������

Kimberly Boyd: Welcome to Pragmatism in Practice, a podcast from ������, where we share stories of practical approaches to becoming a modern digital business. Today's episode is brought to you live from New York at our executive conference Paradigm Shift. I'm your host, Kimberly Boyd. I'm here with Sanjeevan Bala, Group Chief Data and AI officer of ITV, and he's joined by Danilo Sato, Head of Data and AI Services UK at ������. Sanjeevan and Danilo, thanks so much for joining us today. You're both onstage at Paradigm Shift, introducing our audience to data mesh and your transformation journey. I would love to kick off and give our listeners a little bit of a brief introduction about yourselves and the work you're doing together.

��

[00:00:43] Danilo Sato: Thanks for having me. I am Danilo, I'm the Head of Data for UK. I'm working with our clients to help them solve their complex data problems. ITV is one of our very nice clients, where we're applying the concepts of data mesh in practice. I'm really proud to be on stage with Sanjeevan, to talk about the story.

��

[00:01:08] Sanjeevan Bala: Hi, I'm Sanjeevan. My role at ITV is very much being around, how do we help the organization use data to drive value, and in driving that value, how does it then drive organizational growth, in terms of revenue growth and EBIT improvements? A lot of the work that we've been doing together with ������ has been around firstly, identifying a lot of those use cases for how you might drive value from data, but then also thinking about how do we then operationalize some of those things. What does it mean in terms of organizational structures, and what does it also mean in terms of the technologies you might then use? That's very much the work we've been doing together over the last two years.

��

[00:01:46] Kimberly: Sanjeevan, can you share with us, what was the impetus for the transformation journey at ITV? What was prompting the organizational buy-in to say, “This is something we need to invest in and embark on?”

��

[00:02:01] Sanjeevan: I think globally, what's happening in terms of media consumption is the way in which consumers are watching content is changing quite radically. Historically, what used to happen is, you'd watch something on linear television, and then if you'd miss that patch, you then watch it on catch-up service. Now, as all of us know, there are so many now direct consumer services, which means the way in which consumers are watching content has radically changed.

��

That was one of the first external triggers. The second point for us was a philosophy from our CEO around how do we as an organization become much more data-driven in what we do. That was the other organizational impetus, which meant the two forces, both an external factor and an internal driver, then that gave us the remit, if you will, to then start to develop a data strategy at ITV.

��

[00:02:53] Kimberly: Were you a fairly data-driven or data-oriented organization before this began, or was this really going to feel like, wow, this is a wholesale change from where we are to where we want to be?

��

[00:03:04] Sanjeevan: No. If you think about companies like ITV, at our core, we're a content producer-distributor, so we're a really creative organization. When you think about creativity, you think about data and science, the two don't quite go hand-in-hand. What was happening in the organizations, there was a realization that we need to find a way that we can bring both of these worlds together. It wasn't an either/or, it wasn't like algorithms were going to take over, it was very much how do we bring the best of creativity with the best of machine learning and science. That's really the approach we've taken at ITV, this augmented idea.

��

[00:03:42] Kimberly: Actually, I'm afraid we think they seem very counter areas, but there's such a powerful opportunity when you bring the two of them together. Maybe we can dig in a little bit and talk a little bit more specifically about getting to that ideal end state of being that data-driven organization. What does this look like? What is the envision that you had in mind? What were some of the first steps that you took to move towards that?

��

[00:04:13] Danilo: When we started looking at the data strategy- and this is all props to Sanjeevan and his team. When they engaged with us, they already had a good idea of all the value cases of where data could improve their business. What we've done together was really try to map out their journey. What are the things we want to do first? What are the first data products we're going to build? What's the first area of the business we're going to engage? That is really what drove our roadmap. The first area that we've been involved is the marketing team and the idea of bringing data mesh.

