Data Governance and Groove with Tiankai Feng

Krithika Anand Hello and welcome to an all-new episode of the Retail Podcast by Vue.ai. You may have noticed that our intro to the podcast sounded a little different today. If you were nodding your head or you still can't get that beat out of your head, We have our guest today to thank for that. When you're already impressed. But here's a little bit Tiankai Feng, Head of Marketing at DAMA, Germany. With over five years in data analytics and data governance at Adidas, he carries a wealth of experience in data analytics in the retail sphere. After spending 10+ years in data analytics, data governance, consumer insights, and digital transformation, he's now grown fond of the human aspect of data. How to communicate, collaborate and get creative around data. We're certainly amazed at how he's making data fun and we're very excited to talk to him today. Hi Tiankai, It's lovely having you on our podcast today. Welcome. Tiankai Feng Hi Kritika It's very good to be here. Thank you. Krithika Anand Great. Let's get to know what you do better. Tell us about how you started out with data analytics. Where did this passion stem from and how did it lead you to data governance? Tiankai Feng Fan of advertising and was good in math. So when it came to that time to choose what I wanted to study, I chose industrial engineering, which was a mix of marketing and database systems. And from there, after having done my Bachelor’s and Master’s actually slid into the area of marketing analytics and had a chance to directly work in China for four years, really working on topics like web analytics, social media, analytics, consumer insights and all kinds of different things to explain the marketing world better with data. And I spent a little bit more time also in that same area back in Germany, and one of the clients I had was basically Adidas and really moving on then into the client side and the corporate side of things where I continue to work in data and being first, a product owner for a data product, then leading the social media and voice of consumer analytics capability. And lastly, then going into data governance, specifically on the domain of product data governance. I think what I always realized was that although the analyzing part of data was always exciting and if you didn't have the right quality of data upfront to actually do the analysis, then you didn't really have that much opportunities to do anything with it, right? So typically calling it the garbage in garbage out principle, right? If the data is garbage, then whatever you do with it is not going to make it much better. So being and having been a big advocate for data quality and it was very exciting to be able to actually join the space of data governance and to be able to see behind the scenes a little bit about how data quality actually is being achieved and how the data is even being made available for analysts and other data consumers to use. And so, yeah, I'm very happy that I started more from the analytics side and analyzing things to now also realizing how to manage and govern data in the right way. Krithika Anand Wow, that's interesting. And it's been quite the journey, I suppose. It's fascinating how your journey with data went from a key focus on analytics to now accessibility and making it more approachable for the people. Right. Where did this pivot begin? Tell us a little bit about that. Tiankai Feng Well, I think I would say the whole idea behind why I enjoy what I'm doing is really coming from the space of wanting to make data valuable. Right. The whole idea is that we are anyway collecting so much data in the world and anything we do because we live in a digital world, that it only makes sense that we make the best out of it. But if you don't analyze it or you don't manage or govern it properly, or you don't make it part of decision making, then it's really worthless, right? It really only sits there and is not being used properly. And everything I'm doing is following that guidance and that kind of approach that I just want to make sure that data is being used and that we all have a little bit more trust and also fun with data. Krithika Anand Oh, wow. Fantastic to hear this. And I've always thought of this, right? People fear numbers and most people are uncomfortable with data. The mere thought of analyzing data and data driven predictions, can all be confusing and overwhelming for some. But your passion for making data more approachable and fun is commendable. Tell us this. Well, there's so much discourse around data today. There's also misunderstanding and chaos around it. What are your thoughts on unifying governance for data and analytics? Tiankai Feng Yeah, it's a very good point. I think, um, first of all, there's a general reservation or even fear of data, right, as you mentioned. And the more you are not in a data team or not on paper at least data a responsible team, the more you might shy away from it. And part of the reason is probably that we as data professionals have hyped up the area of data so much that we are so proud of it, that we are kind of excluding certain people, that they now feel like it's a very scary place to be in. It's very hard to get in and only experts who deal with data, but that is rather the opposite because in the end, we all need to actually work with data. And the reality is that we are dealing with data no matter what, right? And in a certain way, if you think about any organization just sending emails or working on Excel files or