildings within buildings, right? So we take like a big box and within that, we would create like multi-tier shelving. So within a one lakh square foot building, we would create a racking footprint of 70,000 into four levels, right? So that's like 280,000. So the expansion of the cube and expansion or utilization of the warehouse has significantly changed. Automation has come in earlier. We didn't have conveyor systems and all of that. Now we have conveyors, moving products across floors. Camera systems have evolved. So there's a huge change in the infra.

Dhruv Sharma: How about the cold chain, Ajay? Because, you know, we have a lot more perishables available on tap now than we did back in 2015-16.

Ajay Rao (Emiza): So I think perishables is something that we don't touch as a category. QCommerce has enabled that because QCommerce with their dark stores right next to the next to the sort of the consumption centers, they're able to sort of deal with perishables a lot better. We sort of do package goods. Max shelf life we handle is maybe three to six months. We don't go below that. But it has changed significantly.

Utsav Somani: And something that I read about a variable cost as your pitch as well. So is it like a WeWork for warehouses where you pay?

Ajay Rao (Emiza): That's the perfect example. We are actually like an office or a WeWork of warehousing. It's like in WeWork, you can take one seat, you can take like a cubicle or you can take a cabin or a whole floor. It's the same thing in our warehouse as well. A small brand can come in and start with just maybe 100 orders or 500 orders with one or two racks, whereas or maybe occupy 50% of the warehouse. So we offer all of that. And it's all completely variable. Even at scale like brands, some of the largest brands who work with us today, enterprise clients also operate on a variable model with us. In fact, we are the only 3PL today in the country that's 100% operating on a variable and a multi-client model.

Utsav Somani: And this brings risk to you, right? In terms of CapEx and then you have to, I mean, the price. I mean, it's a financial arbitrage play to some extent. I'm simplifying it. How do you think about that?

Ajay Rao (Emiza): So I think over time, we've evolved as our customers have grown larger. We've brought in minimum guarantees. So we give them the flexibility. Every month they can project what their business is going to look like. And they have to give us a minimum guarantee of about 80%. So they have a play of 20%. So they carry the risk of 80. We carry the risk of 20.

Utsav Somani: Interesting.

Ajay Rao (Emiza): Yeah. And what happens is that most cases they end up beating their projections. So that's, so it really sort of comes where you're short.

Dhruv Sharma: And the variability, so to speak, how accurately can systems now forecast into the future, Ajay? Is it still a black box?

Ajay Rao (Emiza): Still a black box. OK. It's very, very tough because there's so many different factors. But I think on marketplaces, so I think, let me put it this way. There are two things that have changed, right? One is marketplaces are fairly good in terms of the projections they give because they are able to throttle the orders. They know at a facility level what your capacity is. And if they throttle the number of orders, they give you basis your capacity. So that's number one. So that's fairly managed. On the D2C side, I think over time, brands have been able to figure out their spend versus order ratio, which was earlier a bit of a black box. So now a brand knows that if I spend X amount of dollars, that translates into Y amount of conversions and X number of orders. So I think that equation also has gotten fairly strong. So today when brands give us projections, especially for sales, they're usually fairly accurate. Where they go off is typically in terms of BAU and in terms of new SKU launches or how certain SKUs will perform at different points of time in different areas based on seasonality, demographics, performance. That portion isn't solved for yet, but sale events are fairly solved for now.

Dhruv Sharma: And if you talk about, I was going to bring that up. Yes, no, I was doing the same thing. I have to go ahead.

Utsav Somani: Yeah. So was it the automation is expensive and Indian labor economics are different. So how do you think about that in terms of ROI? Because labor is cheap in India and will remain cheap for a while because we've got a very large population Yeah, like automation for automation's sake is not going to work here, right?

Ajay Rao (Emiza): Yeah. I think there are two things. One is solving, see, one is there is a labor crisis in the country, right? Today in May, we saw a huge amount of disruption, especially with LPG issue happening. With lack of LPG, a lot of the migrant labors are dependent on getting LPG gas from the black market. So that supply was turned off. So people didn't show up to work. So that was number one. So labor avail, second is with cities like, with states like Lucknow and Lucknow, sorry, UP and Bihar improving, people prefer to stay there because they're getting employment there, which wasn't the case, maybe five years ago. So therefore labor pools are drying up, right? So there are a lot of things that have changed over there. Third is labor costs are also going up significantly while like Haryana, you've seen recently increased the labor rates by 36% in one shot. Uttar Pradesh followed suit. Now Karnataka is talking about increasing it by 61%, right? So, but while the costs are going up, the quality of labor and the throughput is not going up at all, right? So I think very soon there's going to be a point where that arbitrage is not really working. Labor availability, costs going up and third is the sort of accuracy level, right? So I think all of these are slowly pushing us in the direction of automation. I think it's going to happen sooner than later and it'll happen in small increments. It's, you know, like it started with conveyors. I think then maybe it might move to picking assistance. Like my picker is still picking the items from the rack, but a small bot is following the guy and the moment the crate is full, the bot will take it back to the packing station, right? So that will help. So I think that there's going to be incremental sort of improvements that are going to happen that will sort of drive it up. I don't think you're going to go from zero to a hundred. It'll be as like a like a slow needle moving, right? But definitely it's going to happen.

Utsav Somani: Actually give a shout out to some cool warehouse automations that you've used and the companies behind them.

