Types of Product Management Roles

1. Introduction: Hi, everyone. Welcome to the introductory video of this class. This is Shua Jen. I work in the product team at Poonpe. Powpei is one of India's leading Pin PEC brands that does over 350 million transactions per day and has over 500 million registered users. Prior to this, I have work across many BTB SAS phones as a platform product manager and has over ten plus years of working experience. I started my career as a data scientist and have a computer science degree from Bitspiai. So in this class, we will see various kind of product management roles that exist, and how can you define the career trajectory for yourself based on your skill set, based on your areas of interest, and what is needed for that specific set of roles. Um, the key detailed takeaways from this class would be the types of product manager roles. For instance, for a growth product manager or from a consumer facing product manager, you will be more inclined towards business and revenue side of things. You will also get to design a lot of user interfaces so that the consumer experience is not marginalized. Second type of PMFOL could be an enterprise product manager that deals with a lot of B to B clients that does long term thinking and also has a great deal or scope in client management. We will go into very detailed understanding or concepts of what a platform PM and an AI MLPM means. Platform PM, how do they have to think long term? How do they have to ensure the platform is able to work at scale while empowering a lot of applications on top of it. Where as AIML PM or a machine learning product manager is a more upcoming product management area, and the manager needs to ensure that the business and the engineering or data science team are working in cohesion. The main translation of this business to tech world is done by an AI product manager. In this class, apart from learning the core concepts and frameworks, we will also go into real life case studies so that you can apply the same in your day to day life as a product manager. This class ends with a really cool class project that focuses on you thinking about your career trajectory, the kind of skill set that you have, the areas of interest, the opportunities that lie in the future, the kind of compensation that one might get with, you know, existing PM rules and forces you to think about what you want to do in the future, so that you can share it with your peers and also make them think about the same. So see you on the first lesson, very excited. Let's get through it. 2. Types of product management roles: To the new flector of different types of product management product management roles. Starting with technical product management. These are the people who work on technical products. For example, might have heard of Octa, which is an access based management tool. Couple of other products in big data world are, let's say, a confluence or an Axl data or Atlin. These are the products that are very technical in nature, but it requires a detailed hand holding of a product manager to think it from an end consumers point of view. Um, so functionalities are primarily tech driven. For example, there might be integration, there might be analysis, reporting, various assessment of let's say, security flaws, updates, releases, right? So all these are very technical features or products in nature, and a technical product manager would lead a would leak, you know, features or products or understanding between businesses and engineering counterparts. So from business and clients, they will learn what the end user needs, what sort of navigation is best suited for them and translate the technical world into the world of different end users persona. So if your end product is used by a developer, the language, the navigation will be as per that. If it is going to be used by, let's say, a slightly non technical user persona, which is, let's say, HRT or a business persona, then the language or navigation or even the features built will be according to that. Second is growth product manager. So these are the I mean, people typically who do zero to one projects or do consumer first projects, right? For example, like a dating up which is more it does not need specialization to know the various aspects of the dating world, but it requires a keen, you know, understanding of consumer behavior, consumer psychology, and to be able to drive projects zero to one. Typically typical metrics that growth PMs or consumer PMs look after is acquisition, retention, and couple of business metrics. So these PMs are more closer to business and end consumers and slightly removed from the technical nuances of the product. Because they're primarily working from the business side as a product manager and relaying the same set of feature requirements to engineers. Third category is enterprise product manager. These are the products used by big enterprise businesses. The skill set needed over here is having a technical or sound understanding of technical nuances, if not a deeper understanding and also to be able to communicate to big enterprise clients very professionally. Because as an enterprise product manager, you would be speaking with a lot of your time will be spent on speaking with clients, and these are big clients. So understanding their needs for a big client, which is a company size above, let's say 1,000 people, they might have a very complex ecosystem with so you translate that sort of requirements, understand that, communicate with big