#66 The Path to Building Data Cultures

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI Transformation Journey on datacamp.com.

Hello, this is Adele Neme from datacamp and welcome to Data Framed, a podcast covering all things data and its impact on organizations across the world. According to the new Vantage Partners 2021 Big Data and AI Executive Survey, almost 100% of organizations report investments in AI and data initiatives, yet only 24% of respondents claim that their organization is data driven, with 92% of respondents stating that the lack of data culture is the biggest impediment towards becoming data driven. This is why I'm excited to have Suleman Thopan Mohanchandralal on today's episode. Suleman is the Chief Regional Data Analytics Officer at Allianz Benelux. During his tenure he has deployed many leading AI solutions in the insurance space and has been spearheading the data Culture Transformation program at Alliance Benelux. He holds two Masters in Computer Science and Business Administration with a specialization in Business analytics and Intelligence. In this episode, Suleiman and I talk about the importance of building data cultures, best practices organizations can adopt when building out a data transformation program, the multi pronged nature of data transformation programs and the complexity that arises therein, the connection between data skills and data culture, and more. If you enjoyed this episode with Sudaman, make sure to also check out his recent webinar on datacamp where he goes into much greater detail on how he operationalized data culture at Allianz Benelux. If you want to check out previous episodes of the podcast and show notes, make sure to go to datacamp.com community podcast Suleiman, it's great to have you on the show.

Thank you Adil. The feeling is mutual.

I'm very excited to have a discussion with you on the importance of building data cultures, but before we start, do you mind giving us a brief background about how you got into the data space?

I have been in the data space for quite some time now, currently leading the data and analytics for Allianz business in the region of Belgium, Netherlands and Luxembourg Reporting directly to the regional CEO. How did I land into a data space is sort of a question that I cannot answer simply because I was always in the data space one way or the other. Started with complex event processing for something called an interbank fund transfer processor, moving into pricing environment analytics and release logistics kind of solutioning for IBM and then continuing to work on data warehousing and then furthermore specializing in business analytics. You can see that the complex event processing, the data warehousing and the whole aspect of analytics itself from pricing analytics, all of this put together a super specialization after my MBA in business analytics and intelligence clearly set my path towards leadership role in data and analytics space. So heavily working into data and analytics would start almost from 2012, 13 etc. All the way until now where I've been trying to either work in a data office or trying to set up a data office.

That's great. The role of the Chief Data Officer is relatively nascent compared to other leadership roles. Can you discuss how the CDO role has evolved over the past few years and what is the CDO's mandate today?

It has evolved to a certain extent that it should be stopped calling us nascent anymore. It definitely was like that when I started to do that role. But now I think it's getting into a mainstream as well. But not yet there I would say. I agree it's not yet there. How does it differ? Obviously I aspect of data becoming an asset and if at all there is going to be an asset, it is always managed by the business. Data as an asset should also be managed by the business. And the all the asset management principles that which applies is kind of guardian by the specific CxO. This asset also has to be. There has to be a guardian and a custodian for this particular set and that is to be Chief Data Officer. So businesses based on the strategy look at data as either a culture or an enabler or a tool, whatever. However you look at it, it is one of the main components that which is going to drive the business strategy in the future. So from that perspective this is a role that which will have to share the table with the board, ensuring that the business is going in the direction as required by the strategy.

That's awesome. And looking back at the time that you've spent at Allianz so far, can you walk us through some of the achievements the team has made in making Allianze Benelux a much more data driven organization?