We're not trying to do it all at once across the business, we're pretty much trying to- using the use cases to drive the value and unlocking the benefit with the marketing team. Sanjeevan can talk a lot more about the first results, I guess, from this first use case, but we're really looking at that use case-driven approach to drive the development along not just the technology build to support that, but also all that organizational change and the operating model as well.

As they were building their data function, how are these data teams going to work if we're saying, it's more federated approach where data is owned and used by the business, so how do we deploy the teams and the expertise so they're closer to the marketing team or the commercial team? That was one of the key things that we took into account to plan this journey.

��

[00:05:46] Sanjeevan: I think just to add to that, the approach we took that it was really important, that had to be business-led, the transformation program. What that meant was, the way in which we invested effort resourcing was roughly around 70% of our time was around the change agenda, about 20%, around the technology, and the last 10% was the algorithms, the artificial intelligence aspect. The reason that was really important is because we were really keen that this had to stick in the organization. It wasn't going to come in like a firework really loud and fast, then disappear just as quickly. We needed this to actually stick and become part of the DNA and the fabric of the organization.

That's why we took our time to think about the change agenda, the organizational structure, the operating model, where do you put all the different roles, and then only we got into the aspects of the technology.

��

[00:06:37] Kimberly: Was that something that was a learning from past transformation efforts, because I have a feeling that most folks don't start out and say, "70% is going to be focused on change." They think where the hard part is, is really the tech and the algorithms.

��

[00:06:50] Sanjeevan: Yes, absolutely. I think I've seen a number of transformation programs where they focus a lot on the data, or purely on the technology. You often hear, "We need a new platform, we need a new cloud platform." Or the other thing you often hear is, this idea of this single source of truth or a single place where we can get all of our reporting, and that becomes the outcome. There's a real disconnect, actually, because that's a technical requirement.

What we've tried to do is really connect it to the last mile. If you have that in place, what are you going to do with it? How do you create value? How do we make sure that the resources in the business are actually doing something with it? That's why the whole change agenda becomes really critical for this to become a successful transformation program.

��

[00:07:34] Kimberly: I was just nodding along with those hearing you say all that because countless times we've talked to folks who are on a variety of these digital transformation journeys. Without a doubt, it's always like the tech or the data ends up being the easy part, and then the people and the communication is always the challenge. I think you guys got that mix right. I would love to understand, is there anything that has been a little bit of an "aha," as you've been on this journey of adopting data mesh in the organization and really embedding it as part of this broader change program?

��

[00:08:05] Sanjeevan: Yes. I think the "aha" for us actually thinking it through now was probably dialing up the employee experience of change. All too often, I think what we've observed organization is, you can often run these programs where for the organization and colleagues in the organization, it can feel like change is happening to them, rather than if you walk in their shoes, and you sit on their table, how's it for them, how's it going to feel for them? We really dialed up the whole employee experience of the change agenda such that we could understand what it's like if you're a commissioner or a scheduler or in marketing and really understand their pain points, and that's how we tried to get the change to happen.

We very much focused on the happening within the organization, within the units, rather than a centralized unit pushing the change out, which I think historically, certainly from an ITV perspective, hasn't necessarily always been successful. It's coming back to my earlier point around, we were really keen that the change had to stick. It wasn't going to be like a firework. That was a key piece that I think was our "aha" moment and drove the way in which we focused on the change agenda.

��

[00:09:08] Kimberly: Got it. Was that giving them actual examples of, "Here's how your world's going to look different," because I think if you were to probably to-- I'm a marketer myself. If you came to me and said, "Hey, we're going to roll out data mesh." Great. What does that mean? That might be a little theoretical. What is federated computational governance? I'm not quite sure, but if maybe you told me how that might enable me to work a bit differently, I have to imagine I'd be a little more on board with the adoption. Is that a bit of what the change program was?