clients, manage their expectations, give them a clear vision that requires a more professional, you know, upkeep of your features of your product. So enterprise product managers are primarily technical and more professional in communication style. Fourth is AI product manager, this product manager needs to well versed with the understanding of data science or machine loading point. Now, understanding of data science and machine learning really helps in this case to get how the product or the features will be built. It gives you an idea of the feasibility of these features as well as what is the accuracy of people who are building, let's say, a recommendation engine, right? A product manager has to understand this is a recommendation. It's not deterministic in nature. It can be problemistic in nature. So some understanding of statistics, machine learning, data science, roughly how it functions, what is going on around the world, what are the new updates that are happening? A product manager needs to be well versed with that. Some of the examples would be a let's say in social media on Instagram, the kind of recommendation that you would see is primarily dictated by data science models, or let's say let's say in a news app, right, the sort of recommendation of posts that you would see, for example, even TikTok would be a good example of this. So a lot of businesses actually gets driven with the recommendation that you show. A write recommendation ensures that the users, there will be high DAU. There'll be high MAU, and the sessions also will be longer, which directly contributes to your business methods. Another good example would be YouTube recommendation. So when you listen to music, a good set of recommendations follow, and that makes you listen to or extend your session longer listening to just YouTube music recommendation. Netflix is also another good example. Fifth type is analyst product manager. These are the product managers come analysts that primarily work on data driven insights. Now, this can be these people can be working in, let's say, the stock market world, Site M fintech world where numbers actually make a big difference in the kind of features and the insights that you get. Last one is consumer product manager, which is somewhat similar to growth product manager, but they primarily focus on having users having a good CX or users having good experience on your product, whether it's a physical or a digital product. On top of functionality, a very good consumer experience, as well as they conduct market research or understand user needs the best in the entire team. So they are meant to create excellent products for end users. A good example would be Apple. Apple is or iPhone is just another phone in the market, but it really excels on understanding user needs, right, creating the best experience possible for the end user, which is differentiated from other type of phones in the industry. So these were some of the types of different product roles that one can aspire to, thank you. 3. Platform PM: On. Welcome to the new lecture of platform product manager. So firstly, what does a platform product manager do? So as you can see in this diagram, we have on the top business layer, we have multiple products like a lending product, a PTB transfer product, insurance product, wealth management product. So these are products coming out of the same company. Wealth management is basically your portfolio management. Insurance is giving out life insurance, health insurance to the end consumers. Lending is giving out credit at a certain interest with handling in case of a dispute, in the case of non repayment of credit, what happens, right? These are let's say four different products coming out of the same brand. Now in order for all of these products to work seamlessly, there needs to exist a platform layer that is common across all, which is categorically, in this case, has a payments infrastructure and a fraud leg. Payments as a complete platform or a core functionality needs to power all these business layers seamlessly. So similarly with infrastructure. So since for payments, all of these has actually, you know, a payment or transaction leg involved where the end consumer is paying a certain amount. So you have to do reconciliation. You have to ensure that if a user raises a dispute, how will the company or the end third party repaid back how to handle charge bags. So this is actual money flowing into the system. So reconciliation invoicing has to be done in that aspect. Um, infrastructure needs to ensure that the payments or the visibility of these features are almost instant, right? So there is no latency. In order to build up these features, there are many microservices that you would need to ensure that the products are the system is up all the time. So infrastructure needs to ensure that, all these servers or the requirements of servers that is actually computing for these businesses is, you know, working fine. It is growing at the same pace as the growth of your users is happening. It needs to ensure the system does not go down. In case of a system failure, what needs to happen. So this is the work of infrastructure leg in the platform lay. Alas is fraud. Rod, again, is a horizontal layer over here because Froud can have on any of these products, and it has to continuously improve as the external frass also keep improving. If you see all these three legs are common across all the products. As a platform product manager, your role is to understand the nuances of each product