Sure. It all started in 2018, I would say and now we are in 2021 we have achieved considerable amount of distance. I wouldn't say we have reached the destination yet, because such a thing is not there. So we basically today have delivered at least 30 plus data products. That which is operationalized on the business side, which is a phenomenal achievement in my opinion. And what we have also achieved is year on year we have created operational profits over than what has been set as a target. So cumulative operational profits created by dataoffice, they are in a way structurally underestimated. And I'll explain this in a moment, but just taking that part, we have covered the business case, so we now create more profit situation than trying to cover the investment. So our business case have been achieved. So this is all a great achievement for us, for the business. And indeed I promised to explain you what is the structural underestimation. The way we calculate the value that which Data Office creates is an attribution of the revenue that which the business makes. So every time we kick start an initiative we have a discussion with the business that what is the value that you will attribute to the Data office? And it is more or less far from scientific. It will get scientific in the future, but it is not that scientific. But it is just the attribution of that to the Data Office. So from that perspective, I still think we are structurally underestimated. But even with this structural underestimation, achieving the business case is a fantastic achievement by my team. Pretty proud of the team that I have.

That's pretty substantial. And it's exciting to see how laser focused on value these data endeavors have been. Now one thing I really want to deep dive with you today, which is something that you've been very outspoken about, is the importance of building data cultures. So when looking back at the challenges you face initially in the state of data culture when you joined Allianz, can you walk us through how important was building data culture for you to extract value out of data science at scale?

I think we will never be able to create value out of data if you don't have the data culture. So what do we really mean by this data culture? Right. What do we call as a culture? Culture is nothing but an aggregation of an organizational habit. So if an organizational culture is something that which you express nothing but the regular habit routines of the employees of that organization. Data driven decision making is for an example, a routine which has to become a habit. And this has to definitely replace some other habits which is more instinctive driven. Then in that case you are basically talking about changing the culture. Now if data driven decision making is not there, then you don't really need data. And if you don't make a decision based on data, then there is no value out of any data activities that which you make. So from that perspective, I argue that you will not be able to create any value whatsoever out of data unless and until you have this data culture in whatever maturity level. But at least there has to be this data culture slowly starting to creep inside the company. So yeah, I hope I answer your question. Argument is that you will never be able to create any value of data without this data culture.

I couldn't agree more and I'm excited to deep dive more into that with you. So a lot of organizations are currently investing in becoming data driven and extracting value out of data. There's a range of activities to be done there from culture transformation programs, data infrastructure and technological abilities. Do you find that there is friction between focusing on an organization's technical readiness versus its culture? If so, where do you think organizations should start?

There are really four things, Adil. The first thing is the data culture. The second thing is data solution. The third thing is data enabler. The fourth thing is data tools. So most of the time in an initiative such as data and analytics in an organization is started or kick started, they try to address just as how or where exactly did it start? For example, if it started in the business side, which is always good given the other options you have, one of the four things will get the importance if it is started on the business side. Data as a solution or data as an enabler would basically start getting the importance. While data as a tool or data as a culture remains a little bit adopted. I would say in the same way, if it is kickstarting it, you can imagine that the data as a tool and data as an enabler becomes important. But the solution and the aspect of culture remains as an adopter. So it depends on where it actually starts. In my experience, I would say starting all of these things parallelly is unfortunately the only way you can make progress. The reason is you will have to sustain the investments that which you put. And to sustain these investments you need an outcome that which basically balances all the investments. That's the the returns. And it is also a way to see what is the real value for the investment that which you make. If you want to pivot or persevere on some initiatives, this returns is super important. That's your real feedback. To make that happen. You need some data as a tool thinking. You need data as an enabler thinking you also need a solution because you want to put that on the market and make money. So the only way it can become a solution is when you have the culture to use data in the decision making. So you see one way or the other. Unfortunately, you don't have the luxury of starting it sequentially. You don't even have the luxury of start pipelining it. You have to do parallelly. And that's where the complexity lies.

What are best practices that you found when combining these initiatives? And specifically when it comes to building data culture transformation programs?

