I'm really excited to talk about, you know, how to convert data to value and really demystifying what what are data products and what's data product management.So this is something that is close to my heart. And, just a little bit of, you know, introduction about myself. So I am currently a senior director at Capital One. I work on acquisitions, decisioning, and data products. I have been working at the intersection of data product and financial services for over fourteen years, and really started off in data as a database developer, sort of transitioned into product management, and then, have really worked on building consumer facing data products as well as internal data products, ever since.

So very excited about this topic, and I, you know, hopefully, we can we can answer some questions and have a little bit of discussion as well as we go through it. So, really, I would like to really start off by, talking about why is this topic so potent and important for, you know, organizations today. And I don't think it is lost upon us that over the past few decades, the organizations have really generated immense amount of data, and that is, continuously growing as well. And this is also the importance of data has accelerated, over time due to also the compute that has been provided by, cloud technology to that has given us the ability to handle massive amounts of data at scale and process massive amounts of data. And now with Gen AI as well, we have the ability to create intelligence out of that data in a more sophisticated manner and in a more easy manner. So as a result, you know, data and what we do with it has become so critical for organizations as that has become the key foundational element to help us gain, intelligence, within the organization.

And also this data is really the strategic mode for companies if it is used in the right way. And in order to be able to extract this value out of data, and, you know, it's important for it to be, managed as a data product. Managed as a product, thinking about, you know, what the business value is, what the user value is, and really, work on it in a way like we do with any consumer facing product, to in order to make it more reusable and scalable, within the organizations. And so that's why this the concept of data products and data product management is so critical, especially today, you know, with Gen AI as well, because that is what is will help, organizations create strategic differentiation, as well. So moving on to talking about what is really your data product. Right? I think we all have come across data products within our organizations and also in our day to day life.

So these could be ranging from, you know, your curated datasets to a dashboard or to recommendations that are being served to us through recommendation engines. And, really at the heart of it, it is you know, and this is obviously a definition over here. But what I really want to critically underline is those data products are really those products where data is guiding the development of its functionality. So as an example being a product where which is highlighting any kind of risky transactions. If at the heart of it, you know, it's really dictated by whether or not you have the data, which gives you the ability to identify the risk of a transaction. So when you think about data products and, you know, and and trying to identify whether a product is a data product or not, really sort of bringing it back to whether or not the data is critical to develop a specific functionality and to deliver value, really helps sort of create that distinction.

However, I would argue that now at as we're kind of moving towards more intelligent products and more intelligent organizations, all of our products are sort of moving towards, you know, a paradigm where, you know, they are all going to be highly reliant on your data. So as a result, data products become extensively, you know, more more and more critical if organizations are making this journey towards getting more intelligent. Let's take a look through this. Right? So I think this is, a helpful way to think about data in an organization and also about how data and data use cases increase in maturity. So, obviously, data products are existing in different forms in every organization. And depending on maturity, if you think about your own organizations as well, you will see, you know, where where do you have, like, high value use cases and in what forms are data products being adding value, within your own, you know, organizations.

So it really starts off with your raw data and transactional data, kind of then being transformed into your sort of your curated datasets of your data products, which then are further sort of put together in different, ways to create your analytical products. So these are your dashboards, your reports. And then if you further go down you know, further kind of think about the maturity, then it further transitions to creating more automated insights, recommendations, and predictions. And these, again, can be delivered. These data products can be delivered. These are all different form factors of the data product, but these can be delivered in various modes. So you would see see that these data products could be delivered based on your, you know, use case in forms of APIs. It could be via an insights platform. It could be as a service as well, or it could be embedded within your applications.

And this is really helpful to visualize how your data is power powering different applications, within the organization, as well as what is the maturity of usage of your data. Right? And it is also helpful to understand where in this value chain is your data being used and what other ways in which you can maybe evolve it, for new use cases. Okay. The other thing as well, I think, you know, it's obviously helpful to visualize, right, where, data products have really, added value and see some examples of, you know, various companies that have really done this well. So, again, data products can add value both, in terms of providing direct value, as revenue, as well as indirect value for helping you improve your products, creating new products, reduce costs, or even, you know, help you with strategic decisioning.

So in terms of where you know, some companies that have done it really well in terms of direct monetization that I've seen personally as well has been at, you know, Visa and Mastercard, which have leveraged, you you know, payments data really to create new data products and insights for their, you know, financial institutions and partners, and have really sort of, benefited directly from creating data products.