creating PowerPoints, all that is creating and using data as well, right? And that might be unstructured data, but in a certain way we are all transforming information and communicating information to each other. So it's not too far fetched to think about it that it should be better managed and governed, right? The main purpose of governing data is in the end to make data meaningful, secure, and usable for everyone. Right? And that should be in everyone's interest. And to make sure that that has happened, we need cross-functional collaboration and data governance to make that come true, right? Everything else in data governance, like having the right tools, having the right processes, having the right responsibilities. That is all part of it. But it all has to come from the same understanding that we need to manage data together to make sure that it's usable and valuable for the organization. Krithika Anand Right? Absolutely. And I think I'm going to touch upon this point that you brought up earlier. How important is it and what would it take to maintain a compliant end-to-end view of your data estate with a single model of data governance for all your data, both structured and unstructured? Tiankai Feng I think, generally speaking, you first have to agree on what compliance means, right? Because compliance means there are certain rules in place and it's even the most critical part is to even come to those rules that all of the stakeholders in an organization agree to what the rules are, because you might even have conflicting requirements. Those rules, right? If I want to have table A in this format, but another one wants table B, a table A in this format, then we already have a clash, right? So coming to those rules is number one step and then getting and rallying up everyone behind those rules to say those are the rules that make our data most correct and usable. That is the next very important, but also very difficult step. And basically to make sure that we are all having transparency and an overview of data and also keyword data lineage, for example, or data observability. It's all about working together as a team because you have the data management data governance teams of course in place, but they are only helping to facilitate certain things and you still need subject matter, expertise and different experts from the different functions to actually manage the data operationally and strategically. So in the end it all comes down to a people process and tools, right? You need to establish data governance end to end. You need the right people that feel accountable and responsible for the data. You need to have the right technologies in place to manage data, to harmonize data, and to check data quality. And you need to have the right business processes in place so that we can all hold each other accountable for the different elements that we have and to make sure that we are managing data correctly, Right? Krithika Anand Absolutely. And thank you for those points. Those were really insightful. And I think I'm going to take a different spin on this conversation right now. I came across a piece on creative coalitions where you've taken on the challenge of applying principles of design to analytics, and I'm very intrigued. Can you tell us more about it? Tiankai Feng Absolutely. And so that goes back again to the human side of data that I love so much now. Right. And that you already mentioned also in the intro, because we when we work with data, it's really the key really is that we turn those data into insights and that these insights are then actually being turned into decisions, right? Better decisions than before. And so it comes to the point when a data analyst will analyze something and there will be some findings and some insights, but that analyst has to convince a stakeholder to actually trust those insights, to believe in the insights, and then act accordingly, based on the insights. And at that point, you realize that actually, it's all about convincing stakeholders. And that is very much a parallel to being a product brand, for example, that tries to convince a customer to buy the product, for example. And in marketing or in anything actually related to people, there are design thinking principles that have been established to be an effective and efficient way of developing things and making things more customer-centric, right? And try basically try to really mix design thinking with data work. So and that means in design thinking, you always start with understanding your customers and audience and empathizing with them, right? So instead of starting with analyzing numbers first, we should actually first understand how our stakeholders make decisions, right? When are they making decisions? What do they think about when they make decisions? What are certain milestones where certain decisions have to be made up front? Once you get to all of these answers, ideally in talking to your stakeholders upfront, the better your insights will be because then you know in what direction to work on and not just based on assumptions that you have. Tiankai Feng So that applies to any other data work too, right? As a data governance person. Of course, I'm also thinking about top leaders and subject matter experts, data consumers as my audience and as a data scientist. I try to think about also my business counterparts because they are the ones who are going to use my algorithms and models. For example, as a data engineer, I want to make sure that my analysts, that my stakeholders find