Best practices totally depends on the organizational culture. So that by in itself is the first best practice. Understand your ecosystem, understand your landscape, understand your current organizational culture, understand your current business experts. Because you need to look at it from the angle of what is the cue or the trigger, what is the reward for which there is a routine, that which is in place? Because almost always there is a routine, that which is in place. You will have to understand these things. So that's the first best practice. Understanding the status quo, as I would call it. Then you need to understand the destination rather clearly as well. In the sense the destination is not something that which you would say as an ideal state far away in the future, of course that would be very nice to have. But as much as possible, this particular destination should be characterized in a specific, measurable, achievable, realistic and time bound way. So once you have that, that would be a point of arrival. So from the point of destination to the point of arrival, you will have to define rather clearly a kind of an approach to reach there. Now here is where I would say you cannot have one road, but you need to have those four roads. Data as a culture, road, data as a solution, road, data as a enabler road and data as a tool. And as these roads basically yield into either effectiveness or efficiency, that needs to be tracked as well. So which means connecting all of these things to the outcomes. So that would be the third or the fourth best practice. Fifth best practice, in my opinion would be to have a real commitment principle with the business. So you need to start an initiative not from data side or data office side, but it has to be an initiative of the business in which data is kind of plugging in. The simple reason is because the priorities of the business remains, even though you go with the priority of making a data driven. So instead of adding another priority to the already existing list of priorities to the business, my best practice recommendation would be pick one of those initiatives or a group of initiatives which is listed already as a priority of the business and get their acceptance or commitment that they will basically do it data driven and this commitment extends to the outcome attribution and this commitment also extends to operationalization aspects. These are the best practices that I could immediately think of.

One thing you mentioned here is the need to work with business experts and leaders because organizational priorities are essentially business priorities. Can you elaborate on the nature of getting acceptance when you work with business leaders and whether there is friction sometimes when gaining buy in from organizational leaders on adopting new data science methodologies and how to best navigate instilling this mindset shift within the organization?

Technically speaking there is no reason for a friction because you are basically trying to do the same business initiative of the business. You're adding more resources to it rather than creating hurdles. I say technically simply because there is of course the other side of it, that there could be politics at play, there could be other kind of challenges at play, and even rightfully so. The business might be thinking of focus drift, that which is affecting their ability to deliver the target because they don't really understand the data world of it and they don't see it as resources, but they see it as sort of hurdles. So these things are always there and you could call those as friction. But indeed this is where the data literacy part of IT or data education comes into play. You will have to spend some energy in ensuring that they are able to see data as data as a resource rather than as a hurdle. And to do that shift, and like you correctly said, it's a mindset shift, that is the only way you will be able to achieve a frictionless discussion. In any case, there is going to be frictions one way or the other which would be healthy ones because of identification of issues or roadblocks or gaps which has to be filled. So that's why I would call it healthy. But at least a good level of understanding will enable those so called hard frictions to disappear.

Now you mentioned as well here the importance of data literacy and data skills. Can you walk us through the link between data skills and data culture across the organization, whether it's for executives or individual contributors. How do you view that as an organizational priority?

I'm going to take reference to our data academy within Alliance Benelux. We call it Accelerate Data Academy. We divided the data skills into three levels. So to say the first level of attainment is bronze. We call it data bronze. And everyone in the company, irrespective of what they do, where they are, there is a certain level of literacy around data that they should have. And that's exactly what this particular level of attainment delivers. So they, we call them as data citizens. I don't want to confuse the term of data citizens to some other definition. It's just that anyone who kickstarts and has the minimal knowledge that they should require to have about data, in terms of data ethics, data privacy, giving importance to the data, the confidentiality nature of it, and including all the other techniques, the recent techniques that we just come in, data etc. The high level understanding of those things, things, the analytics aspects of it, very high level. And in this literally whatever role that you are playing in the company, you are an underwriter, you are an actuarial person, you are a receptionist, you don't care. Whatever is that role that they are doing, they should be aware of. That's the bronze level. We call them data citizens. The next level is of course silver. And when you are in silver, you go a little bit more deeper. Then there are streams, of course, if you are an executive, you're able to basically understand the data value extraction process. I would put it like that. You are also able to look at business initiatives where data can add value to. I mean you're able to identify them if you're an executive. The second stream of course is business managers, etc. Where they go a little more deeper and learning how to build this cases where data adds value. And then of course the third level is the technical people who are basically building the models using data etc. So at this level we call them data masters simply because they are able to really create value out of data. And in the third level, that's the level of gold. And this basically delivers a certain attainment at the level that you actually have a point of view or an opinion about what is going to be the future, given the current circumstances such as this, in terms of data and analytics or in terms of the initiatives, the business initiatives, which is heavily data and analytics intensive. So this one is at the level where you're able to visualize what kind of an interruption will yield much better or improve or increase the outcome. And also you are able to voice your opinion with facts and you're able to pitch and participate in strategic discussions of the company. So we call them as data ambassadors. So in my opinion, direct answer would be it depends of the organization that you're talking about and the business that they are doing. But not to sound very generic about it, but in an attempt to give you an answer, I think the business should make a clear decision of how it would want to use data, do they really want to use it as a competitive advantage, or do they really want to use it as a kind of defense mechanism in which whatever is the decision that they make, then they can basically use the levels that I spoke about accordingly. And that's the kind of strategic workforce planning that they will have to do.

