Receive exclusive updates and insights directly in your inbox.
<h2>Mrinal shares,<p> Walmart has been in the business for like 57- 58 years now. Data has been front and center for us to be able to build on such a large operation. We run about 11000 stores worldwide with close to 4600-4700 stores in the US. So data is an absolute integral part of it and there’s a lot of places like our shopping experience, mobile apps or desktop apps and supply chain and a lot of other things that we do in terms of merchandising decisions and pricing decisions to a lot of it is driven through data science and machine learning algorithms. I think the way I internalize it is I think dugs challenge to us is to think all across not only some of these critical high transaction king of systems but think about every place where we are building platform services how do you make data as an integral decision making in terms of your architecture choices. So that we can leverage machines, where machines are good at, and leverage human Intelligence with our large workforce and deep retail experience; where we leverage humans and our associates’ knowledge to make decisions where machines might be imperfect. Things which way machine can with the right amount of the data trading and expect that can drive a lot of right decisions, that just put that. So a lot of our application designs, apps, or services are now data science has become an integral part of it. In fact, if I look at my team, in the last year one of the single job families which have the highest amount of growth has been the data science and machine learning engineers. The core of a retail business is for us to be able to provide competitive prices to our customers. <strong>The supply chain is a disproportionately high cost</strong> and especially in certain kinds of merchandising like fresh produce and something like that it is absolutely critical because you need to have temperature control, cold storage, and proper transportation needs to have. So we are disproportionately deploying IoTs in our refrigeration platform. in managing. We are getting their feedback from our thermometers in terms of what temperatures are and is really helping us sort of not only maintain the freshness but also the amount of energy consumed to keep them fresh, and not sure if you know that we have made a promise to get rid of 1 billion tons of greenhouse gases from our supply chain by 2030. So this is a big initiative, we are doing a lot of things beyond just this, but coming back like IoT is our front and center. Our stores where our fresh produce is kept; the supply chain where our fresh produce is in transit; our trucks so that we continue to not only be efficient but deliver the freshest produce to our customers. All that we do is a lot of planning and execution rigor that needs to happen because, unlike general merchandise, right like a poster or something like that, it has a long shelf life and they are not perishable. So at what time they are picked from by the farmers, how much time it takes for the product to ripe at the shelf, all of that planning needs to sort of come to probe into the process. In cold storage, there is a certain trait each of those vegetable needs in terms of the amount of moisture, the amount of temperature, and things like that, all of that comes into the picture to make sure that by the time it gets to your fridge or when you come into our grocery stores, It stays fresh it is at the right optimum place. And then again we have bundles where, for example, if you buy bananas, we know when people are buying bananas are like a dozen or so, the consumption will carry on for days or so, like there as some other things like watermelon, chances are, you want to consume it in a single time. So all of that comes into the planning process for merchandisers, or buyers or suppliers into an integrated supply chain that we come to know that an item is coming into an ecosystem even before it comes into it. So that we can do our own planning process, what would kill this whole thing is, if we have lagged in our processes that the item is sitting too long in our warehouse or the sitting too long in our supplier’s warehouse. The supply chain becomes an extremely integral part of it and coming back to your question about IoT, this is where IoTs are coming in place and helping us. https://youtu.be/zEQI6UDRe4U</p></h2>
Read More
<h3><strong>Gaurav Shares</strong><p> The challenge in today’s world with data is that people are on two ends of the spectrum, there will be some products/ companies that say I don’t have data for my use case. In my previous start-up, we were in that bucket, how do we get data for the application that I want to build? There are a bunch of people on the other end who have lots of data. Having data is not the challenge, making sense of the data is the bigger challenge. For example, there are CCTV cameras today that are generating a lot of data, a humongous amount of data that we cant process. I think that was the genesis of the data platform – can we centralize it, make sense of the data, make it easy to process and extract value out of it. I think we were not the ones who started that trend, US companies