How the Bundesliga adopts an agile and cross-functional data strategy

Hello, everyone. Welcome to our session today. Really excited to be here with you. And along with my colleagues, Toby and Bastian, really appreciate your time.

I would like to start with the question, how many sports fans do we have in the audience? Let's raise hands, how many soccer fans do we have? Ok, quite a bit good audience. Just for the disclaimer, we are going to be interchanging some of the wording. Uh we are coming from Europe, so we may call it football. What we are really referring throughout this presentation is soccer, so not get confused on that one. All right.

So we are going to start with how we at AWS work, work with our sports customers. How do we help them to address some of the industry trends and challenges that we see? And secondly, we will focus on our partnership with Bundesliga, Germany's premier football league. And then um thirdly, we will talk about um the data strategy, how we at AWS um approach, creating an effective data strategy, cross organization. And then finally, we will hear from um DFLC uh the digital sports CEO Bastian to talk about how they approach creating cross functional data strategy in their organization.

So my name is Elif Togan. I am senior engagement manager and professional services lead for Bundesliga. And which means that I get to work with um sports customer like teams leagues and others. Um right holders innovate on AWS. So let's get started. Hope you get a little bit glimpse on how excited and passionate and enthusiastic about our job at AWS.

We do a lot of cool stuff with our sports customer. But before we dive in, let's take a few steps back and remember what happened in 2020. Yes, I'm gonna take it there. Uh the outbreak of Covid has changed pretty much everything. Our social interactions were restricted and um our work environment has completely shifted and changed. So is the sports industry, the fans couldn't go to stadiums to see the game, they have to watch it on tv and uh at their homes, no social interactions and um the most famous athletes and talented players had to play their games in an empty stadium. The remote consumption and interaction has become the new normal.

So what we see here is some of the key industry challenges and trends that we work with across all sports customers at AWS when I was growing up, maybe like some of you in the audience pre social media era. Um we would either go to the stadium to see the game or we would wait and at home in front of tv to come up uh your favorite game that you want to watch. But now there are so many ways to consume data and uh really content through many delivery channels and services like mobile apps, social media and uh web websites. This changing the behavior has created a new complexity for providing continuous access through many services and devices. This is probably an area that most of your uh yourselves can relate to.

Um there are huge data and increasing amount of data sets. When we talk about data in sports, we are really focusing on three areas, fan data, radio, data and performance data. This explosion of data creates a huge opportunity to create an effective data strategy uh and create an actionable, meaningful insights for your business.

Like the third part is the legacy media and broadcasting. We see a lot of disruption there because there is an increasing demand from um customers to get more personalized content to do that is available anywhere anytime on any devices. Now you need to in the past, you will need to produce only one feed. Like when you produce the game, you will send it to your broadcasters. But now you have to produce multiple feeds that will go to international broadcaster, domestic broadcaster or advertisers.

We also see a lot of demand from our customers to meet with their sustainability goals. Our customers are looking for ways to um understand their option, utilize technology to understand their option and uh track their progress and report on their progress.

We AWS as partners and our customers. You may be familiar with our approach. We always start with customers and work backwards from that. We always seek to understand what is it our customers needs and their desired outcomes. In order to create solution, many of the world's leading sports organizations are working with AWS, AWS professional services and AWS partners to d to design, deploy and uh deliver um innovative solutions that are reshaping the sports industry, whether it's predicting the next probable outcome of the next big play, optimizing ticket prices after a winning season or improving on field performance technology is changing the game.

So how do we work? A AWS help our sports customer, we accelerate innovation. AWS is changing the game of uh sports by and helping our customers to solve through business problem by accelerating pace of innovation and changing how we collect, analyze and leverage data for fans leagues and teams.

We increase efficiency and return of investment with the last area of services and future. AWS makes it easier, faster and more cost effective to uh more move existing applications in the cloud and build nearly anything you can imagine. AWS helps eliminate it, uh eliminate that cost and align demand better with the it cost and um reallocate resources to higher value projects much faster.

We elevate fun experience. AWS helps sports organizations um to deliver high quality, low latency, personalized content that is accessible anywhere on any devices by transforming the entire match day experience both in venue and online. Um we help our customers to increase engagement and deliver meaningful outcomes for their customers or fans.

We turn data into insights. We do that by increasing uh by helping our customers to design and deliver meaningful, impactful and actionable insights.