the data that I've created usable. And so it really applies to any element of data work. And yeah, I'm really trying to advocate for it because I think it's going to make our work much better. Krithika Anand Absolutely. And I think what you've really done with this is, you know, bringing different teams together to sort of build that cohesive story around data and storytelling itself. And I think this is. Very, very fascinating. On that note, everything we spoke about, you know, sort of highlights the importance of streamlining your data ingestion and management. Right. But how can businesses successfully do that? And where do you think I can help out here? Tiankai Feng Yeah, it's a very good point. AI is all the hype right now, right? I think, though, what we have to realize. So first of all, let me talk about the importance of streamlining. So basically to manage data well, we have to avoid data issues, so to say, and as early as we can avoid data issues in the data lifecycle, the better it will be because then we are curing the root cause of data issues and not just fixing the symptoms, so to say. And all that starts with data ingestion, right? If you put the right rules in place and you harmonize in the right way when the data comes in in the very first place, then you already can tackle most of the data issues up front and everything else afterward might be only smaller technical bugs that you have that you could fix ideally. And the thing is I first of all have to work with the right databases to be trained to work right. You need historical data so know what it's actually supposed to do. And if the data that's being fed to AI and that is being trained with is not good, then I will also learn the issues of it and it will think that the issues are right, which means it would amplify and recontextualize all of the wrong data as well. And that is really not good. So what that means is if we want to actually make AI work, then we have to first establish kind of a training set and almost a role model and an example of how correct data looks like. So I can actually support us to make data better. And lastly, I would also say to that point that there are certain elements where you think about AI that could autocomplete data input as well. Right? But it really depends on how the data is created in the first place, because in many cases you still have human beings who type in data, right? That's for example, a product manager who creates a product for the first time in the database. But you have also an automated system that has greater data, like for example, in retail, when you just create a transaction, right, you scan something, it goes into your system and then you basically create that transaction in a retail store. And that is automatically created data. Ai can help in both, right? It can either complete a little bit the human thinking, or it can have quality checks on the automatically generated data. But I think the future is brighter. So as long as we don't lose control over completely and completely, 100%, completely 100% trusted and don't check anymore, we should just always have a point of view on it and just observe it, but we can definitely start using it to be more effective going forward. Krithika Anand Right. And I completely agree with your point. You know, it's more like bad data and doesn't go well together. And it is true in most cases. And having said that. Right, the role of here has found its footing across industries and it is certainly here to stay. What are your views on how companies are undergoing their digital transformation journey? I mean, it goes without saying that digital transformation is no longer a question of if, but when. And the sooner companies embrace it, the sooner and smoother they're going to ride this wave. What do you think or what do you feel is the first thing that organizations should keep in mind when they embark on this journey? Tiankai Feng Um, it's a very good point. I think that, first of all, there are many factors that are driving digital transformation no matter what, right? And that makes it not optional anymore or mandatory. So, for example, um, we live in a world now where governments are regulating data, things in things much more heavily going forward, and that will not stop, right? So they're going to put a lot of more requirements on transparency and doing the right things with data and AI, which means as a company you have to be able to be showcased that you are doing the right things. And if you're being audited or being checked, then you can showcase that you are compliant with the rules. And if you don't transform digitally, everything will just be transparent and you will not have anything actually to show which puts you, in the end, even a worst-case scenario into a legal risk area. Right? The other way is that customers and consumers, of course, are much more digital and they are actually the ones who are adopting new platforms and channels much more quickly. Let's, for example, mention threads as the reason Meta launched a Twitter alternative, for example. And it's growing users as fast as never any platform before. Right? And if you are not part of it as a, let's say, consumer-relevant brand, then at some point you are leaving out one of the most important channels and one of the most critical channels. And lastly, um, every other organization is doing it too, right? So if you are not doing it, then you are definitely going to left behind and you are going to lose market share and that will just make you less