Machine Learning Zoomcamp: Free ML Engineering course. Register here!

DataTalks.Club

Articles Slack Events Podcast Books Courses

Chief Data Officer

Season 4, episode 9 of the DataTalks.Club podcast with Marco De Sa

Did you like this episode? Check other episodes of the podcast, and register for new events.

Transcript

The transcripts are edited for clarity, sometimes with AI. If you notice any incorrect information, let us know.

Alexey: This week, we'll talk about the role of a chief data officer. We have a special guest today, Marco. Marco is the chief data officer at OLX Group, which is the place where I work as well. So we are colleagues. Before OLX, Marco worked at Spotify, Twitter, Facebook. I probably missed a couple of companies. You also have a PhD, right? Welcome, Marco! (1:18)

Marco: Thanks for having me, I'm excited to be here and to be a part of this community. (1:46)

Marco’s background

Alexey: Before we go into our main topic of the role of a chief data officer, let's start with your background. Can you tell us about your career journey so far? (2:12)

Marco: I'll start all the way back. I'm Portuguese, I grew up in Portugal, I got a degree in computer science. Then I realised that I liked people more than I liked computers. I hope I don't offend anyone with this. This made me move more into trying to understand people and how they use computers. I still love technology, more so than just spending all my time with technology. Because of that, I ended up getting a master's in Information Systems, which is a way of getting information about how people are interacting with systems — this is now called "Data Science". This was a long time ago. Then ended up getting a PhD in related topics, moving towards a qualitative understanding of users, like user research. That's my academic background. I had a lot of fun, I worked on quite a few different areas during my studies. (2:24)

Marco: I ended up moving to the US where I joined Yahoo. I worked as a research scientist that included quite a bit of data science: pulling data and making sense of it, understanding opportunities to build innovation. To be totally transparent, it was quite detached from the product development happening in the company — looking further ahead, like 5-10 years out. I worked on things like augmented reality on mobile devices. This was over 10 years ago before Pokemon Go. Also, I worked on live video production before periscope existed, and a bunch of other fun projects. The goal was to publish patents, which is fine, but not a lot of impact. (3:23)

Marco: I ended up moving to Facebook. At the time it had less than 1 billion users. I worked on trying to understand how can we expand on the products that we have and how can we build more value for our users. I was leveraging both qualitative and quantitative data and figuring out how can we solve user problems in the best possible way through technology, software, and design. I had a lot of fun and stayed there for a couple of years. (4:09)

Marco: Then I moved to Twitter to continue doing the same thing, but scaling up my impact through management and building teams. I spent a few years at Twitter. All of this was in the Bay Area. In Twitter, I worked on the launch of a ranked timeline that leveraged ML to organise and rank tweets based on someone's interests versus that typical reverse chronology. This was a pretty controversial move by Twitter but ended up working really well. It's still maintained. (4:40)

Marco: Then I moved to Spotify in New York. There I led a bunch of data science teams, marketing insights, started to build marketing analytics, user research, growth strategy, and a few other disciplines. My goal was always to use data to make smart and informed decisions that are going to benefit our users. (5:17)

Marco: Nine or so months ago, I moved back from the US to Europe, and I joined OLX Group, where I have the role of chief data officer. It's been a fun journey so far. (5:41)

Alexey: Thanks! That's quite a journey. I haven't met many people who say that they like people more than computers. Usually, it's the opposite. (5:54)

Role of CDO

Alexey: Can you tell us what do you do at OLX as a chief data officer? (6:08)

Marco: The chief data officer is a fairly new role, not just within OLX, but across industries. It's still being defined as we go. In my case, I'm responsible for defining our broad data strategy. That includes quite a few pillars: all the way from data infrastructure and data governance to "What kind of data do we need for us to continue expanding on our business?", and "What's the value that we bring to our users?", "How do we make sense of that data?" and "How do we make it accessible across the different organisations that are part of our group?", "How do we leverage that data to make better products through artificial intelligence and machine learning?", working with different business lines, business units, and teams that we have across the group. I want to make sure that we're prepared for the future, and we're delivering the best possible value for our users and anticipating needs that they might have. (6:15)