So let's look into our partnership with Bundesliga next. So, you know, sports is a very emotional product and in Germany, uh the love of football or soccer uh is overwhelming and it's often called as Germany's favorite child. And uh Bundesliga, Germany's premier football league, um pe they people go see in stadium, more than 20 million people in every season, they go physically see the games in the stadium. There are about 1.1 0.5 billion people interested in Bundesliga all around the world that are following actions, 32 million.

So uh more than 32 million uh fans are following on social media and sharing their enthusiasm about Bundesliga. Our partnership with Bundesliga started in 2020. And since then, we've been innovating with Bundesliga together with AWS.

Bundesliga is using a iml analytics compute database and storage services from AWS in order to deliver real time statistics that improve insight into the game and uh help predict outcomes and recommend personalized match footage for fans across online streaming television broadcasts and social media channels.

So let's look at some of the innovative projects that we achieve together. So I'm gonna be focusing on what we've done so far. And later, you're gonna hear some of the exciting stuff that are coming up.

This is a Bundesliga interactive feed, without a doubt, the top tier products, sports products will be more immersive and personalized in order to meet with our uh fans and customer demands, the interactive feed um that you just saw what the experience, how users can interact is an award winning product where fans can interact with footage through an overlay, they can switch between games, they can select teams and players that they like to see specific stats. And this is pretty revolutionary and they won award with Bundesliga.

We build real time statistic called Bundesliga match facts um to create a better game experience, of course, for fans and fans love it. In an international survey. 95% of fans said that they have a better understanding into the game and 88% said that they uh know the the details of the game, they understand it uh better.

So how do we do that in every Bundesliga stadium in Germany? Um you'll have about 16 to 20 positional, the cameras that are located in the stadium, they sent positional packages 25 of the players of the position and which is about 25 times per second for every game. We collect about 3.6 million data points through those cameras. Once we collect those data points, we process it through serverless architecture using uh serverless architecture, using like AWS lambda and other uh AWS services to create these analy uh the these statistics, we also use a w uh Amazon stage maker to um to the uh produce these analytics.

As of today, there are 15 Bundesliga match facts. As you just saw, there are some that is go probability and attacking zone, set piece trad and so forth that are being broadcasted globally in every Bundesliga game without a doubt a i is gonna play a very important role in media production and consumption as um we are getting more personalized. And also we are looking for ways to really utilize a i to create automated comment, automated translations or maybe producing the game through the uh robotic cameras that are automated, recorded around the stadiums.

So that's already happening. But I'm not going to focus on that. What I like to do is to talk about the consumption piece. So what you see here is the official Bundesliga app, which is one of the most important distribution channels for advanced analytics with millions of users plans can leverage Bundesliga app to um get an additional insight.

Um and they can customize their experience accordingly, they can get insights during the game before the game. And after the game, what you see on the left is um the the team forming. So this is like a one aspect of the app and then what I'm about to show on the right is the match facts that you just saw how we display, oops did it again, um how we display.

So in addition to advanced analytics, we also worked in a very strategic um project with Bundesliga to increase the user uh interface of the app and experience of uh of the Bundesliga app. As you know, football is a highly emotional product like other sports. And i personally struggle and that i don't have a favorite team. I usually support like 34 teams depending on the season. And I want to know every time my teams are winning, losing or they're scoring like someone scoring against them or they're scoring a goal.

So what we find out even though people support their favorite teams, they want to know at least three other teams that they want to follow. So on average fans are interested in following like four teams. And what if you do the math in bundles together? There's 36 clubs. So there's about 1.4 million combinations that club combinations that people wants to see.

So I don't really care about the rest of the league, but i wanna know like those four teams. So on top of that, there are also demographic difference, regional differences, user preference, what you want to see uh and what you want to know about those teams.

So what we did is we use Amazon personalized to collect all that data and process it to analyze and then deliver a personalized um recommendations in the Bundesliga app with the help of um some of the a b testing, we wanted to see the effectiveness of this approach. And um what you see on the right is some stats, there was clearly increasing engagement from the fans and um there the fan were more engaged in the app and uh through this personalization effort that we did together.

Ok. With that, I would like to invite um my colleague Toby to the stage who will now focus on how we had AWS approach. uh creating a data strategy.

Thank you all so much for listening. So, are you facing many challenges when you try to create and implement the data strategy in your organization? And you can't find the right prescriptive guidance to help you with that. I'm here today to walk you through an agile data strategy method that is easy to adopt. My name is Toast and I'm a solution architect at AWS

Based on a survey from PwC. Highly data driven organizations are three times more likely to report a significant improvement in decision making. And that's just one of many surveys out there that show that it is valuable for an organization to be data driven. So the goal is clear, you want to have a data strategy, a data strategy that helps you become more data driven in your organization.