��

[00:09:37] Sanjeevan: Yes. We're a content business, right? Say to quote Fight Club, right? The first rule of Fight Club, [laughs] is you don't talk about data. You don't talk about tech. What we did in our business, we didn't even talk about data mesh, right? We didn't talk about federated and decentralizing, those things. We purely had the conversation around, how could this help marketing? If we were to build a model that allows you to predict viewers that are likely to churn, how might you then activate your marketing differently? If you were to do that, what would that then do in terms of our corporate KPIs?

We had all those conversations with the business. Which then helped us understand in marketing language, what does this mean for you? Then we only got into the technology and the data and all the other-- It was very much a distant second, third or fourth. We literally didn't open with data or technology in any of our conversations with the business.

��

[00:10:25] Danilo: I like the way you described to me the other day, that we see a lot of data transformation being about data literacy programs within the business. How do we make everyone understand data a little bit better? The approach that Sanjeevan is championing is almost the other way around. It's like you've got folks with data skills, how do we do business literacy for those people so they are able to have those business conversations? I know you've been quite keen, should we say, like, let's not talk about data mesh. Let's not talk about source aligned data products. We're doing those things, but when we're talking to them, that's really not the major, most important thing.

��

[00:11:03] Kimberly: The secret for driving data literacy is to never talk about data. [laughs]

[crosstalk]

��

[00:11:07] Sanjeevan: It's true.

��

[00:11:08] Danilo: [unintelligible 00:11:08] you can do with it.

��

[00:11:09] Sanjeevan: Often we challenge our team. We do the so what, now what test. You've done source aligned data products, or you've done a decentralized federated model in the governance of data mesh or so what, and now what does that allow the business to do? Our teams, because they've now trained to think about the so what, now what, when they're engaging with the business, they then describe the now what scenario. Not the data mesh, not the source aligned data products, not the API or the microservices, or low latency data or anything like that. They really focus on the now what, and therefore they start to describe it in business language and business outcomes. That's what I think captivated the organization.

��

[00:11:47] Kimberly: I love that. So what, and now what. It keeps you really focused on why you're doing this and not getting sidetracked in all the background details perhaps that are enabling it, but not the primary focus of what you're trying to achieve. Curious so- obviously marketing was the first use case for what you put into practice here. Have there been any learnings around what use cases are ideal or not, or optimal for trying to implement this type of change or is it really truly how you frame it and making sure you have the change program in place and the use case itself is less important?

��

[00:12:23] Sanjeevan: We went down this approach of co-development with marketing. I think we had our initial hypothesis of a set of use cases and value cases that we loaded into the program and secured funding for. I think as we started working with marketing, I think we realized one, marketing objectives often change, and secondly with what marketing was trying to achieve, how do we make sure that everything we are doing actually connects all the way through to the last mile? What I mean by the last mile is everything we are doing in the mesh needs to connect to a CRM system or some outbound customer data platform.

Technologies that marketing can then activate audiences with and actually do something with it, that was the key thing because otherwise it felt like it was sitting in a silo. They had all these different models they could work with, but actually, they need to do something with it. That was a big piece that we focused a lot on. Almost the, what are you going to do with it first type of conversation, and then figure out how we might use that. That was one, the second area I think was we probably underestimated the level of training and upskilling and development that was needed.

Within our marketing department, you've obviously got some marketeers that are very digitally literate because they come from a digital search background or optimizing paid media, but also we've got marketeers that are brilliant brand marketeers. They do some amazing promos for ITV. In that conversation as well, how do we help them better understand the role data might play in what they do? That's where we tried to take a more holistic approach across marketing to understand how could it play a role for all aspects of marketing. That's much more of a holistic co-developed approach to strategy.

��

[00:13:59] Kimberly: When you're embedding it in across the entire value chain of marketing too then it feels more relevant. Like we were talking about, you're not leading with data you're leading with, what's the impact for brand? What's the impact for your paid efforts? You talked a bit about- obviously it was very much like a co-development effort with marketing on this. Curious what some of the other partnership pieces were that were involved in this effort and how you maybe made the decision on what partners to work with as you were on this transformation journey.