Alexey: That's quite a lot of things. I was also wondering what a CDO does, and I read an article before our char. It was about the role of CDO. Here is the summary. There are four main responsibilities. The first one is governance, which is monitoring and governing enterprise data. I think you mentioned governments as well. Then, operations — enabling data usability and availability. The third thing was innovation — driving cost reduction in generating revenue. The final one was analysis — supporting analytics and reporting. Is it an accurate description of the role? Or something is missing? (7:17)

Marco: It covers part of it. I think that's a more traditional, slightly outdated view on the chief data officer role. As we move forward, the integration with innovation, product development, business strategy becomes more noticeable — that's such a crucial part of any business. It indicates how your business is performing, which ties it to analytics. But it also powers a lot of product development and product work. Having the right data to enable ML applications to solve different problems touches on the access and usability of that data, touches on the collection, safety, and use of that data. It also touches on understanding the problems that you're trying to solve — not just from an efficiency perspective of collecting, storing that data, making it accessible in an efficient way and innovating, but also understanding our user base — our customers, both external and internal, so that we can use this data to solve their problems. (8:21)

Marco: That's the direction that we're moving towards: having people in a position of being able to drive the data, infrastructure, and organisations, but also understanding the real business problems that can be solved through data. Often I see a bit of a disconnect between business leaders and the more technical side of the organisation. They want something to happen, they have clear problems that they need to solve, but the technical side is usually focused on technical things instead of using technology to solve business problems. (9:32)

Marco: One point is not clearly articulated in this list: it's the responsibility to look forward and design for the future and for data. Data is going to continue powering innovation. The more data you have, the more you'll be able to do it. That could be as simple as asking a user question during the signup flow or collecting some contextual information such as location when an interaction happens. A lot of companies are doing that, but often they think of data as a consequence of the products, not a driver of the future products that they can have. That's part of the responsibility of the Chief Data officers to think, "Three years from now, what data will enable us to do something that we don't do yet?", and "How can we get that data in an ethical, responsible and safe way?". (10:19)

Keeping track of many things

Alexey: So you need to think about all these things? I don't know how many things you mentioned, five or six? How is it even possible for a single person to keep track of all that? I imagine that you don't just go there and then analyze data, you also don't go there and develop infrastructure. You delegate the work. But how do you keep track of all that? How do you know which things need to happen and which don't? (11:40)

Marco: Well, they're all, to some extent, related. The production of data, the collection of data, the access to that data, the utilisation, modelling that data so it can be used in different contexts, thinking about those contexts. Ultimately, it comes down to working with different people. No single person can have the answers for all the questions that have to do with data, and all the applications of that data in whatever business context. Building the right teams, having the right teams, having the right people that know much better than me what to do. I can offer the context, the resources, or the support for them to drive aspects of the vision that is created jointly. That's the key to success in the organisation where we have a complex set of goals that interact with each other — having the right people that share that vision, can communicate that vision, and work together to deliver it. (12:11)

Alexey: For each of these pillars, or responsibilities, you have a person who leads it. For operations, it could be a director of data engineering. Then your role is to keep everything in mind at the same time and make sure that they are not disconnected: everyone has the same context and understands why we're doing certain things. Is it an accurate summary? (13:11)

Marco: Yes, I think so. To some extent, the organisational design varies depending on many things. Companies are more and more distributed. Teams sometimes should be together, sometimes separate. There are various types of organisations. But absolutely, my role is to enable, extract the information and division from those different teams and try to articulate it into a single vision, a single strategy that makes sense for the business and also continuously delivers value to our users — which is the biggest driver of the business. (13:47)

Becoming a CDO

Alexey: Thanks! We already have a question: "What skills are necessary to go from head of data science or analytics to CDO?" This is something I also wanted to ask you. At some point, you were probably also a head of something, and then you eventually become a CDO. I'm curious what the journey looked like for you. What did you do to become a CDO? (14:24)