But many organizations already try to be data driven. And a subset of them also has tried to create a data strategy before but still over 60% of organizations report that they don't get value out of their data. So how do you create a data strategy that actually improves your organization?

Let's have a look at some of the common challenges that we see when organizations try to create a data strategy. At first, the organization see a high upfront investment, they see a lot of effort they have to put in before they actually get something out of their data strategy.

Next, they create data strategies that are not aligned to the current business objectives. So they are not working backwards from the goals they have right now. And last most, mostly the data strategies are focused on technology and they don't focus also on the other important domains that you have to tackle in your data strategy.

So the question is what is actually a good data strategy and how do you get there? A good data strategy should have the following five characteristics at the baseline. A data strategy should accelerate your organization in creating value out of data. And the responsibility of your data strategy is to accelerate you. It's not the responsibility of the data strategy to actually create the value.

Next, you should always work backwards from your current business objectives so that you make sure that your data strategy actually helps you in achieving those goals. The data strategy should also be agile because the context, the business context, your capabilities and your goals, they are changing all the time. So you should also have a data strategy that is able to adapt to those changes.

The data strategy should also span multiple domains. It should also tackle mindset people, processes and technology. And lastly, the data strategy should end in aligned actions for your organizations because it's very common that you have many teams involved in being more data driven and you want to have those teams aligned across their actions.

So now we know what are the characteristics of a good data strategy. But how do you get actually there? How do you accelerate your organization? You should improve your capabilities and you can start by the following, you can bring all the teams together and then you're gonna assess everything you have right now. And also understand where do we want to go to in the future?

Then you try to align all those teams and the whole organization to create one strategy that is also focused on the whole organization. And lastly, because this is very exhausting, this should last for three years. You already see there are many downsides with this approach.

At first aligning many teams in an organization that takes a lot of effort. Second, the teams that are involved in that they don't see direct outcomes from this because the data strategy is in this case focused on the whole organization and it's not focused on a single team. Also the context as i mentioned before is changing all the times. So it's not really realistic that you have a data strategy that lasts for three years.

So we need to have a different method, a method that helps us to prioritize the capabilities we want to improve next. And that is also aligned to the current objectives. So let's start at the team level. This time, we just start with a single team and this single team, it's gonna have some business objectives right now.

Working backwards from this, this team can envision ideas in domain of data, what they can do next in direct alignment to their current business objectives. At this point. it's optioning that you can also invite an expert from AWS so that this expert helps the team to understand what is actually possible. What are the next ideas, what we could do over time because the team will repeat this, they will be trained to do this on their own.

And based on that, based on the ideas that are based on the objectives, the team's gonna select capabilities they want to improve next. And for this, you're gonna need a catalog and this catalog can be the cloud adoption framework for a i machine learning and generative a i that i would be talking more in detail in a second about.

But what's important is that this cycle, the scope of this is just multiple weeks, it's not years. So actually, in each cycle, the team's gonna select a small number of capabilities they want to improve. And through that, they're gonna see direct short term outcomes of the data strategy that is also aligned to the current business objectives. And this will lead to just the higher motivation of the team to participate in this data strategy. And it makes it also easier for the teams to integrate this into their day to day business because they can now integrate that.

For example, if they run, if those teams work in edo scrum sprints, they can add that to their sprint planning and could take 1 to 3 sprints as the typical cycle, the typical length for one cycle, as i mentioned before, one option for a catalog of capabilities that you can choose from so that you actually know, ok, what are the capabilities that i could improve? Next is the a cloud adoption framework for artificial and challenge and machine learning and generative a i.

This framework gives you a description of 51 capabilities across six domains, naming business governance, people, security platform and operations so that you can select capabilities from this catalog for at your team level, what you want to improve next. Additionally, this framework also gives you an opinionated order of capabilities so that you can use this as an inspiration. What could be a good next capability to improve at the team level.

So now we have a data strategy method that works at a team level. But how do you scale that to the whole organization? The good thing is as it starts at a team level, we can now incrementally adopt this data strategy method. So we can select which teams to start with and also add teams later to this method.

Then you're gonna set up a meeting at the beginning of each cycle where a representative of each team is coming together to talk about which ideas did they envision and also which capabilities they want to improve by that, they can have sharing of knowledge between the teams. But more importantly, they can understand where do we have dependencies between the teams because still some team might want to improve something where they are dependent on another team. And this is the meeting where they are going to discuss that

At the end of the cycle, you're gonna have another meeting where each of the teams are sharing their outcomes so they can share the capabilities, they have improved the outcomes as they what they have planned as ideas. And this is gonna help other teams to be inspired to also go ahead and select the next capabilities to improve.