Similarly, Walmart Luminate is a great example as well. They've also done a, you know, a great job of leveraging data strategically and creating insights for their suppliers on top of the, you know, commerce data that flows within their ecosystem. And then when we talk about indirect monetization, I think, you know, we've seen how companies have created strategic advantage, with likes of, you know, Netflix and Amazon using data to create personalized recommendations and really create that, you know, the stickiness with their customers by improving their products, by, you know, really understanding their customer and personalizing the experience through data, for them.

And another great example that I've always thought of when it comes to direct monetization is the example of Starbucks, which has used a combination of or has been using a combination of location data as well as social media data to determine, you know, which are the best locations for them to open new stores.

And that is a great example as well of how companies have leveraged data internally, to create, you know, differentiation and, get indirect benefits out of it. I think, you know, this is there are quite a few considerations when you think about converting data to organizational value. But I'd like to highlight in today's conversation, you know, five key principles, which, are quite critical when you think about, you know, how do you scale this value, and get value out of data. So the first one, similar to building any products, it's really critical, to start off with your business goals and business strategy. And, and especially so because data products and creating data products and data initiatives, they do take relatively longer, and they're not cheap. So it's really important to understand as an organization, where do you want to head with your data.

And having this sort of top down approach, of, you know, starting with your strategic goals, it help it really helps because it also helps you get buy in across different organizations where the data product may be creating value. You know, it's being created, but not really, you know, it's it's being reused in different, forms. So it's really helpful to kind of get the buy in and make sure that it is delivering value. Because if it's not delivering value, it's, you know, it can equate to, you know, fast and painful deaths for your data product. So it's really important to do that. The other component of this is it's also important to understand where as a business you're headed. Right? Whether you're headed towards creating more, whether your goal is more internal value generation or is it more external value generation?

Because the shape that your data products could take both in terms of the form factor in which it is delivered, the scale at which it's delivered, and the teams you need to deliver it could vary based on those two goals. So as an example, external data products may require more go to market teams. They may require pricing teams as well, and commercial, teams as well to in order to make them successful. The second principle over here, again, I think this is a little bit, I mean, it's very intuitive that you do need to keep your user at the center, at the heart of building your product. But for data products, especially if they don't have an interface, sometimes this can get, you know, this is not does not get enough emphasis. But it is pretty critical as you're building data products to also keep your user at the center, whether it be sort of, you know, human user or a system user.

It's it's important to kind of keep that in mind, because it can really influence different factors for your product. So it can influence the form factor in which you're delivering your data product, whether, you know, it's an API or is it a dashboard, is it a report. It can influence the level of extraction you want to have for your data product, you know, at the level of detail that you want your data exposed. So as an example, you know, an operational analyst may require more, you know, a higher level of detail than, say, for example, a business manager who may have the requirement to really sort of extract the the details and really provide the insights. So that really varies as well. So which is why keeping, you know, your user at the center of even building your data product is so critical, because all of these different factors another one is also frequency at which you want to deliver your data.

That could also vary based on the user. The other thing as well that I found is that, you know, doing this also helps you uncover a lot of other use cases and new problems to be solved. Sometimes you go in with one assumption, but based on conversations, you may find other net new users of your data, that you haven't really thought of before. So that's that which is why, you know, this is another critical factor, when you're considering building your data products. The third one is an extremely critical one, which is, you know, you know, have building trustworthiness for your data product set. And trustworthiness is sort of an amalgamation of different, components. And each of these are important as you're thinking about building your data product.

Because if, you know, people don't trust your data or your data product, they will not use it. And it takes one breach of, you know, one data quality issue. And then regaining trust in your data product becomes very difficult. So it's important to kind of build your data product and making sure that, you know, you have these key components in mind. Some of these are, you know, explainability, making sure that, you know, your customers and your customers and your users can understand what the data means, what an insight means, how has it been derived, and building that into your data product through different means. Ensuring that, you know, there are any kind of privacy guide guidelines that you may have within your organization are maintained. Any sensitive data is, you know, sort of anonymized or removed from the dataset, or you're getting consent from your users before you're incorporating that in your dataset.

The other one being here that, you know, was both calling out is obviously data quality and accuracy, ensuring that, oh, you know, your data and your insights are error free. They are fresh. They are relevant, to the context of, you you know, your user as well as the use case that you built the data product for. And, you know, obviously, making sure that, you know, the security and access around the dataset is, only the use users that are authorized to get access to that data, are getting the appropriate level of, access as well. And then depending on your industry, you know, the compliance with regulations, there's obviously a lot of different, regulations depending on the industry you are in that, you know, data products need to comply with, GDPR, HIPAA, PCI compliance. So that would vary obviously based on, you know, the industry you're building your data products. The next two, essentially are talking about the first one really is about building and scalability and performance as you are thinking about your data products.