Marco: It's going to take quite a while to map exactly the things that led this path. The key differences are not in terms of skill set but in terms of where you spend your time as you progress in your career and take a more strategic role. Becoming more strategic is the requirement for moving in the direction of a CDO. That means prioritising aggressively to focus on the big picture: the vision, the distance, objectives that you have, versus the technical operations and how to execute that. The key in any leadership role is the ability to empower and enable people to do the best, hire people that can do that much better than you, being able to scale yourself, and not get too attached to the solutions and the execution of that vision. In the end, you're going to depend entirely on people that can do that with you. My journey has always revolved around surrounding myself with people that can do amazing things, working more as a facilitator and enabler of their fantastic work. (14:54)

Marco: I don't want to mischaracterize the requirements of this type of role. That's not specific to CDO, it could also be CTO, CPO, etc. Having a good understanding of the craft and some background helps you guide people who are more technical than you. Then you bring them along the journey with you. Having constructive conversations around that is pretty hard to do if you don't have any domain knowledge. If you don't understand the craft, if you don't understand the technology, it's hard for you to drive (but not impossible) a group of complex organisations, and be passionate about what you do as well. It's achievable by anyone as long as you focus on delivering the best possible impact for your business, using the tools that you have in your hand. (16:28)

Strategy vs tactics

Alexey: So, to summarise: you need to become more strategic, always keep the big picture in your mind, have this vision. I'm curious, when people say "you need to be more strategic", what do they mean? Then we also talk about strategy vs tactics. Tactics is something that you do immediately, right? And strategy is something that you plan long-term? (17:37)

Marco: Strategy is your broad plan to get from where you are to where you want to be, and it's more aligned with where you want to be. The tactics or the execution are the smaller steps that you take along that journey. (18:21)

Alexey: So strategy could be setting up some KPIs or some goals, like "This is what we want to do in three-five years", right? Then you tell about this vision to others, you convince them that this is the right direction to go, you try to explain it, and then it's up to others to actually decide how to go there and what needs to happen to actually be there in three years? So, tactics is the sequence of steps we need to do to achieve our goal in 3 years, right? (18:42)

Marco: I think that's a good example. I would also think if we have the right people to execute those tasks? Am I enabling those people, giving them the right resources, setting them up for success? I also need to ensure that we have the right people to build some of those strategy blocks. You can break down a strategy into smaller pieces of that strategy, that then are executed using some tactics. (19:18)

Marco: The line is blurry between the two. But it's exactly the case: looking ahead, setting some ambitious goals, and then deciding how to break this down to make it happen, and aslo working with the right people to share that workload and finding people that can execute it better than you can. (19:50)

VP of Data vs CDO

Alexey: There is this role called "VP of Data". What's the difference between VP of Data and CDO? And, in general, we have CTO and VP of engineering, we have CPO (Chief Product Officer) and VP of product, and we have CDO and VP of data. And to me, CDO and VP of data look pretty similar. They're quite high in the corporate hierarchy, both are quite strategic. They both need to think about "where we want to be in three years". So what is the difference between these two roles? (20:17)

Marco: The difference depends on your organisation and the size of your organisation. It ultimately boils down to the scope. I'd say a CDO is a level above in terms of the strategy influence. You're working closely with the CPO, the CTO, the CFO, and the CEO. You're thinking about all the angles of how data is beneficial for your business and how it's used to drive that business. A VP is typically responsible for a piece of that. That can be attached to a business client, a domain, or one of those pillars that I mentioned. They could be more focused on the governance and collection of data and infrastructure, could be more focused on the data science portion or the analytics portion. Whereas CDO has to have the horizontal perspective and the business component because they're part of the executive team. (21:04)

Alexey: I'm trying to think of an example. The goal could be "make X amount of money", or "X amount of market share" or something like this. This would be a strategic goal. This is where we want to be in three years. Then you think "We have this goal. How do we set our strategy to be there?". Then you look at all these pillars, and try to understand what needs to happen — from the data point of view — to actually get that market share, right? (22:29)

