How Data Literacy drives Data Culture – with Steve Prokopiou, First Central

Carsten (00:00.727)

Steve.

Steve Prokopiou (00:02.576)

Hi, Kostin.

Carsten (00:03.96)

It's a pleasure to have you on the Data Culture Podcast because you have a very long data career in a way and have been in your current company now for a couple of years. And maybe you can just guide us through that journey and what you basically did with data and data analytics.

Steve Prokopiou (00:25.762)

You're making me feel very old, but thanks very much. I'm glad to be here, Karsten, and nice to talk to you. I spent my entire career in data and analytics. I've worked for major software companies like Business Objects, more recently Tableau Software, and I've always been most interested in helping organizations and people work well and effectively with data. I've been at my...

current company for a couple of years now. I joined just over two years, I joined September 2021 and my initial title was Head of Data Products and Data Culture. And so there was two elements to the role. The data products part was that to set up a data products approach in terms of how we deliver solutions to the business. And the data culture part was the data culture, data literacy, data fluency.

You know, those kinds of things in terms of helping people feel more capable and confident working with data. So my, my career and my journey at this current company, I work for a UK based motor insurer has been interesting, fascinating, challenging in all measures. And yet I'm happy to share whatever you would like to talk about in that, in that journey so far.

Carsten (01:43.944)

I think it's a very interesting combination, data products and data culture. And you are actually the first person that I speak to in this podcast. And actually that I know that has this combination. So let's, let's talk about both, I would say, because I think obviously data culture is super interesting, but data products is also on top of mind for many companies now, some combine it with a data mesh approach, which has been in the

market only for a few years now, but the idea has been very becoming popular even before the term data mesh or the idea of a data mesh has been developed or described. So I think data products is a very interesting idea and approach. And maybe we could start with you explaining us how you understand data products. So what it is and what's different about them.

Steve Prokopiou (02:43.266)

I will do my best. This is a very interesting question, and I'm not surprised that this is a topic of interest for people because I've spoken to a number of my peers, both in my organization and my network, about well, what does data products mean for you? And the inevitable response has been a sort of scratching of the chin and going, oh, well, the classic, it depends. And it's, well, what do you want it to be? So for my organization, what we wanted it to be was a,

Carsten (03:05.112)

Mm-hmm.

Steve Prokopiou (03:13.106)

a more targeted approach in terms of how we deliver solutions for our business. And so it may be, is not a classic definition of a data product or a data product approach. But what we needed to do in this organization was move from a effectively a ticket delivery machine in the data and analytics team towards a more focused, targeted and business savvy approach to.

producing valuable data outcomes for our business. And so when we talked about data products, what we decided to do was align our team to certain parts of the business. So we had, we ring-fenced a team that was solely responsible for working with our marketing organization. We had another team that was responsible for delivering data solutions for our underwriting team and then another team for claims and fraud.

So by that segmentation and that ring fencing of a leader in that group, data engineers, visualization people, analysts, solely targeted into a certain part of the business, we started to move the needle in terms of more effective deliveries for our business users because they were focused on a particular part of the business.

Carsten (04:39.224)

Mm-hmm. And if you say we, was that like a data team in the IT department or where was that organizationally situated?

Steve Prokopiou (04:49.066)

Yeah, so we're a data and analytics team that report into the CIO within the company I work for at the moment. And so, yes, that's where we sit. And I should say that I'm representing my own views in this podcast, and I'm going to need to keep the Legal Eagles happy in that. What I'm talking about today is really my view of our journey, my personal journey. So...

Carsten (04:53.886)

Mm-hmm.

Steve Prokopiou (05:15.558)

there is, I was asked to make sure that we included that caveat in this session. Um, yeah, but yeah, we report in, we, we report into the CIO who, who then reports who's part of the executive committee here at the company that I work for. And, um, yeah, we built out the team over, over time. So since September, 2021, we, the team I inherited, myself and my director inherited, he, he started a month before me, we inherited a team of eight.