competitive going forward. So I think the first step to do it right is to understand that you have to do it and you have to understand how can you realistically evaluate where you have to go. Like, what does digital transformation mean for your company? Is it more about data, right? Is it more about capabilities? Is it more about your touch and sales touchpoints, for example, that you have to become more digital? And but once you know where you have to go, you can define a target state that you want to have, let's say, in the next 3 to 5 years, and then you assess your current evaluation, evaluate your current status and you know where you have the biggest pain points and the biggest potential. Right? What is the most urgent for you to accelerate on to actually transform digitally and then basically take the whole organization with you to do that digital transformation together? Krithika Anand These are some interesting points. Right. And now that we're talking about AI, I also want to touch upon a little bit about generative. Ai or generative. Ai is taking the world by storm as everyone starts to see its potential capabilities and what it can offer. Do you think the future of data governance is one where generative AI plays an integral role? You know, navigating data, governance, transparency, and trust in a generative world? What are your thoughts on this? Tiankai Feng Uh, I definitely think that generative AI can help with data governance, but it will not be able to take over the most critical tasks, I believe. Right. So what I'm saying is that a big part of data governance is to documenting things and making things explained right to the rest of the organization. For example, KPI definitions or a data model. And although you can probably let AI help you design things, create models, and write text, you still have to make sure that it's the right one, right? So you have to prompt it even with the right input so it writes to you actually and generates the right content. And that is basically only going to shift a little bit of the responsibility from writing it yourself to prompting it in the right way. But it's not ever going to fully be automated on its own, right? You cannot just wait for AI to automatically generate things that are in your head happening, right? And so that means it might change slightly the tasks and the nature of the tasks of data governance professionals, but that just gives us actually more time to do the more important critical thinking work, I would say. Krithika Anand Great. Great. Really interesting to hear your thoughts on that. Right. I think we've come full circle with my last question for the day. How did you discover rather bring together your passion for data and your love for music together? Um, and maybe you could sing for us here as well. Tiankai Feng I mean, we already sang via the music in the intro, right? But I'm very happy to talk about it. I feel like it really was a lucky coincidence. So, I mean, just for context, I did, um, play a lot of piano and sang. Um, but it wasn't until 2019 that I first started applying my musical skills to my profession as well. And the first song that I wrote about data was called The Digital Analytics Anthem. And it really came from a place of frustration and dissatisfaction because I realized that digital analytics was really misunderstood and in the world a little bit deprioritized and so I wanted to make a statement about how cool it is to work in digital analytics, how important the analytics is, and try to use like a fun genre like rap to do it right. So I did it and I didn't know that it would have such an impact as it all of a sudden had really inside my organization. But even on LinkedIn, outside and on YouTube, and so on, really gaining a lot of traction from people all over the world. And that really showed me that there are so many ways of being creative and making the topics you're working on accessible and fun. And for me, it might be music, but for everyone else it might be other things like drawing comics or telling jokes or all these kinds of things, right? That whatever you feel comfortable with and being creative with. And so another way of looking at it is just to bring your own personality to work, right? Like if you are good in something and you know you're passionate about it, why not apply certain manner, and certain aspects of it to your day-to-day work as well? You're spending so much time at work anyway. Why not bring your authentic self to it? Right? And for me, that just means to do music. And if people outside and inside my organization really like what I'm doing, then even better. So I feel even more motivated to be myself. Krithika Anand Wow, that is super, super interesting. And this was absolutely incredible. Um, that's the most exciting way we've concluded a podcast, to be honest, to know more about you and to understand how you bring two things together, right? I want to thank you once again for joining us here today. We're looking forward to all the fantastic work you do. Tiankai Feng Well, I mean, if anyone is more interested in data governance or AI or my music, even feel free to contact me on LinkedIn. I'm happy to connect and also have a YouTube channel where all my music stuff is there, so feel free to check it out as well. Krithika Anand And that was the end of today's episode. We'll be back with another episode of the Retail Podcast by Vue.ai, where we chat with leaders, pathbreakers, and game-changers in the retail space. Until then, I'm your host, Kritika Anand.