Marco: Yes. (23:09)

Alexey: So you decompose the strategy into multiple things, and the VP of data will take one of the pieces? (23:13)

Marco: Yes, to some extent. On the CDO level, the expectation is not just say, "Here's what we need to do from the data angle to achieve that goal". It's also "here are some goals that will drive the business in this way". Not just think of data as a supporting function to drive the business, but as a driver of the business itself. "If we do these five things from a data angle, the business will grow by this much." This proactive component is likely to be more present in the responsibilities of CDO. Whereas the VP is typically attached to a specific component of that strategy. (23:22)

Marco: But it depends greatly on the organization. If it's an organisation with 200 people and with a very clear domain or a clear problem to solve, then maybe the role of a CDO in that context is smaller than the role of a VP. The VP has a bigger context where they own a specific angle of the strategy. So it depends, but I'd say, the distinguishing factor is looking across multiple points. (23:34)

How many VPs of Data could be there?

Alexey: Do CDOs usually have multiple VPs reporting to them or just one VP of Data? (24:55)

Marco: It depends greatly on the organisation. In many cases, they can have several, depending on how they structure the organisations. Let's use a practical example that is not necessarily attached to any of the companies that I've worked at. You could have a company that is split into producers and consumers of some sort of content. Then you can have someone who's responsible for the data strategy for producers, and someone else is the VP of data for the consumers. Whereas the CDO has to work with both those people. And then there might be other components as well. It's not very common to have just one person in the reporting line. It also depends on the career ladder of the organisation. Some organisations have a long and steep career ladder, where you have senior manager, directors, senior directors, VP, Senior VP. Some have a more flat hierarchy where they'll have manager, then director, and then VP. (25:03)

Splitting the work between VP and CDO

Alexey: A few companies back at work as a company where we had a CTO, and the CTO had one VP of engineering reporting to him. That situation seemed pretty strange to me. It looked like the VP is doing all the work. And the CTO actually just manages one person. I was always curious how are we actually split the responsibilities between each other. If you look just from this organisational structure: CDO -> VP of data -> a bunch of people reporting to him or to her. (26:11)

Marco: If you have to spend a lot of time thinking about "What kinds of data will we need in the future?", "How do we tie and enable the business to move in this direction?", "How do you comply with governance models from a specific region of the world?", etc. And, at the same time, you also need to execute against building this tool, building this framework, building this infrastructure, then the division of work can be fairly simple. There's still a ton of work to do, that is covered by the CDO. But that's dependent on the size of the organisation and how the roles are defined. This applies not just to data. It applies to CPO and VP of product, to technology as well as other functions. So it depends on how much is there to do and how many people do we need to do it successfully and continue building upon that. (27:01)

Difference between CTO, CPO, and CDO

Alexey: Do you know what's the main difference between CTO, CPO, and CDO? When I look at the top management — it's all about meetings, setting strategy, planning for budgets. These responsibilities look pretty similar. So what's the core difference between all these roles? (28:02)

Marco: It depends on the type of company and the type of product. In some cases, the product is more technology-related — if you're building a framework or a platform that other companies are going to use, versus building something for usual users. It varies quite a bit. Typically a CPO is more focused on determining the product vision. What is the product supposed to do? CPO then works with CTO, who is focused on "how do we build that" and "how we use technology to make that vision come to life". The CDO is more concerned with "what data do we have that enables us to build that, to inform the direction that we're going in, and to continue pushing that vision forward". That could be a high-level theoretical division of responsibilities. But it depends a lot. For some companies, their product is data. Their CEO can likely fulfil both roles. I've seen places where all those roles are combined. And I've seen places where they're clearly differentiated because that's the nature of the product and their technology stack as well as the data they produce or consume. (28:30)