So while we were transitioning to this data product approach, we were also recruiting the team and building the team. And so we didn't immediately go to one data product team per group. We went to where the file was burning brightest and set up the ring fence group for marketing. And then we had the everything else bucket. And then as we went through the quarters and we recruited more people, we created more of the team. So

It's evolved over time for us. So as we built out the team from, I think in that first year we went from about 80 to maybe about 25, 30 people. So we were evolving and growing as we went along.

Carsten (06:26.42)

And if I understood you correctly, you are basically bundling the know-how you need to build a complete data product. So from data acquisition to visualization, maybe building a dashboard on application. Is that correct?

Steve Prokopiou (06:42.302)

Yes, pretty much. And then with the purpose of the ring fence team is that it increased the collaboration with the business to say, look, our capacity in this team. I think we had teams of eight, just for the sake of argument, is we've got eight people and therefore our capacity dedicated to you is X. And so what it did promote was the ability for the business to prioritize where the team spent their time. It wasn't us telling the business what we could do. It was them saying,

This is everything we want to do. And this is what, and then we saying, this is what we could do. And, and, and then the prioritization was driven by the business rather than the, the data team. And, and so we got the benefit as a result of the business feeling more in control of what they, what they wanted us to work on.

Carsten (07:31.668)

Yeah, and in a way they would have exclusive resources, wouldn't they?

Steve Prokopiou (07:36.542)

Exactly, they had exclusive resource dedicated to them to do what they wanted, within reason of course.

Carsten (07:43.336)

Yeah. So they didn't have to go into prioritization meetings with other departments who gets the resources first and so on and so on. So I completely understand that should be very popular with the business. And, and, and just a question here. So what often I see companies thinking about joint teams, they say, well,

Steve Prokopiou (08:00.411)

I'm sorry.

Carsten (08:11.9)

One of the biggest problems is the lack of domain know-how, lack of know-how about the processes. And we need to include that in the team. Is that something you considered also?

Steve Prokopiou (08:24.138)

Domain is in data domains or business domains?

Carsten (08:27.028)

No, on business domains, yeah, so that you need certain know-how about the business processes and the terms and what's really the problem and so on.

Steve Prokopiou (08:37.17)

That was one of our hopes for byproducts of that engagement, that closer engagement with the team is that we're data experts, the business are the subject matter experts, the domain experts, to use your terminology. And therefore, by the closer integration and discussion, often daily, many times daily, that domain knowledge starts getting transferred over. As you said, another part of my title is data culture, data literacy.

And we talk about the businesses needing, business stakeholders needing data literacy, but I also believe that data teams need business literacy, which I think is where you were going with your question. And so it's something that at the moment, definitely is a learning point, is that more understanding of how the business works would be something if we had the chance to do again, that we would definitely put more emphasis on. Because things get lost in translation, because people, the data guys don't really,

Carsten (09:31.561)

Mm-hmm.

Steve Prokopiou (09:36.118)

don't 100% understand all the time what is being asked for. And it takes a lot of backwards and forwards to define that.

Carsten (09:41.748)

Yeah, absolutely. And data products are often also a way to change ways of working in terms of that you stop thinking about projects and start to think in products with responsibility, a product management, modularization, and so on and so on. Was that also part of your data product idea?

Steve Prokopiou (10:01.346)

Hmm.

Steve Prokopiou (10:10.042)

At that time, that was part of our North Star, is that that's where we wanted to get to. My role recently changed in July, and I'll explain more about that in July 23 on that. And I would say for those first 18 months from sort of January 22 through to July 23, we didn't really achieve that kind of product management approach, the culture of, here's a...

a beautifully curated data product that gets continually refined and improved by the data product team, which is maybe a more classic definition of a data product. That's ultimately our ambition, but certainly in those first 18 months or so, we've never really got there. We did see some improvements out of the more closer engagement with the business stakeholders.