And this is also because for data products, it's important to think about, reusability. And I think a lot of you would, you know, agree with this and may have already seen that in your organizations. You will find that there are multiple pipelines that are kind of extracting value from the same dataset. And this is a common scenario in organizations. And what that leads to is really sort of obviously, it leads to inefficiencies because you're, you know, wasting your resources and your technical resources as well in building those, you know, redundant pipelines. But what, you know, more importantly, what it results in is, you know, it, it leads to different definitions and understandings of the same data elements and potentially even different types of decision making, based on the same data. So it leads to economic inefficiencies and then leads to, you know, higher cost, for the organization.

So one thing that is also important is to think about, you know, how your, data product can be made usable across multiple use cases. And this also ties in with sort of the prior point around making sure that, you know, your data product, gets, you know, is aligned with sort of your business goals and your business strategy, which helps in making it reusable as well. And making sure that it can scale, is important because as you're solving more use cases, you want to make sure that, you know, it is built the performance is not degrading over time as more and more use cases are being added onto your backlog and your roadmap, and more users are, you know, sort of using your data product for for to serve their needs.

I see we have only five minutes. So I will not break the point about continuous and holistic measurement. I think that's pretty critical as well. But what I would really want to, you know, also call out is that, you know, what we spoke about in the beginning was, you know, how we're with Jenny, I right? We all organizations are thinking about how can we build intelligence into our products and to our organizations. But all of this, you know, this unless the foundation is robust, you will build a very shaky sort of house, or data you know, your kind of intelligent organization strategy. So Gen AI, in order to be successful, also needs reliable data. And as we're thinking about, you know, what are some of the strategic data assets within the organization to get to move towards an intelligent organization.

This obviously is something that, is pretty critical to evaluate. And then the other aspect as well over here is that, Gen AI is going to enable, you know, new types of data experiences with natural language UX. And it as it becomes the interface of choice, data products are really sort of behind making those interactions intelligent, accurate, and actionable. And so eventually also I feel that data product managers will also need to think about, you know, how to design not just for sort of the analytical use cases, but also for some of these, models of interaction where you're not just designing for screens, but also for language, and making data products that are sort of trustworthy, contextual, and natural language ready would be also one of the important factors to consider.

With the three minutes that we have, I just want to leave behind an emphasis on these four, aspects, which is basically, you know, products are becoming intelligent. So as a result, data products, quality data products, quality being sort of underscoring that, are really going to be the foundation for organizations to be able to move towards that paradigm. I think as product managers, there will be a new skill set that would emerge, which would be combining product, data, domain, as well as aspects of, you know, how do you think about building responsible data products? So we will see sort of, you know, people in these different domains can kind of bring together, these skill sets. And then, you know, thinking about building strategic data assets in a very intentional manner, would be would be critical for, you know, people in the organizations as well as, you know, thinking about, embedding trust, privacy, and governance. But, yeah, I mean, I that's some of the thoughts that I wanted to share in today's conversation. It's very less time.

It's a huge topic, but, hopefully, I'd love to connect and happy to, connect with folks that are interested in this topic. You can find me on LinkedIn. But, yeah, thank you for joining. I see a few questions. I'll try to see. Yeah. That's a good question, Jitka. Maintaining, maintainable models. Absolutely. I'm happy to share some thoughts, but I think it goes back to, you know, sort of getting alignment about, you know, which are those, key models that you want to maintain within your organization and what are some of the use cases. I think it's really about building that business case to show, that these are, you know, things that are reusable and getting kind of buy in, that you don't want to build that over and over again, and you want to use those specific ones that you've identified. So, yeah, we can we can connect, and we can talk a little bit and share thoughts. Resources for deeper dive onto this topic. There's so many different resources here. That's a great book, Ashley, on data product management. You can definitely refer to that. It's escaping escaping my mind right now who's the author.

But, you can look it up. It's unpacked. You can speak you can refer that. So it's a great reference. But, honestly, like, a lot of this, that's more solely on kind of data product management, but there's other components like, you know, governance, etcetera, which are a bit more, you know, they're not included in that. But, yeah, hopefully, that this is a good good starting point for you. Great. I think that's my time. Thank you all. And, yeah, love to connect. Thank you.