and power core platform capabilities to all. The skill set needed over here is technical understanding of each of each of these layers, which is payments in fraud, have a overall understanding of how the system architecture works and still keep a profound or, you know, acute view on what's happening on the business side. So this was a couple of ideas or areas that a platform PM needs to be aware of if they wish to pursue in the same direction. Okay. 4. Platform PM: Dimensions: Hi, one, let's look at what TASA day to day job looks like for a platform PM or what are the skill sets needed to be a platform PM. Firstly, they need to be interacting with different types of stakeholders in the ecosystem. Hence they need to have a profound understanding of technical systems and system architecture. Because one of the additional stakeholder are technical experts who communicate or understand things in a different from a different perspective so since it's a highly technical role as well, you need to get a thorough understanding of your technical systems. Alongside the usual stakeholders that a PM has, which is business, other product folks, design, engineering, that goes on in Palo. Second, is ensuring the protection of four platform interactions. These platform functionalities are highly critical in nature. If you change one small aspect, it can lead to a system downfall and to even identify what caused it, how to remedy it is a big task in itself because there are many uh, microservices running under these core platforms. So any decision that is being made, the gravity of making that decision, the impact on the criticality on the existing systems is huge. So as a PM, since you would need to take multiple such calls, you have to ensure that it is protected in the more secure Third is scaling with public APIs and securing access with private. What this means is that your platform, since this platform has to be used by different business layers. Hence, your APIs need to be public in nature so that any business or any application layer can use it directly. But you need to ensure the internal application or internal connections to your system within your platform layer is highly secure in nature. That means if a lending or any other application is trying to trigger or invoke an API, which is not intended to which can lead to a potential downfall, those need to be protected as highly secure private APIs. So this might seem like a technical nuance, but the way a product manager would structure this that these are the site of public APIs, this is the use case and rest everything is private, needs to be highly modularized. Fourth is microservices support structure. Like I said, this can get complex really fast. Uh, the set of microservices required to have these big core functionality systems running can be highly complex in nature. So a product, a platform PM needs to ensure that this is highly abstracted. The technical from the use case part is highly abstracted out. That means if within my core functionality, my Microservice one is calling Microservice two. We need to ensure the templates are same, restructuring are same, and the technicalities is left upon the technical folks or engineers. So Microservice one cannot be very different in terms of architecture and scale from Microservice two because it will continue to lead to complex structures and very different taxonomies will be a nightmare to scale. If the features that are meeting long sustainability goals, like I said, that the whole infra piece or the whole platform piece needs to ensure that it scales over time with a scale of users. Long term sustainability goals would mean that the database that is u you know, holding the most the transaction data or the user data. A, first, it is highly secure. And secondly, the database will keep growing in size on a day to day basis as the users or number of transactions increase. So how do you ensure that the features that you're building or the textag that you're using is meeting your long term sustainability goals? So even the tech stack has to the features has to follow the textag that is being used. And hence for a product manager, the features or the kind of complexities or latency they can expect really depends on the kind of textag that is being used. Another one is future proofing during platform updates. When you give out a system update or a version release, right, you have to ensure it's backward compatible with everything that you have done. Now backward compatibility also needs to ensure that whoever, whatever application that is using your core functionality is not breaking when you are releasing new updates in the market. Because each of these public APIs can be used by any number of system. Hence sort of documentation, outside of API documentation, proper versioning control is needed to execute this exercise. Different platform monetization model. So most of these core functionalities has a good potential to be monetized or since IFA or these core functionalities does account for a big cost structure in your PNL. So you have to be aware of A, the cost is that is getting incurred by the business players is how much? And how does it increase with time? How does it increase with each business? As well as ensuring that if I want to monetize this core functionality in future, what is the cost structure looking like? Last is focus on flexible platform first culture during the shift. Now, this is important to ensure all of the above. The culture here becomes really important where