I would say that we tended to still deliver projects even within those product teams. So that's why I said it's sort of a data product, but maybe it's not the perfect definition in terms of what most people would say would be a data product. Yeah.

Carsten (11:11.549)

Mm-hmm. Okay. Yeah.

Carsten (11:22.036)

Yeah, I think so. But anyway, it sounds like a good starting point.

Steve Prokopiou (11:28.962)

Yeah, it was definitely a good starting point. And when you look at where we were when we first started the data product teams and where we were even at the end, 12 months later, at the end of 2022, I think it's fair to say we did make some really good progress and delivered some good solutions, some good projects for the business. What we didn't do so much is the shift in the cultural shift, both within the data team and our stakeholders about, well, what is a data product really all about?

Carsten (11:58.592)

Yeah. Let's shift the conversation also to the cultural topic. And you mentioned now a shift in culture. Was that the reason why this or your title included data culture? Or was it like, was it a new title? Was it something that has a reason why someone said, well, let's get someone that actually looks after data culture. Or was that already in the company before?

Steve Prokopiou (12:27.858)

It wasn't in the company before. It was a new title within this particular organization that I work with. And the reason why it was included in my title is that when I was being recruited with my director, I said, yeah, I'm happy to do the job, but it has to include both of these elements. Because as an individual, I've always been, I think, as I said, in my introduction, I've always been very passionate about connecting people with data solutions, making sure that it meets their business need.

Carsten (12:44.385)

Mm-hmm.

Steve Prokopiou (12:56.254)

making sure that we're not just frisbeeing software out of the door and saying to people, right, there you go, good luck, we've done our job, now get on with it. So the whole way of holding people's hands into using data effectively, using analytical solutions within their role has always been very important to me. And since the term data literacy, data culture got coined, what was it, the last 10 years or so, then.

It's one of those things where I've probably always done it and always been interested in it. It's just that now you have this phraseology, then that's why I wanted to include it in my title, because as an organization, we previously did not have an embedded program of how we're going to help people feel capable and confident with data. And that's why we created the job as part of my title.

or the title is part of my job, whichever way around, and to start the process of engaging a community of people across the business with working more effectively with data.

Carsten (13:50.327)

Mm-hmm.

Carsten (13:59.6)

Yeah. So would you say there was like a need to work on the data culture or to improve it?

Steve Prokopiou (14:07.01)

Definitely, it didn't really exist before in terms of an initiative like any organization. We're a 1300 person organization, 1300 person organization, and we're an insurance company. Okay, so data is at the core of what we do. We have more data than we can believe. And there are absolutely people in the business, even before I came along, who are very skilled at working with that data. You know, but in...

typically like most organizations, those data professionals, if you want to call them, they're 10, 15% of the business. You know, what about the other 85% of the people in the business that work with data, but maybe they don't even think they're working with data, they're just looking at a screen and making a decision. And certainly, so helping the widest population around with their data skills and abilities with data was why that was part of my role.

Carsten (15:04.656)

Mm-hmm. Makes a lot of sense. And now when you started your role, how did you approach it? I would think you would first take a look into the organization and try to identify what might be the most beneficial initiatives to improve data culture. How did you approach the topic?

Steve Prokopiou (15:30.386)

It's a challenging, you want to help as many people as you can, but it's a really good question because you can't do everything. Yeah. And the ambition is to help everybody, but you couldn't do that. I couldn't do that. And so my approach in the early days has been quite a general approach, which is creating learning opportunities and making them available to anybody. So what we did was we

in the middle of 2022, for example, we created a data learning at work week. And so it was a sequence of webinars over a week on a number of different topics where we partnered with our internal comms team to communicate out that we were running this data learning at work week and we encouraged sign-ups across the business to come and learn about

Power BI is our visualization tool of choice. So we did tool specific sessions. We did more general sessions about why data, why now? Why is data important? Why is it important for everybody to be able to speak data? So it's this, we look to cover as much as possible. So to help the data citizens, if you like. So those 85% understand that actually, you know what?

you're a data person, because do you use your phone to look at the weather, you know, to look at the information on your screen and say, do I need to take an umbrella today? You're actually reading and understanding data there and taking an action. And so we did storytelling with data, all different things around the data space and definitely, definitely not just tool-specific webinars. We tried to cover a cross-section.