Alexey: To make it a bit concrete. We talked about this example of "get X market share", or "that amount of revenue". Probably this is a goal that the CEO would set together with the entire team. Then it would be up to the CPO, CTO, CDO to actually decide how to decompose this goal and see what kind of data initiatives could happen, right? Like, "we need to make this, this and this data available to make these things possible. Everyone thinks "what we can do from the engineering point of view?", "what we can do from the product point of view?", "what we can do from the data point of view?", right? Then it cascades down to a VP of data, who would need to think about a particular chunk of the data strategy, up to engineering managers, and so on. (29:52)

Marco: Yeah, VP of data — or whatever the titles are in that organisation. I would again add that the responsibilities of those three C-level types of roles are beyond just executing or breaking down the strategy from the CEO into the specific functions that they represent. It's also to provide input and offer a vision of the future that the CEO can put together and help ensure that he enables the other C-level executives to deliver that. It's more of a proactive role, where you're not just delivering against something that is asked of you, you're offering opportunities to the person that can help you make those decisions or make those decisions for them. (31:00)

Breaking down the goals and working backwards from them

Alexey: You mentioned at some point something like "if we do these five things, then we will achieve this goal". I'm wondering, how do you come up with these things? How do you know if we need to invest in a data platform? Or how do you know if we need to enable our analysts to process data better? Or how do you know if we need a better machine learning platform? How do you know where the company should go to achieve this goal in three years? (31:50)

Marco: That's a broad question. There are many ways of approaching it. I approach that by first aligning on the goals and then working back from those goals. I try to understand what are the blockers for us to achieve that. Then, break those blockers down by the magnitude and impact of that blocker and optimise for things that are going to help us get there faster. What are the enablers that are going to help you drive that vision forward? The vision itself is also organised around the impact that it brings: the value for users and the value for the business. What are the biggest bets that you can take to help your company, colleagues, and users to have a better experience, to get more value out of the product, and be able to achieve their goals faster? (32:26)

Marco: That's how I set goals. First I understanding what are the big problems around us from a user perspective, from a business perspective, from an organisation or technological perspective. Then I see which ones that we can solve will drive the most impact and get us closer to that shared vision — not just the data vision, but the shared vision for the company. Then I work back from those and understand what we need to do to get here. For example, we now have multiple platforms. That slows us down. If we have a common one, then we agree that we all can move faster. So let's build that together. (33:19)

Marco: That's how you break down goals and work backwards from them. Try to come up with simpler things and simpler chunks, then you can provide full ownership and accountability to people that can drive them further than you can. (33:56)

Alexey: You care more about blockers and problems than specific solutions, right? Then your team thinks, "For this blocker, this could be a solution. Because we're doing this thing not in an optimal way. To make it more optimal, we need to do this." And you trust these people to do their job. You don't interfere. You just check that we indeed move in the direction we want. Right? (34:12)

Assessing if we’re moving in the right direction

Marco: Exactly. An essential component of that is aligning on how to assess whether we're moving in the right direction. What are the metrics that help us understand whether we're achieving our goal? Of course, you have to have some level of trust in the people that are going to do that. But also you need to understand and have visibility into how it's progressing so that you can unblock, can enable people, and also correct if necessary. Your experiment with something. Are you moving in the right direction or not? You have to make some of those adjustments. And then also you have to bring the right context. If all these pieces have some level of interaction and people don't have visibility across, it's also my responsibility to create that visibility, and to find synergies where they could exist. (34:43)

Alexey: Would you say that one of your responsibilities is measuring the effectiveness of different teams, projects, initiatives? Maybe not measuring, but making sure that it can be measured. Let's say, there is a new initiative, and we define a KPI first, and everyone agrees on how it should move. And your job is to make sure that this is indeed measurable, everyone agrees on the KPI, and everyone is aligned. Right? (35:28)

Marco: That's the responsibility of any leader or manager to set clear goals, and have clear ways of creating ownership, accountability, and measuring whether you're moving in the right direction. That becomes even more relevant for a data-focused leader because we live and die by data. We care a lot about what types of data are we looking at, how we're collecting that data to ensure that we have a good view on whether we're progressing in the right direction or not, generating more data through experimentation and through different trials and errors. (36:08)