like Facebook showed that there has to be a data platform and we were the first to adopt it in India. Flipkart very quickly realized that data would become a differentiator. It is not because Flipkart has been around for a long time that it will win but it is because Flipkart has understood what these users want, what works, what doesn’t work and how we create a better experience for the users day in and day out. We started that, in today’s time I will be surprised if any startup says that they don’t have a data platform, data is the key. We are discussing hybrid cloud vs single cloud, so yes this is a question that is on the top of the mind for a lot of CTOs and CIOs, I hear a lot about this question in other forums too. We are at a stage in the evolution of technology, cloud came in and we realised that we can focus on what matters to the company and build our application, our functionality and offload availability, scale and reliability to a cloud provider. I still remember in 2012 we had a massive catalogue in Flipkart which powered our website, we were storing it in MySQL. There was no relational database service or cloud providers who did that. So, there was a large dedicated team that ensured we could run the data at that scale and keep it up. There were investments in specialized hardware to store this data. It is a deep, highly specialized, skill-oriented work, hard to get people who do this. You need to be on it day in and day out. Now cloud comes in, so startups on day one can go to RDS, you start small but you can quickly scale to 1000 requests in a minute, it can support that with all the data coming in. So, the way cloud started off, people were really overjoyed with it solving their problems, businesses could focus on what they needed. Generally, clouds are reliable, they have a concept of SLAs – what percentage of time can one say the cloud will be up and available. We measure them in nine’s – 4 nines, 5 nines and so on. These tell you, in a month what is the maximum downtime that can happen, we are looking at most 10 minutes of downtime in a month. What this did was now most of the companies are on cloud today and sorts of applications are coming on cloud and this is where this question comes – now that you are on the cloud, are there any challenges or downsides to being here? I believe the biggest thing you end up seeing eventually for a company which grows massively like udaan has a huge footprint on the cloud. Even at Flipkart, we had a huge footprint on our own data centres. You begin to realise that if things go wrong you have very limited control because someone else is doing it. They are not expected to go down, if they do go down something really bad has happened and anything can go wrong. For instance, if a company’s data centre is in a country that is prone to flooding, it would be just water, which is out of your control. If things go wrong, it is usually bad, most of us face that challenge – we have to figure out a way to mitigate the downside or impact on the end customers. There can be instances where it is very little you can do, the provider is fixing the issue. This is one kind of challenge you see, second challenge is that there are multiple cloud providers, some are good at something, others at something else. As the dependency on cloud platform services starts increasing, you realize some platforms are much better than others on some of these services. The third challenge is you see that the cost starts becoming material over time. Like everything in life what is good eventually ends up showing the challenges on the bad side. What is working to your advantage on the cloud, for example, is to empower my engineering team. You need to build a new service and scale it out and it is just a command or a click on the UI. What it also means is that at scale you need the right governance otherwise your cost will just blow it up. So, that’s the third thing that hits companies. We’ve seen that at scale most companies have gone to their own cloud strategies, however, there have been exceptions. All the large players like Facebook, Google and others started depending on cloud providers. However, Google eventually externalised it like GCP is available as a platform to others. And then there are companies like Netflix who decided that they will not have their own cloud, it continued to depend on AWS. But they have very deep connections, so there is a lot of custom work happening for them to solve these three challenges – cost, availability of features, capabilities and uptimes just to make sure they have the least number of challenges. A lot of other companies like Alibaba in china ended up building a cloud for themselves. Flipkart built its own cloud, we had a cloud team. My take on this is a multi-cloud or a hybrid cloud strategy, that is one of the ways to mitigate it. You create infrastructure which allows you to take the best of what is available to you or certain services. For example, you depend on cloud API to do OCRs or to do accurate facial recognition and one of the players has a much better performance at this time. Why wouldn’t you do what is best for your business? So the availability to go to multiple clouds – I would say the availability to have something of your own which allows you to do deeper integrations with your data. For example, if you are storing sensitive data like Adhaar card for KYC, you require native support like hashing and encryption, ensuring the data is more secure. You don’t want to put this data on the cloud, keeping these options is the thing to do. But these require investments and do not come for free. One of the biggest costs of the cloud is to take the data out. If you are on two providers, you’ll have to figure out how to bring the data out. </p><h3><strong>Watch the full podcast here:</strong></h3><p> JTNDaWZyYW1lJTIwd2lkdGglM0QlMjI1NjAlMjIlMjBoZWlnaHQlM0QlMjIzMTUlMjIlMjBzcmMlM0QlMjJodHRwcyUzQSUyRiUyRnd3dy55b3V0dWJlLmNvbSUyRmVtYmVkJTJGbUdmRHJzUXRfQ00lMjIlMjB0aXRsZSUzRCUyMllvdVR1YmUlMjB2aWRlbyUyMHBsYXllciUyMiUyMGZyYW1lYm9yZGVyJTNEJTIyMCUyMiUyMGFsbG93JTNEJTIyYWNjZWxlcm9tZXRlciUzQiUyMGF1dG9wbGF5JTNCJTIwY2xpcGJvYXJkLXdyaXRlJTNCJTIwZW5jcnlwdGVkLW1lZGlhJTNCJTIwZ3lyb3Njb3BlJTNCJTIwcGljdHVyZS1pbi1waWN0dXJlJTIyJTIwYWxsb3dmdWxsc2NyZWVuJTNFJTNDJTJGaWZyYW1lJTNF</p></h3>
Read More
<h3><strong>Sonia shares,</strong><p> Another thing that Atlassian does well which has helped us in this aspect is, that we won't hire like three engineers and sort of leave them on their own to get something done right. We will hire a full triad of a manager like an engineering manager, product manager, and designer, and will give them some level of ownership on the roadmap so that you are excited about what you're doing and you can see the impact of your work. We don't have a hub and spoke model, where all decisions happen in one location, we operate like it's a pretty global company. Every office has its own set of areas that they are driving, which has helped a lot as well so we have multiple leaders in India for example who globally lead teams, whether it is product, design, engineering and I think all that makes a huge difference, I think that has helped us huge to attract talent. So we do follow the triad model, basically, the triad is Engineering, Product management design, and then you have the triad at many levels so like my triad has a few triads under us so it's like a pyramid. They all sound very perfect it's funny I kind of joke sometimes my triad does not work, the triad is great because it's not always easy you know sometimes you have a level mismatch but having said that we do lean on triad a lot, and when is a lot of trust between the triad members, it makes a huge difference as I have seen like how I am such an engineer sometimes I go and talk to my design counterpart just to balance out my company because I feel like design comes with a very and definitely my triad partners both in the marketplace and here, they have helped me think very differently from how I would normally approach a problem, so from that point of view as well it really helps when you have the diversity of thinking and perspective, of course, you know that the trial like the success and failure is all tied to the triad so it becomes easier to save your team, the team is going to make this happen I think that has helped us quite a bit as well. I think my Urban ladder experience helped me a lot because I got exposure to you know working with customers service, operation, and marketing like I got to work with the head of the department where are you learn so much of their craft which I don't think earlier and had exposure and I definitely cherish that. Microsoft for example, when I got in it was 60,000 people and today Atlassian maybe 7000 I think even that scale makes a massive difference in how you operate and of course, every company evolves and there are pros and cons. It is very hard to say this vs that, how you operate still at 7000 is significantly different as well at Atlassian what i have seen is you know we do try to ensure that like you want every decision to be made with the information is present and take it at the lowest level by your maximum information and that definitely helped; that goes back to you know to empower your team, unleashing a potential. How do you unleash the potential of a team that empowers people, are empowered to ensure that they are able to make those decisions. So we do like there are certain practices we follow, definitely that lean into Agile a lot. There is definitely a fair amount of autonomy within the team. In fact, how zebra operates is very different from how the team might operate. I think that autonomy is critical you know because that specific problem you solving right and what you need to like that judgment and I think that is how, where tech is different right. You value somebody’s ability to judge, to make the right judgment calls and i need to lean into this looks interesting and this is the area where we should focus more right. At this scale i see we still manage to do that relatively well, there is a big because we follow the triad model, you know design especially plays a very key role in bringing in like User experience and User experience research that we have a pretty big team that does use experiences research as well and they bring a lot of insights around that having said all that you know I think as engineers especially it is i mean the onus is on all of us. </p><h3><strong>Authored by Pratheek. V</strong></h3><p> For more information, please reach out to the <a href="Marketing@purplequarter.com">Marketing Team.</a> To find out more, watch the entire podcast here: https://youtu.be/GWOBw2ug4JA</p></h3>