someone is creating a new microservice, they are following the right taxonomy, they are using the right things to ensure the complexity does not increase multiple folds just with the introduction of new feature in one of the microservices. So all of these are on top of the skill set or the items that you have to keep in mind as a platform PM, all of these are in addition to what a usual PM or a generic PM means. This might seem like very highly technical in nature, but if you extract it out, it basically needs first principles understanding to execute all of this. 5. AI/ML Product Manager: Hi, Ron. Welcome to the new lecture of AI product manager. First, let's see what does it mean or why it has become a more nuanced or another vertical of being a product manager. A machine learning product manager focuses on building products that originated from artificial intelligence or data science technologies. Some of the examples that we already discussed are, let's say, a YouTube recommendation a module, a Netflix recommendation module. Again, these are becoming increasingly important because it feeds directly into the revenue pools, right? For example, let's say, if you take the case of TikTog, the kind of recommendation that you get from public accounts, right? So if I am, let's say, the app figures out my interest areas that I like to, let's say look into sports or news or weather, right? It has to figure that out from the basis of my historical content that I watched. Once they figure it out, they will be able to better recommend those public accounts who cater to do this type of interests. And on top of it, they can actually recommend me products that I could buy based on these interests. So you have to really ensure that the user attention is pied or it's still there when it comes to e commerce shopping. Or if I like to play sports that I might be interested in buying, let's say, a table tennis rackuet all of that has to work seamlessly and systematically in order for a user to purchase a product from your platform that ultimately leads to revenue. Now this requires a sound understanding of technical systems because all of this information needs to flow seamlessly and you have to make, you know, some analytics or some assessment out of it. Strategic thinking, again, is important because you have to a product manager has to think from A, a business and end user perspective and translate those requirements to technical and machine learning stakeholders. Third is strong communication skills. Um, some communication skills are anyway important, but it becomes more important here because the kind of stakeholders you would have a data science person will be more talking in terms of their terminologies that they use. For example, an accuracy recall precision. These are the terms that they use on a daily basis. A product manager is expected to understand that and communicate accordingly, right? And when they're talking to business stakeholders, some of these technical jargons cannot be put in the business meetings. It has to be translated into a world that they understand. Increasing trends that are becoming popular and hence the need of AI product manager. First, this increased focus on explainable AI, like I said. So as we all know, that machine learning works or blends as primarily being a black box that you put in one input and you get an output, and what is the reason of that output is a complete black box, right? Now, people want more explanation of why this input has come given or why this output has come, if I have given this input. Input means that as a user I go on TikToOput means, why did I get this recommendation? Explanation is becoming increasingly important A to improve your decision making, as well as to your external stakeholders. You need to communicate what exactly is happening in the product. Second is growing interest in AI driven personalization and automation. Systems are getting more efficient, systems are getting more personalized. Personalization, again, is important to achieve maximum efficiency or maximum revenue potential. Hence a quick interest or growing interest in AI driven, especially when in the world where there is open source chat GPT, you can create applications on this big data model that has been created from an open source point of view. Increasingly, the servers are becoming cheaper or the servers are are working at a scale we can actually crunch humongous amount of data. So all the other external factors also are adding into the development or growing interest around this field. Or lastly, but not definitely not the least is ethical practices around AI and responsibilities. So as the world continues to evolve at a very fast fast pace in terms of development of these technology, we have to be very ethical about the use of this. And as a product manager, even as a business stakeholder or CEO of a company, you need to make very effective calls to not compromise on the ethics of the use of AI. That basically means you cannot, you know, work upon the human vulnerability to buy a certain item, right? There has to be like stops at various points in the product that they are ensuring that the ethics are not breached. And especially now in this world, if it's a black box, and hence, transparency it becomes important and with transparency, better management around ethics on the continuous use of AI becomes even more important. So these were, you know, some of the skill sets or why AI product manager is becoming more relevant these days. These were some points