Dealing with many meetings

Alexey: I'm curious how many meetings per day do you have to be able to do all that. To align with all these people, your day must be full of meetings? (36:45)

Marco: It is pretty full. I, fortunately, benefit from having someone who helps me manage my calendar. I also have an amazing set of teams that I work directly with — either under me or across our group that I collaborate with. This helps me minimise the meetings. But, nonetheless, I have around 8-10 meetings today, on a regular cadence. I'm trying to reduce that — to have more time to think long-term and articulate some of those visions and strategies for the rest of the organisation. But right now, it's a pretty jam-packed schedule. (37:03)

Alexey: How do you not go crazy with this amount of meetings? Or this is a trait that a CDO must have? (37:44)

Marco: I have a few things that work well for me. One is I like some of the meetings that I have. Some become tedious. But, in general, I like the meetings that I have because I typically learn something from them. And I also like the people that I work with. I'm very lucky to be part of a broad team that has fantastic people that I enjoy — even through zoom — spending time with. That's one of the benefits that I'm lucky to have. (37:54)

Marco: I also try to plan ahead. Again, I benefit from someone who helps me manage my schedule and organise it in a way that makes sense and I can prepare for. That helps quite a bit. I don't have to switch too much context. A simple example is having one-on-ones with my reports in one day, more strategic discussions with external people in another day, with internal people — on another day, team meetings — different day. Having a bit of that structure helps quite a bit. (38:22)

Marco: There are more techniques like documenting, taking notes, making sure that there's an action that takes place after the meeting so that the meeting is fruitful — when we move forward, we don't have to revisit the same issue over and over again in subsequent meetings. (38:59)

Marco: Also, everything that I do outside of work helps. Sport helped me separate from the zoom world that we live in right now, clear up my mind, focus on different things. Taking different projects within work helps as well — projects that have to do with diversity and inclusion. I'm passionate about them and they are not so related to the things that I do on a daily basis. It helps me switch context and have my brain thinking about different problems. That helps. Otherwise, you get too closed up in one specific problem and it limits your ability to think through it. (39:15)

Being more effective

Alexey: Thanks. Do you use any productivity tools? You mentioned somebody manages your calendar. Do you use any tools on top of that to help you be more effective? (39:56)

Marco: There are some traditional communication tools that I use, like chat rooms and platforms to communicate with my team. Documentation is a pretty powerful way of sharing vision, collecting feedback without necessarily having to spend time in a meeting. Doing so offline, asynchronously helps. Also synchronously. Being available to have a quick conversation sometimes can solve a problem and avoid a meeting — just by chatting with someone for a couple of minutes. Those tools helped my productivity. Basically, it's just keeping a structured schedule, and being able to take short breaks helps quite a bit. (40:16)

Alexey: Interesting. I never thought of chats like Slack as a productivity tool. But it is and often it can save you time. Maybe this meeting should be just a chat in Slack. (41:10)

Marco: Quite often multiple times a day, I'd say. (41:25)

Alexey: But these meetings still happen. Or not? (41:28)

Marco: If you can solve the problem through Slack, which often is the case, you don't have to have the meeting. But in some situations, Slack will expedite the meeting, because you come in with some context already. You discuss, read some documentation. This is something that we're adopting now — reading whatever context you have before the meeting starts, or right at the beginning of the meeting. It expedites the conversation and discussion quite a bit. (41:31)

Building the data-driven culture

Alexey: Thanks! We have quite a few questions from the audience. Is the CDO responsible for building a data-driven culture? How a CDO can achieve that? (42:02)

Marco: Absolutely, I'd say it's everyone's responsibility to build a data-driven culture. (42:21)

Alexey: You should promote it, you should explain that this is important, right? (42:28)

Marco: Absolutely. (42:31)

Alexey: Otherwise, maybe the product people don't know it. (42:32)