Read More
<h2><strong>Natraj Shares</strong><p> Yes, blockchain will very much be a part of cross-border payments, at present cross-border payments are a pain to put it in a subtle way. Most of it depends on two or more people, who are basically the intermediaries who move the money and make swift transfers. The problem is, one you need accessibility to banks or financial institutions second the cost of the transfer is prohibited and third is that it is slow. Cross-border remittance is a huge business with about $500-600 billion in remittances each year across the globe which makes it the right time for disruption. Some years ago companies like Ripple, which started in 2012 and Stellar (started a few years later by the same founder of Ripple) have done a lot of things around that. In countries where there are sanctioned sort of banking systems, there is also adoption, for example, the largest amount of B2B crypto transactions happen in countries where you least expect it like Nigeria to name one. As of now I still think it is in its infancy, while in theory Stellar and Ripple can solve the problem there are a few problems around it like accessibility and conversation to a fiat currency. While moving money from the US to India, by using the Stellar network, it transfers immediately, it is fast. Stellar unlike bitcoin is a free mind it doesn't take a lot of time. Bitcoin on the other hand is not for transferring money, it becomes very expensive but Stellar and Ripple is not. In theory, you can, but for example, if you do it in India, the challenge is what is the liquidity? are there enough anchor points? Anchors are the on-ram and off-ram, if you buy a currency, the transfer happens and they move from one point to another. Are there enough anchor points, is there enough liquidity for such a trading pace and things like that. I think it will happen, the industry is right for disruption, although I think it will take a little longer to mature, and completely replace the swift system but it will democratize it. Right now it is available to people who have access to banks and other things. But there are a lot of countries, not just India but also Africa and other places. For instance in Africa, Mpesa has done really well and become a huge success. That's why I don't see why this will not, this will bring in the security of the blockchain, the distributor's trustworthiness of the blockchain and disrupt the system. Even at Zolve, as of now we handle this through traditional mechanisms and we have a partnership with Transferwise and SBM in India to move money and they are fabulous partners. Both have solid technology, we use them to move money and provide the experience to our consumers. At some point we will also see how to get into it, as of now we don't have it, right now Stellar looks like the best option especially in countries where we have anchors available who can actually convert it into a fiat currency and there us enough liquidity, we haven't made any concrete plans yet. We are planning on implementing blockchain at Zolve soon, during the pandemic, the entire concept of blockchain went from being exotic to the mainstream. I think the kind of innovations that have happened, are happening is phenomenal. I think there will be a bit of consolidation as changes will happen but I think the consolidation will happen and we will have a much more stable, trustworthy, cheaper and faster way of moving money. Something that is accessible to people without very formal or large banking systems. To find out more, watch the entire podcast here: https://youtu.be/JQ2LEVuwC4A?list=PLPXxlEzxWZditWn76bh46eLUWgoCjfScT</p></h2>
Read More
<h3><strong>Thippesha shares:</strong><p> Today, AI is used more as a buzzword in conversations, however in reality it depends on the company when, where and how they use AI and ML. What matters is what stage is the company at, that is what determines if the use of AI and ML is apt or not. For example, if the company is a startup, then the said data can be managed with the use of excel sheets and if the company is mid-sized then relevant engines can manage the data. One does not necessarily need AI and ML models to manage this level of data. If a company even then goes to apply AI/ML there is a high possibility that it might not yield the right results as there isn’t enough data to train these models. Even at Amazon, I didn’t use AI/ML on the first day, for the system I first scaled the business to billions of dollars only then did we begin to learn and implement the system. We took regular feedback on the engine and using that trained the data to know what to deposit, how to classify the product, etc… all of this can only be done after a point in the business. With the cost of competition and storage coming down along with companies like Amazon, AWS, Google, Microsoft, etc. that provide AI and ML features; the overall cost of AI/ML has reduced