This is very insightful, especially when tying the business objectives with the learning objectives. Now, you mentioned how data culture is intrinsic to the entirety of the business. As such, when it comes to creating a data culture and galvanizing it, who do you think is responsible and accountable for this culture transformation program?

The board is accountable for the culture of the organization more broadly, and so they remain responsible and accountable for the data culture as well. Now, more precisely, it is the CEO, because CEO is always the responsible person when it comes to culture. The reason being, if there is a strategy, and that's something that which the CEO is responsible for, culture is the vehicle that takes the strategy to its destination. Now, you cannot only have something that which is packaged and ready to move, which is the strategy, but you should also build a vehicle that which is necessary for you to take it at the right time to the destination. And hence the CEO remains super critical. Of course, the CEO will have a CDO who will execute it and will help even to build this vehicle itself for the strategy to move to its destination. But I still think the role of a CEO, none other than, but the CEO himself, is super critical in making data culture a reality in an organization.

Do you think that functional and business leaders should also be responsible for galvanizing a data culture and carrying over that message for the remainder of the organization.

Absolutely. The CDO, along with the CEO will have to ensure that they are enabling the other CXOs and their respective reportees to play a role of taking accountability of this culture. Because you should see the CDO's role here as a catalyst or an enabler role, but not really an action role. The action is always with the other CXOs who are running the business, in fact. But it is the job of the CDO to make all the things that make the ecosystem and the environment suitable enough for them to be able to do that. But the real responsibility and accountability lies with the others or the CEO mandates that as well. This is so that it should not be the case that we are building an ivory tower under the cdo. And that is not the way to go. In my opinion. It doesn't work. So hence my way. Hence, like you correctly said, the Accountability of ensuring that the data that the culture basically accommodates the data as well is something that which will become the accountability for the other business leaders as well.

That's great. And one thing that you spoke about recently on a data camp webinar is operationalizing a data driven framework framework for developing data culture or what you call the D3M framework. Do you mind speaking briefly about it and how you went about it?