��

[00:14:34] Sanjeevan: Working with ������, one of the big questions, as you get to any of these data strategies, is which cloud provider to work with, and interestingly in ITV, we work with all three of them. We've got Azure running, we have Microsoft and we have Amazon and Google. What ������ really helped us understand was actually, you see this pattern happening in the market where a lot of clients work with a single cloud-first, then they go multi-cloud. Then they come back to single cloud. You have this sort of decentralized centralized sort of pattern happening.

The way we worked with ������ to assess this sort of understanding, actually, where do you have core abilities or competencies in your organization? Where have you got skills? Where do you have already services that are wrapped around a particular cloud provider, that if something falls over at three in the morning, it automatically kind of resurrects itself. What we found was actually the way we configured our business, one particular cloud provider we were quite strong on because we had a lot of services, we had teams in place and that was really what helped shape decision.

What was really interesting as we went on that journey, it wasn't about the technical aspects of the cloud provider. Going back to my earlier point about technology, it was much more about the operating model we already had in place and therefore acknowledging that operating model saying, "Well, that's your primary cloud provider, you should start working there." Don't get me wrong, we still maintain our other relationships, because I think as we all know, one cloud provider can be better on language processing, a different one can be better on vision. We've got the flexibility to still use the different components of technology but we've landed on a primary one because of the way our structure was set up.

��

[00:16:07] Kimberly: It's still that common theme of the model versus- permeates even when you're thinking about what partners you want to have with you. If I were to have the ITV marketing team in this conversation with us today, what would their kind of take and perspective be on this journey? Like, what benefits would they be able to speak to since embarking on this data-driven transformation that you're leading them on?

��

[00:16:38] Sanjeevan: There's been like a cultural shift in the organization. The whole idea of having a common set of customer segments and a common set of customer kind of language in the business is one of the things the program has delivered. For example, across our marketing teams, commercial teams, commissioning teams now, you'll internally at ITV, you'll hear them all talk about the socializer, which is one particular segment that we're going after in terms of our customer thinking, and why that's really powerful because, in the past, everyone had a different definition of a customer in the organization.

For the first time now, we'd be able to pull all those together, and we've got a single unifying way and language we describe the customer. That's the first thing that's been a massive benefit, because you can now have a conversation where the whole business understands what you're trying to achieve.

��

[00:17:23] Kimberly: You're talking apples to apples.

��

[00:17:24] Sanjeevan: Exactly right. The second one is time-saving. What used to take marketing about three months to pull together some data, create a segment and then look to activate that segment, promote some content on Facebook or Instagram or TikTok for example, now takes some three minutes. Now that saving is huge for marketing because what it now means is we can now start to learn a lot faster. With every campaign we send out, we have control groups, we can understand what worked, what didn't work, what do we need to change? While the campaign is live, in-flight, marketing can now dial up or dial down the wait of advertising, the wait to marketing in response to what they're seeing.

That agility is something that we just haven't had the ability to do in the past, and increasingly, given a lot of our views now more digitally first, it's really important we can have that capability to flex really easily and start to dial up and dial down our marketing activity in response to that. That's been a massive gain for us organizationally.

��

[00:18:25] Kimberly: I have to imagine that has to be incredibly powerful because just think about how quickly content comes out these days, how quickly it can go viral, you want to be able to respond to it. Like you said, shift that mix and being able from three months to three minutes is definitely I think a big boost to what they're trying to do there. I'm sure they're big fans, even though that maybe if you would've told them, "Hey we're taking you on a data literacy journey," they wouldn't have known they were going to be this excited about it.

I bet they're all on board now. Now that you're further into this journey and perhaps thinking of listeners who might be contemplating doing the same, what are some of the key learnings, that advice that you'd have for individuals who are focusing on data transformation, implementing a data mesh type of mindset in the organization.