across the board. But I’ve seen that when it comes to third parties, i.e, especially startup companies, they come up with solutions where AI/ML is used in such a way that they aggregate the data, normalize the data and test the model. As a single company, you may not have the desired data but as a group of companies you will have the collective data and you can train the model without compromising on the security of the data. Thus you will be able to provide solutions to other companies. In many areas such as consumer behavior and B2B companies among others, AI is already adopted, it is not a gimmick any longer, it has long become mainstream. Whether you know it or not, you and I use AI/ML every day either in the form of talking to Alexa, watching something on Netflix, or the form of online shopping; you name it and it’s all AI and ML. AI and ML are working behind the scene but no one is labeling it as such. Thus it has become mainstream. At Upstox we have scaled, we started with spreadsheets, whereas now we are at an 8 million customer base company, which makes it difficult for us to just add manual resources. After a point it is not a viable solution, what we did is create a data strategy, we started collecting data from different sources and collating it in one place. All business units have access to and can use this collated data. Now from that data, we get customer insights, segment insights, trends and where and all can AI and ML be applied. We have already begun to apply AI/ML in service areas, where we can serve our customers better. We are looking at the recommendations, that is how we recommend the right content to the right consumer. We are also looking at some personalization experience, we are doing various experiments in this space as well, definitely, AI and ML will be a part of Upstox and continue to build the system and provide scalability.JTNDaWZyYW1lJTIwd2lkdGglM0QlMjI3MDAlMjIlMjBoZWlnaHQlM0QlMjIzOTQlMjIlMjBzcmMlM0QlMjJodHRwcyUzQSUyRiUyRnd3dy55b3V0dWJlLmNvbSUyRmVtYmVkJTJGRlhiWjVxVHhZb1ElMjIlMjB0aXRsZSUzRCUyMllvdVR1YmUlMjB2aWRlbyUyMHBsYXllciUyMiUyMGZyYW1lYm9yZGVyJTNEJTIyMCUyMiUyMGFsbG93JTNEJTIyYWNjZWxlcm9tZXRlciUzQiUyMGF1dG9wbGF5JTNCJTIwY2xpcGJvYXJkLXdyaXRlJTNCJTIwZW5jcnlwdGVkLW1lZGlhJTNCJTIwZ3lyb3Njb3BlJTNCJTIwcGljdHVyZS1pbi1waWN0dXJlJTIyJTIwYWxsb3dmdWxsc2NyZWVuJTNFJTNDJTJGaWZyYW1lJTNF</p></h3>
Read More
<h3><b>Abhinav Shares:</b><p> Essentially supply chain is the backbone of any business, be it fashion or pharma. The pharma supply chain particularly is a massive challenge, especially in our country, if one were to zoom into the pharma supply chain, one would find that the supply chain is extremely fragmented, what I mean by that is it's a traditional supply chain with limited penetration of information technology. Statistically speaking, there are anywhere between 7.5 to 8 lakh independent distinct registered retailers of medicines in the country. All of them are disconnected from each other, so what I mean by that is, let’s say there are XYZ medicals in your neighborhood and then there are ABC medicals, both are 500 meters apart and are totally disconnected from each other in the sense that they do not have any idea of what medicines are available at each other's stores. What it leads to is that as a patient when I walk into a pharmacy to get my prescription filled, in our country chances are that only 60% of all the medicines will be available in one shot, this is the country's average. Now if you see, we are fortunate to be living in big cities where we have access to multiple pharmacies, the moment you start to step into tier 2, tier 3 cities it becomes an acute problem, there are limited pharmacies with limited space, they can only keep so many medicines. Today roughly, there are around 300 thousand stock-keeping units of different medicines available in the country, and a normal, even good pharmacy can only keep anywhere between 5000 to 10,000 SKUs in stock, around 2,90,000 SKUs have to be procured from somewhere else. You will want to know where is that somewhere, it is basically a distributor. There are 8 lakh retailers and then there are anywhere between 80 thousand to 1 lakh distributors in the country, these are again independent isolated distributors with very less penetration of technology. Let's take an example, a lot of people have experienced this in a pharmacy while getting their particular medicines - “sir I don’t have this medicine right now, I can deliver it to you tomorrow or the day after or whenever I can’t procure it”. When you look at the process behind the scene what happens is that the retailer calls his distributor to procure the medicines. It is possible that this retailer in hand is working with two, three, four, or maybe even five distributors over the last decade or two, this is entirely relationship-based. Now, the retailer calls his distributor to ask about a particular medicine, and the distributor says sure, you will get this medicine tomorrow or the day after. This further leads to a different problem because even the distributor has very little idea of what all SKUs and quantity of these medicines he has with him. Thus again because of the lack of technology the two invested parties here