See, it did not occur to us immediately that this is the way it should be done. It is over a period of time that we basically kept on fine tuning it to the level that we are today. And I'm pretty proud to have achieved certain things with my team. And like I said, I'm pretty proud of my team as well. Coming back to the discussion around data culture, we wanted to understand first of all, what are all the cues and what are all the rewards that the current existing routines are helping our business to achieve. And of course you can go about thinking like that on a very generic terms across the whole Allianz Benelux. But then we decided that's too much. It's an ocean boiling story. So let's go to just one area and start from there. And we started with sales. The question that we asked them was okay, what are all the different decisions which you make? And we listed them as per the business impact that they create. And then we said for these decisions there is a reward. And so we call these decisions as the cues and the reward we basically started to list and then we started to look at the routines, were they data driven or not? Now if the routines were data driven, good news. But if the routines were not data driven, that's where we started to act upon. So you see we started to make it hyper relevant, right? So for example, we discussed about one specific queue called brocus tiering. And we said that for broker steering to achieve the reward is basically nothing but more lead generation, lead conversion. So that's the reward. The routine was more instinctual and routine was more comfort dependent and more in the heads of the sales people even though there was a data reports available but seldom did they use it. So that gave us a good cue, good starting position and from there on we started to look at, we did surveys, we did interviews, we learned and then we realized that okay, if the data literacy improves and if data education is delivered then and the data quality also is improved, the possibility of also changing the routine from whatever it is today to data driven decision making will see the broker steering go data driven. And of course the impact of it was there and we started to manage. So you see, we now are able to even calculate the lead generation as a percentage towards this interruption that we created through data culture. So that's how we operationalized it. Now you can imagine that this can be scaled to other sides of the lines of business and other areas. Now this is a very intrusive but then a very hyper relevant and quantifiable approach of delivering data culture to an organization. This is how we operationalized it.

That's awesome. And you mentioned the use of surveys and key areas to evaluate when looking to operationalizing data culture. Can you walk us through what are the main areas to test for when organizations should start evaluating and building out a data transformation program?

Again, totally dependent on the organization itself. For us, what felt better were the areas around data. By data I mean the data collection, the business processes that we basically bring in the data and the areas around the infrastructure where you store, where you compute. And we also took into account the accessibility, the findability, the trustability, the understandability nature of these things, which means the documentation, the governance mechanism, the management aspects and more. So also the routines of the organization, right? I mean, how are the decisions being made? In my previous answer, I was referring to it as the routines based on a queue and expecting a reward, the routines which were followed through. So these were different dimensions that which we chose to answer your question, Adil, it's definitely something that which each and every organization is very different about. And a generic approach cannot be definitely subscribed to it. Such a thing does not exist because what we are talking of is a real culture.

I really appreciate your input on the uniqueness of culture and how organizations are different. But before we wrap up, do you have any other call to action or best practice you'd like to share with other data leaders looking to build out a data culture program?

The best practice that I saw really working is work from the top. It is not going to help you to build data literacy and data education just like every other literacy and education action is performed throughout the whole employee base. But I think the best practice, that which worked for me, has the highest potential to work to work from the top, which means from whichever is the topmost part of your organization. But at the same time, don't forget the middle. And also there has to be a clear conscious differentiation of delivery and the content of what you do across the organization. So when you work from the top, do not work with the top alone. Bring the middle as well in the same room while you Work with the top. This is very critical for ensuring that there is a dialogue between the top and the middle. What do I mean by that? Is when I'm working with cfo, I also bring all those who are reporting to the CFO in the same room and I'm working together as well. This is super critical because we are the enabler, so we should enable the dialogue. The dialogue of data driven decision making has to find its place and it will only start between these two because it has to happen there. So that would be my advice. One strong best practice would be to yes, work from the top. Don't just think about the general big democratic education, that which goes to everyone. That has to be done anyways. But focus and work from the top. Work with the middle as well. But remember that the success lies when you bring the top and the middle together because there is a dialogue that really is super critical. This dialogue also feedbacks into how you set up the content and the delivery for the next level of education and also helps you to explain the vision that which the top and the middle have to the people who are working pretty much in the lower layer as well. So that would be my best practice. Adam.

Okay, that's awesome. Suleiman, thank you so much for coming on the show and for sharing your insights. It's really appreciated.

You're welcome. Thank you for the opportunity.

That's it for today's episode of Data Framed. Thanks for being with us. I really enjoyed Suleman's insights on building data cultures and the unique nature of culture transformation programs. If you enjoyed today's episode, make sure to rate us on itunes. Our next episode will be with Elisa Visnich, CEO of YLABS on MLOps and why it's so important. I hope you'll find it useful and we hope to catch you next time on Dataframed.