��

[00:19:17] Sanjeevan: I think as we look back I think there's probably four key things that I would certainly suggest you look for. The first is absolutely be laser-like focus on your value based outcomes. That then gets the business talking about- in the language that they understand, but it also means that you get the backing of your CFO because fundamentally you're driving revenue growth and growth in the business and you really need that kind of backing. Absolutely be laser focused on value based outcomes. The second then is what I probably call the golden ratio, the 70, 20, 10.

70% of your efforts focus on the change agenda, 20% is your technology and the 10% is the AI aspects of what you're trying to achieve and all the complexity of the modeling. That ratio means that you focus on the right amount of change in the organization, certainly for a traditional business going through a transformation program, that's absolutely the right mix and the way you need to think about this as a program. The third is that classic the so what and now what, but it's around the how you then deliver. It's not the what you deliver, but it's how you deliver.

This gets into the employee experience because if you're just pushing all this into the business, you've not really sat in their shoes and think about- actually if you're in their shoes and this is coming at you, what's that experience like? How do you think about that and really dial up the employee experience of change? That's the third aspect I think. Then the last one is I think culture. What's been really interesting is we spent time looking at where the organization had come from and what's the DNA of the organization. At the end of the day we are a creative producer distributor. All too often you hear in these transformation programs that suddenly these companies proclaim they're going to be a data company or a technology company. It's like, really?

��

[00:21:00] Kimberly: You're entirely changing what business you’re in? [crosstalk]

��

[00:21:01] Sanjeevan: You're going to change everything? I think it's that understanding your where you've come from and the culture of what you've got, and then think about how do you enhance that? How do you co-create with what you've got, because those are your core competencies as an organization, so you don't throw it all away. How do you complement that and- almost a fusion if you will, and that's probably the fourth aspect of all this.

��

[00:21:23] Kimberly: I love that, because I think we hear about that often. People are like, "Oh we're a tech company now." No, think about how you want to fuse technology into your company now, but perhaps not totally change and become a tech company. Danilo, from your perspective, being a partner to an organization on this journey what key learnings or advice do you have for folks?

��

[00:21:46] Danilo: I think this focus on value is something that I don't see with every client. There's still a lot of people thinking about- even like with data mesh we got a lot of people interested and it's usually about the technical aspects of it. How do I build this self-served platform? How do I implement this governance? We are doing those things with ITV, as you heard even from the conversation today, it's not the main topic of the conversation. That's part of the how we're doing it. I think that is definitely a key learning and I'm actually taking that learning with other client conversation that I'm having because this 70/20 rule, it's not something that people are asking about.

They're always focused on, but how do I do this? What technology do I use to build this? Or even worse, which product do I buy that will give me a data mesh like that? That's completely the wrong kind of conversation, and I think it was definitely a learning that just validated working with ITV, that's definitely not the focus that people should take.

��

[00:22:45] Kimberly: Yes, it's what business value do you want? Not, here's some data mesh. Final question, what's the latest content that'll be coming out from ITV that we should get really excited about? If you can answer, or if there's anything that as an American might be hitting our streaming platform soon that we could check out.

��

[00:23:05] Sanjeevan: We did bring you Love Island USA right?

��

[00:23:07] Kimberly: This is true, this is a gift that keeps on giving.

��

[00:23:09] Sanjeevan: I think it's fair to assume that we'll continue giving

��

[00:23:14] Kimberly: Fantastic. Well Sanjeevan and Danilo, thanks so much for joining us today on Pragmatism in Practice. I really enjoyed our conversation.

��

[00:23:21] Danilo: Thank you.

��

[00:23:21] Sanjeevan: Thanks a lot for having me.

��

[00:23:23] Kimberly: Thanks so much for joining us for this episode of Pragmatism in Practice. If you'd like to listen to similar podcasts, please visit us at thoughtworks.com/podcasts. Or if you enjoyed the show help spread the word by rating us on your preferred podcast platform.