So I think in that first week, we did about eight or 10 of these webinars over a five-day period with varying levels of sign-up, but it was very successful and at least started the process of engaging people around data. So yeah, so we did that and then the other, that was a very general approach. And then what we also did was a more specific engagement with the Power BI community. We were already using Power BI within the organization.

Carsten (17:36.385)

All right.

Steve Prokopiou (17:51.35)

And so together with my BI team, the business intelligence team, the Power BI experts, we created a community using our collaboration tool, using Teams. We set up a Teams channel and then we sort of published, how-to questions on there, encouraged collaboration from there. We created Power BI doctors. So somebody working with Power BI could book a...

a time, a one-to-one time with an expert to say, I'm struggling with this Power BI question, can you please help me? So we did a general approach and then we did a more targeted approach for the Power BI community.

Carsten (18:34.216)

Mm-hmm. So obviously a focus on literacy, so meaning education. And, um, but I mean, there are other possibilities to, to influence the data culture, like, um, working on the principle, how data is available and also generating more transparency over, for example, available data sources and what you can do with it.

Steve Prokopiou (18:41.314)

Hmm.

Carsten (19:01.876)

Was that also a topic in your organization?

Steve Prokopiou (19:05.374)

It is a topic in our organization and a peer of mine is the head of data governance is was working in 22 and is still working on pulling all of that together, like any organization, we have bundles of data across the organization. So knowing where to go for what data, how accessible is it, et cetera. We've got the classic challenges of.

how to find and use data and get access to data. So it's definitely improved over the last two years. And there's work to go. You know, we're considering, you know, use of things like a data cataloging tool, for example, so that we're more organized and it's easier for our end user community to get access to the, find and get access to the data that they want. Yeah. So that's what's happening on that side.

Carsten (20:04.224)

In all this, how important is the leadership? We call it data leadership. So also encouraging people or allocating resources for data work, not only in a data team, obviously that grew quite a lot. So there were new resources, but also in the business departments. Yeah. Could you talk about that?

Steve Prokopiou (20:21.03)

Mm. Look at this.

Steve Prokopiou (20:29.742)

Yeah, I mean, it's massively important. And when I first joined this company in September 21, I joined as a contractor. So after I left Tableau in 2018, I did contract work. I was freelancing for a number of years. And that's how I came into the company I work with now. Then when the opportunity was offered to go permanent in January 20,

22, I'm trying to make sure I got my dates right, is what I did, well, I think what anybody would do, which is, okay, what's, what's in my favor, you know, in the data space, you know, what, what's working well in terms of a date approach to data for this particular organization. And I've been very lucky to see a lot of data strategies in my time. And I've seen a lot of not so good ones. And I've seen some good ones, right. And

Number one is, okay, well, what's the approach to data? You know, is there a data leader in place? Is there an executive committee that are behind data? And I'm pleased to say at this organization is that both those boxes were ticked in that our whole business strategy is underpinned by data. Our strategy on a page mentions data four times, in it the specific word data.

And I've been around enough times to know that's not normal. That is not, you know, and so that gave me great confidence in the discussion. And the other thing to my advantage was that I'd already worked for a few months within the organization. I could see that it wasn't just words. It was actually believed from the executive committee, from the CIO and the data director, that this really is something that was funded, believed in.

core to what this business is looking to do. And so, you know, there was lots of ticks in the boxes in terms of being in a good position to, the opportunity to create some great value for this organization around data and analytics. And so that's why I signed up, because it was a challenge. There was lots of things that weren't here two years ago. They are now guessing they're now, and we still got more things to do. And

Steve Prokopiou (22:49.014)

when you're a data guy like me is you want to be delivering and creating and building and I'm pleased to say I've got strongly got that opportunity on throughout everything that we're doing myself and the team I work within here at this company.