You can also add some gamification at this point because if you select the catalog at the base for that and you have like 51 capabilities, you can set goals like how many of those capabilities have to be improved until one of the teams is getting something special. And by that, your organization is going to accelerate over time because each team is improving their capabilities. And you're gonna add more teams to this process over time. And you can even measure that because in each cycle, a team is at first creating ideas and then it's selecting capabilities over time. Those teams will already reuse capabilities, they have improved before.

So there will be cycles in the future where a team might not be, might not need to improve a capability before they actually can directly go to their ideas and create what they want to create. So that's something that you could measure even now you might say, well, but there are some capabilities, for example, setting a north star for the organization where the organization wants to be in the future in the domain of data of being data driven that is typically done by a management team. And that's right. And the management team, and this method is just one of another one other team like the others.

So it's also one representative of the management team that's going to join into the alignment and into the sharing outcomes meeting. So that the management team can also understand when the other teams have a dependency on them. For example, they say we need a north star and then this management team can also tackle that in the next cycle.

Let's have a look at an example of this method. Let's say we have a website team and a website team has the current business objective to create a more personalized experience based on that they decide. Ok, we want to do our idea is to have a better segmentation of our users so that we can adapt our website to this, to those segmentation.

Next. In the method, they go to those catalog of capabilities and they understand, ok, we already have a tool or another team has a tool that we can use for this. So we are fine with that, but we don't have a process that how we work with that data. And we also lack the skill in the people that they actually know how we can analyze this data for the the skill part. They just decide, ok, we need to go to a course, we need to learn something about how do, how can we analyze the segmentation data.

But for the process part, they want to have something that they can use on their own to go deep and analyze the data. So they want to have, for example, in bi tool that they want to use. So they go to this alignment meeting. And before this cycle, there was a different team that was responsible for providing this bi tools. But this team decide, ok, this is something we want to do in the future and we want to keep that.

So we want to have a self service bi. So they gonna go to this alignment meeting and then they're gonna tell the other team, ok? We now have the dependency that we would like to have a self service portal for doing some bi. So now the other team can also take that with them and plan that into next cycle to provide, uh provide a self service bi tool for them.

And then those teams can simply go and build what they have planned to build and share their outcomes at the end. And this will over time then be repeated and accelerate your organization and being data driven.

Now, let's hear from Bastian Soba, the CEO of D Digital Sports about some insights about what Bundesliga will do in the future based on our partnership and also based on this agile data strategy method,

I just need to drink some start. So hello, everyone, good to see you. It's hard to see you because there's a lot of light but happy to happy to have you here on board today. Um and let's move on.

So before we move into some details, as Toby just said, i would like to elaborate on two topics which might not be too obvious to everyone, but which we believe at the DL, which is the governing body of the Bundesliga in Germany, comparable to the NFL for football in the US.

We highly believe that those need to be addressed if you want to have excellent outcomes. And if you want to be successful in implementing a strategy, be it an overall strategy or be it a data strategy? And I'm talking about culture maybe not too obvious for a lot of you guys, but we highly believe that culture is very, very important.

And i me personally, i truly believe in that quote from Peter Drucker. And we also believe in this one which somehow resonates in our data strategy and in our culture. So we truly believe in this one without data, you are just a person with another opinion. And this is more or less the outcome we have achieved now after 67 years. Here is the remaining transcript formatted for better readability:

And let me give you some examples of what we have done to what i believe is a successful culture now.

So the first thing we have done seven years ago is we have put the importance of data into our corporate values. So most of you may have some sort of corporate values in your companies, right? Mostly they are on a billboard in the elevator and no one cares about we do it quite differently.

So we have put that into our corporate value and we live that day by day to enhance that what happened as well. Six years ago, we have put the increasement of our data capabilities end to end as one of our five key strategic priorities.

So it's not just the billboard and the corporate value that we have, we have put it as one of the five key successes and strategy goalposts into our strategy. So this gave the entire team a clear understanding that this is not talking about data is not for the data guys or it's for the product team, right? It's for everyone in the organization, everyone has to deal with that to foster that we have implemented data driven frameworks.

On the one hand, for example, the OKR framework, right? Those of you who are familiar with, OK, when you do the key results, there need to be measures in their objectives, which is data, right? And to get to this data, which is somehow a hypothesis for the future, you have to analyze data.