catering to that, and people can pursue their career in this direction. It is a highly, you know, sought out field. As the technology will keep growing from 6. AI/ML Product Manager: Details : S one, let's discuss some of the responsibilities that is different to a machine learning product manager from traditional PMs, starting with technical skills. So these PMs would need a solid understanding of artificial intelligence and data science in general. A, to be able to work with data science and engineers and also to assess when you ask a data scientist to create a model, what it really means, in which cases a heuristic model would work better than a neural network model, right? What is the scale of data that is needed before even you can propose a data science model? How does the transition from simple heuristics to manually crunching data evolve into a more sophisticated model, right? How to use once a model is made, how do you make out what this model output is? So if a data scientist tells you that the accuracy achieved with this model is 95%. So firstly what does this accuracy mean? What does it? What does the recall number mean or the precision metric mean? Uh, and in different industries, it means very different things. So let's say, one of the classic use case is to detect a disease from imaging, right? So if you conduct an X ray exercise with a certain certainty, you can tell that, you know, this disease is, um, this disease the probability of occurring this disease is very less versus very more, very less versus more. Now, when you say that my model is 95% accurate, this means 95% of the times, my disease is predicted accurately. Now, what about those 5% of the times, right? Now, this is a very critical flow. A person, uh you know, or a doctor working on the accuracy of this model might not be highly accurate for those 5% of users, the repercussions are really high in this case, if you fail to identify these 5% versus something that is not very critical. For example, my recommendation module works, 95% of the accuracy. Now this number, the same number has to be dealt very differently in both these industry. One should understand the true meaning of these numbers while working with Theta scientists and use that number in accordance in the flows. Second is data analysis. Again, it becomes more important for an annual PM, because most of the decision making, whether it's on recommendation, whether it's on the accuracy numbers, whether it's for different use case, a lot of this decision making has to be done on numbers. So A, what numbers are really important in terms of business context, in terms of user needs, and also just to have a handle on understanding big scale data numbers. Third is strategic thinking. Again, it is common across all product managers. But here, since you need to form that bridge of converting technical understanding to consumer and business needs, strategic thinking in the reverse route, whereas you are able to identify business opportunities. Example, if there is an increasing trend of, let's say, a squid games that is getting popular on Netflix. How can you utilize that into an AI product that can lead to more views or more eyes on your platform? This can be done easily, let's say, with an AI model or an AI product. What are the other trends that you need to analyze and feed it to your product? What is the feasibility or what is the TTL? Of introducing such an exercise. Hence, the feasibility or technical understanding becomes really important. As well as strategic thinking in order to convert a business need or a growing tread into an executable or an actionable insight, again, becomes important. Ethics, like we discussed with growing concerns around the ethical practices of AI, it is the responsibility core responsibility of a product manager to ensure transparency, fairness, code proper code of conduct, regulations being followed as per the government standards. Success metrics, like we discussed, now the world does not no longer believes in just black box way of thinking. They need to be supported with more robust arguments to make out what basically an output means. AI integration. Again, it's a more technical nuance, but the integrations that are needed across in the products, which requires components like personalized alcoholthms. Like I said, that a user based on this historical context, need to be shown different things than another users, prediction engines. That means, um, even predicting for a user who has been an avid user on Les TikTok, who has purchased four different items through that platform. What is the prediction of the lifetime value a user can provide to the platform through AI modules? They can actually calculate that if I have spent let's say $100 up till now, on Tik Tok by buying products. What is the lifetime value for this user? This again, becomes important to do budgeting properly to predict revenues and PNL properly. And lastly, conversational interfaces. Now, with many different, um, outputs or formats of AI becoming popular, which is, let's say, voice assistant or text based search or text based conversation. AI is no longer limited to just recommendation based on images or text. It has come about following or introducing itself into different channels can benefit the same product. So identification of such opportunities as well as integration of those channels into your existing pod. Now, these were there some of the difference in responsibilities to traditional product managers that ML product manager needs to have.