Ajay Rao (Emiza): So I think in India, we've spoken to a few, but like there's Unbox Robotics, there was Gray Orange at a point in time, but none of them had been able to come out with solutions that are what we really need today. So we are actually now looking at countries outside or looking at like China and stuff like that. And I've got a visit planned next month. We're going there to visit a few companies there and to come out with these specific solutions that we need, right? Because everybody who talks about automation today goes for the extreme. They go for ASRS and very capex heavy structures, which I don't think will work for us because we are handling so many brands, so many different variability. Those systems work within a very well-defined environment where there's a fair amount of predictability. But different products, different SKUs, I don't think there is a solution there, which is why I said it will go through an increments. Whereas, you know, we solve for simple problems, like I told you, like picking, you know, moving product from the rack, from the picking area to the table. So there, there are a couple of companies in China called Geek Robotics. They're doing some fairly interesting stuff. In India, there's also Adverb, who's also doing some fairly interesting stuff. So, but like I said, it's got to be horses, horses for courses. And we don't want to go from zero to 100. We want to go from zero to 10, 10 to 30, 30 to 70 and 70 to 100.

Dhruv Sharma: I think we should lock in a part two episode right away so that Ajay can do a trip report when he's back from China. And Ajay, can you also tell us, you know, these giant warehouses huge, involve huge capex. How do people generally think about site selection?

Ajay Rao (Emiza): So, so I think it depends on what hat you're wearing. If you're a marketplace, like a Flipkart or Amazon, you would think about it very differently. If you're a brand who's, you know, you think about it differently. So it really depends on what perspective you're thinking about it from. From our perspective as a 3PL, we think about it very simply from two points of view. One is we need to be fairly close to the logistics ecosystem. Where do the courier companies have their hubs, right? Or where do the marketplace have their hubs? Because they're going to pick up from our facility, right? So that's number one. Number two, we think about labor pool availability, right? Where are we going to get labor easily? Can we, is there any public transport close by? Can we run a bus service? So those would be the two things. Third is from a safety point of view. We don't want to be close to any petrochemical type of facility or alcohol type of facility or maybe where there's risk of flooding, right? So these are the two or three.

Dhruv Sharma: How about utilities? Is that a consideration as well?

Ajay Rao (Emiza): Which one?

Dhruv Sharma: Utilities like power.

Ajay Rao (Emiza): Power, I think is fairly, because if you're selecting a place which is close to a delivery warehouse or a new part warehouse.

Dhruv Sharma: It's a small problem, yeah.

Ajay Rao (Emiza): They're going to get power and you're going to get power, right? So, yeah.

Utsav Somani: And in terms of, I mean, I'm guessing you serve the full stack, right? D2C, e-commerce marketplaces, quick commerce even.

Ajay Rao (Emiza): Correct, yeah.

Utsav Somani: Which customer is the hardest? Economic same for all of them or economics as well?

Ajay Rao (Emiza): So for us, right? The way we think about it is that we are enabling multi-channel fulfillment. So if you're a brand, if you're doing, like some brands are more D2C heavy, some brands are pure quick commerce, some are traditional businesses. So it doesn't matter to us, right? We enable all channels. So for us, the unit economics, what happens is that channel to channel, my revenue realization for transaction changes, but my unit economics or margins remain the same. So let's say for a B2C order, if I would charge, let's say eight bucks a unit, the same unit in a B2B would be two. So it's a form factor of one is to four, but my margins as a percentage would be the same between the two. But in absolute terms, it would vary because of the size of the transaction.

Dhruv Sharma: And Ajay, for any entrepreneurs who want to be, who are or want to be building in the logistics space, if you were to leave them with some unsolved problems in the first mile, the middle mile, the final mile, what do you guys worry about the most?

Ajay Rao (Emiza): So I think the biggest challenge is around returns, being able to A, on the logistics side, A, getting returns back into the warehouse quick enough because most of the courier companies have built a very strong forward leg, but the returns by design or by default, they don't get that level of priority. So if it takes two days to get to the forward location, it may take six to 10 days to get back, right? So that piece is something that needs to be solved for speed on the return because that's inventory. That's lying sort of in transit, right? Which needs to come back in. Second is on the warehouse side. There are a lot of challenges. Each marketplace is making it more and more difficult in terms of claim settlement. They're raising the bar just to protect themselves. So a lot of problems there in terms of how do you get your returns process within the warehouse in such a in time and in such a manner that your claims don't get rejected. So these are two big problem statements. Another problem statement would be really around inventory forecasting, inventory planning. How do you get the right product in the right quantity at the right place, right? I think that's and how do you get your fill rates upwards of 95%, right? So these are some of the things that are problems that are yet to be sort of well solved.

Utsav Somani: Three years from now, where do you see Amiza as a national warehouse network? Or do you see a future for Amiza where you become that asset light software enabled layer where people are building warehouses and you're sort of helping drive demand and digitize that.

Ajay Rao (Emiza): So I think we definitely owning the asset has a huge amount has a huge advantage because that is the one the asset ownership. So, you know, in an AI world where businesses are getting disrupted left, right and center, I think asset ownership is critical and that is what gives us our edge. So I don't think we're going to let go of that. We're going to continue to build assets out. And for us, that's the that's we make our money on the asset, right? So for us, we want to keep on doubling down, adding more warehouses, more facilities, building a network of multi-client facilities where I thought maybe 27, three years from now, I want to be at 75 because to me, this is an I've created assets with recurring revenue, right? So that's the way we're thinking about the business.

Utsav Somani: All right, Ajay, thank you so much for coming on the show. Have a wonderful weekend. Thank you. All the best for China. Yes, thank you. Take care. All right, listeners, that's it from us. We'll see you on Monday at four o'clock. Have a wonderful and a safe weekend. Bye bye.