Carsten (23:03.164)

Yeah, sounds good. Let's spend just two more minutes on this topic of a data strategy. You mentioned a few things that where you said, well, that's a good setup. I understood it's the link to the business strategy and a business strategy that really focuses on data. And you also mentioned the organizational setup. So there is a data leader and you also got the impression.

Steve Prokopiou (23:09.749)

Mmm.

Steve Prokopiou (23:24.375)

Hmm.

Carsten (23:32.268)

that it's not only on paper, but there was, let's say across the organization in the different management levels, there was a, let's say a belief that data is helpful or however you would call that. What else in from your experience makes a good data strategy?

Steve Prokopiou (23:57.006)

That's a really good question. A good data strategy, I go back to what I said earlier, is that when you can see that it's not just words on a piece of paper, and we want to do data, okay, well why? Oh, because everybody else is, right, well, okay, that's not a good data strategy. Let's turn it around and say, what do you need to deliver successful data outcomes for an organization? Right, yeah, you need executive sponsorship.

Carsten (24:07.48)

Mm-hmm.

Steve Prokopiou (24:26.41)

You need a data leader. You need belief. Tick, tick, tick. You need funds. If you're gonna really do it seriously, you need funds to build a data team out. You need funds to perhaps adjust the tooling that's already in place. You need funds to develop people and that cultural change. Because if you don't take the people with you, then your likelihood of success is less. So...

I think I've answered your question in a bit of a roundabout kind of way, but to deliver an effective strategy, you do need that belief, but you also need the dollars or the Deutschmarks or the euros or the funds behind you to actually really deliver that strategy because you can't deliver a value just on a shoestring. That's why lots of CDOs struggle because, well, what do you mean you can't do it all by yourself?

Well, because this just doesn't work that way.

Carsten (25:29.4)

Very good points. Thank you. Let's, let's go back to data culture. You already mentioned a lot of interesting initiatives you took, but maybe you could sum it up in a way, what were the most effective ones? What were the initiatives that worked best? And

Steve Prokopiou (25:33.857)

Okay.

Carsten (25:51.24)

You already feel it. The next question after that will be what didn't work at all. But let's look at both because I think we can learn from both.

Steve Prokopiou (26:00.666)

Yeah, what works best is getting started, doing something. I observed that some organizations spend a lot of time thinking about what they're going to do around data literacy and data culture, and that delays the benefits that people achieve because that thinking time can take months and years. So the reason why I went with a general approach for

to get started was, okay, well, I'm going to put some stuff out there and at least start the ball rolling in terms of engaging with the widest part of the business. And then in addition to that, the more targeted approach was the Power BI community. So getting started would definitely be one thing. Establishing a sense of community around data, making data accessible, making it fun.

is something that has worked very well at this organization and also in previous scenarios where I've done this before joining my current company. And so I think those are two elements of get started and create a sense of community and engage with others. I would definitely say that. And what was the other question? What didn't go so well? What didn't work was my ambition was

Carsten (27:19.668)

Yeah, what didn't work.

Steve Prokopiou (27:25.51)

over above my bandwidth and the bandwidth of the people available to me. So I was very ambitious when I first started this program and I had to cut it back in terms of what I wanted to achieve and by when. So we've made really great progress and actually another thing to do is to measure your success in terms of how many people are you actually touching. So in the first year we at least

We measured that 30%, 3, 0% of our organization and my colleagues actually went through one of our data literacy modules. So from a standing start, that was amazing. That was really, really good numbers, you know, so in the hundreds. And then this year we've moved to 44% of the organization has been through at least one of our modules in data literacy. So being able to measure and report on your success is important.

not just within data literacy, but also in your wider data strategy is that there's always going to be the people that turn, well, what value have we got from spending all these, you know, many, many pounds or euros on this data strategy. So being able to measure and report on your successes and the value that you've added is really important at both levels.