who make a promise to the actual person in need are frankly unaware of what’s available in the market and what’s available with them. This creates some massive access problems in this fragmented supply chain, and we are trying to solve these bigger problems at PharmEasy. We are not only trying to solve the problem of immediate access online we are also solving these problems for the retailers and the distributors. We happen to have our biggest marketplace of retailers and distributors; on one side you have the retailers and on the other side you have the distributors, our technology is kind of embedded in these entities and we present a real-time view to both the retailers and the distributors of what’s available in the market, what are the service level agreements (SLA) on which one can get these medicines, what discounts are available, what schemes are available, where is the order when placed, if you have to make a payment, where it is stuck, you can make a payment on the platform and a lot of other things from logistic to warehousing, demand visibility to price visibility and discount visibility, one can have all the information available on one single platform. If compared this is somehow similar to what the likes of Uber and Amazon are doing to us as a consumer. I mean 10 years ago, hailing a cab would be a nightmare right? You have a flight to catch at 5 in the morning and you are at the mercy of the driver, you make a call and you have to literally beg ‘Sir, I have a flight to catch at 5 in the morning, please be there at 4’ this was the challenge that Uber solved, all the information that you needed as a consumer to make a decision is right at your fingertips, from who the driver is, the rating of the driver, the condition of the cab, you get the information about the route, the cost, which just makes things highly convenient for you to make an informed decision. It's the same thing with the likes of Amazon for instance or Myntra or Flipkart, right? They are consciously making information available to the consumer so that they can seamlessly decide, this solves the problem of the supply chain in a different domain. We are solving that problem in the disconnected domain of pharma where the need is critical. This is a genuine need, this is not optional, so getting it right and making these distinct pieces connected with technology is our vision, we want to make sure that every independent entity in this pharma ecosystem - the <strong>patient</strong>, <strong>doctor</strong>, <strong>retailer</strong>, <strong>distributor</strong>, and the <strong>manufacturer</strong> are connected, who in the present scenario are highly disconnected from each other. The doctor, for instance, prescribes medicine to you but the prescription is handwritten in most cases and this handwritten prescription is not digital so you cannot look at this prescription at a later point in time, you cannot take this prescription to a different doctor. Let’s say I happen to be in Banglore and my doctor has prescribed certain things to me and for whatever reason, I have to go to Delhi or someplace outside the country, that doctor I consult at that place has to start all the way from the beginning noting down all my allergies to all the conditions that I may have had in the past. This all is a result of a non-digital, non-technology based disconnected ecosystem; we are building capabilities to digitize the doctor practices as well so that the doctor gets the benefits of referring things from the past and also extend those capabilities to you as a patient, so you also get access to your own prescriptions, your history, your symptoms, and your diagnosis. All of this now becomes available to you in a connected ecosystem where even the doctor is digitized, when the doctor gets digitized, the retailer is digitized, the distributor is digitized, now the manufacturers are also directly reaching out to people in the ecosystem using the same connected ecosystem. So, fundamentally from an overall supply chain perspective, the basic problem is that you have five key entities and these have largely been non-digital and they have been disconnected from each other for all this time and in an economy like ours where there are 1.3 billion people and there are only 0.64 doctors for every 1000 individuals in our country, in that kind of a set up a totally disconnected ecosystem would mean a lot of inefficiencies. That is what we are trying to solve by building technology that sits across all of these entities together. Technology, in general, is able to talk to each other better compared to humans and processes, it’s a simple and straightforward thing. Once you build it properly, these pieces can talk to each other in their own languages and the technology does the translation between them and thereby giving you a highly connected ecosystem where every stakeholder gets the visibility of every transaction they are making in the system. </p><h3><strong>Authored by Pratheek. V</strong></h3><p> For more information, please reach out to the <a href="Marketing@purplequarter.com">Marketing Team.</a> To find out more, watch the entire podcast here: https://youtu.be/ydvLBY_c09Q</p></h3>
Read More