Carsten (28:28.825)

Mm-hmm.

Carsten (28:40.776)

Yeah. What else did you measure? So people participating in the educational programs. I think that's a very good measure. Do you measure anything else?

Steve Prokopiou (28:47.826)

Yep. Yeah.

on the data culture side of things or more generally. Yeah, so we measured things like number of sessions that we ran. So how many community sessions did we run? How many learning events did we run? Either instructor-led training or webinars. Number of people that were engaged. We measured the number of people that registered and the number of people that actually attended to see what our attendance rate was like. We...

So in order to try and get a sense of, well, we've put these opportunities out there. What's working well? What do people really engage with? What don't they engage with so much? So we measured a number of things on that. Yeah. And we use that to inform, you know, what we did next. We asked for feedback on every session, you know, so that again, we could iterate and improve and...

Carsten (29:35.036)

Mm-hmm. And in the...

Steve Prokopiou (29:45.91)

And it goes back to what I said earlier, which is just get started and learning as you go along, because you will get feedback that will help you refine what you're doing.

Carsten (29:55.796)

Yeah. Do you also measure things like how many data products you build or how many people in the organization start to do or to start data projects by themselves? Like, I'm sure you also have a self-service approach with Power BI. Is that also something of interest?

Steve Prokopiou (30:18.115)

We're not measuring that yet. It took a bit of time to uncork the Power BI usage log, for example. So in previous lives, I mean, the great thing about a product like Tableau is that it has an inbuilt analytic capability, which is who created reports.

who consumed the reports, what's the most used report, all those kinds of things. Unfortunately, Power BI, it's not as easy to get to that information. So literally in the last six weeks, we've finally got access to that information about who's using, you know, who's creating Power BI content, who's consuming that content, how often is that consumed, what's the most used report. Because then again, that gives you the ability to go and have a conversation.

with people in the business about what is this report used for? Is it the most performant report? Has it been written so well? Is there a learning opportunity there that you can educate people better on tool usage or even how to communicate with data? So yeah, these are all good inputs to then go and have a conversation with people.

Carsten (31:35.654)

Unfortunately, our time is almost up. So let me come to my last question, which is really forward looking, take a look into the future. So what's next for you or how do you plan to develop this topic of data culture further in your organization?

Steve Prokopiou (31:43.036)

Mm.

Steve Prokopiou (31:52.01)

Yeah, well, that time went really quick. Yeah, so our ambitions for 2024 is that we're going to be move, still continue with our general approach in terms of our use of data across the business and emphasizing how important data culture is for our organization. So, you know, data culture is for the organization, but then data literacy is really for individuals. And so we're going to be creating, in the process of creating data profiles or data personas.

Carsten (31:54.325)

Yeah.

Steve Prokopiou (32:21.102)

to offer more specific learning pathways for specific groups within an organization. So if you think about even within a data citizen, you can have people that are primarily consumers of data and they need to be able to see data, understand data and then take an action. But then you could have data explorers where they look at a piece of data go, hmm, that's curious. I'd like to interact with that data a bit more and be able to ask and answer some of my own questions.

And so they need different education in data to be able to be able to do that. And so we're identifying these data personas and then we're going to be creating specific learning journeys for those personas. So getting more specific and targeting it and intentional about how we hold people's hands into data is where we're going.

Carsten (33:14.336)

That sounds great. So I wish you a good luck with that and all the best to develop data culture further in your organization. Steve, it has been really a pleasure and I hope to speak to you soon again. Bye-bye.

Steve Prokopiou (33:28.132)

Thanks very much, Karsten. It's been my pleasure talking to you today. Thanks a lot.