Marco: Anyone in the organisation has to make decisions. You can always make better decisions if you have some level of information and data or evidence. That enables you to make a smart decision. It's especially the CDO's role to enable a company or a culture to be data-driven. First, people need to understand that data plays a role and be eager to have that data — to help them make great decisions. The CDO and the data teams have to make sure that when people need data, they have that available. They also think ahead and provide data that will be useful at some point. They make sure that people are aware and have access to that data. (42:37)

Marco: That comes back to data usability. It's not just for algorithms and products to use this — it's also for people to use to make decisions. It ties it very closely to analytics, which is another component of the role. That's the CDO's responsibility to ensure that it's set in place: the data is well democratised, accessible, easy to use, and quick to use — in a way that you can make fast decisions. That means not just looking at data, but also producing and leveraging that data to generate insights. (43:21)

Alexey: So it's everyone's responsibility, but CDO should make sure that there are tools for doing that, everyone can use those tools, and people are educated as well. (43:56)

Challenges of working remotely

Alexey: Do you feel that remote work is challenging your role? Especially when you manage your team? (44:12)

Marco: Absolutely, yes. Full transparency, I haven't met the people in my team. I've never been to an office. I joined OLX after the pandemic and after this remote work started. It has not been limiting. It hasn't prevented me from doing the things that I want to do. But of course, it is a challenge. It's hard to build relationships, to get to know people on a more personal level — which helps some discussions to take place, understand people's motivations, what makes them tick. I think it's a really important component of working with others. Doing that through a little window in your computer is much harder than doing that in face-to-face situations at the office where you spontaneously meet someone, a conversation comes up. You have ideas, you exchange ideas. It is challenging. Absolutely. It hasn't prevented me from achieving the things that I wanted, but it likely has caused more work and more investment into some of those relationships that I have to do in the virtual world that we live in right now. (44:20)

Alexey: So you need to communicate more now remotely. (45:35)

Marco: Yes. And provide extra context when you have virtual conversations. There's a lot of research on the added cognitive load of having virtual meetings and reading contextual cues from people's gestures and language and so on. That is attached to us living in this world now where everything is online. (45:39)

Alexey: Surprisingly, from what I saw, remote work is not that bad. It kind of works. Of course, it's bad not to be able to see people and meet them in the office, but I thought it would be worse, actually. (46:11)

Marco: Absolutely. It brings advantages as well. It's not just problems or challenges. You can now have access to talent that would never want to relocate. You can employ them and benefit from their amazing skills, creativity, ideas, and perspectives. That is amazing. It opens doors for a lot of things — for people to live where they want to live and still work for the companies that they want to work at, that have the missions that they're passionate about. It's a fantastic thing. (46:30)

Marco: I am also very sensitive to the fact that it's not the same for everyone. Some people love it and will thrive in that environment. Some people — and I know quite a few people like this — still crave that human connection that they get out of the office. Also, the fact that it gets them out of the house, they get to spend time with others, go for lunch during your lunch break. So it brings a lot of options and it has a lot of benefits that we're still starting to observe, but it's hard to debate that it also brings some challenges, especially for certain people. (47:03)

Alexey: Interesting. By the way, I am in the office right now. I do like this routine of biking to work, having lunch, and accidentally bumping into colleagues, and have these little chats. (47:42)

Does CDO need deep technical skills?

Alexey: Do you think that the CDO role requires deep technical skills? What kind of technical expertise is needed for the role? Machine learning, programming, SQL? Something else? (48:04)

Marco: You can always benefit from having the technical skills. But the role relies more on being able to look across, find touchpoints between different parts of the strategy, different organisations, different pillars. But data and data science — from applied ML to data engineering, to analytics and insights — all those things are continuing to grow as we collect more data and learning new ways of using it. So it benefits us more to have a good technical background — at least it has benefited me more. It helps to have a good level of understanding of the technical portions of it. You don't necessarily need to be too deep into one specific area. That can be an advantage in certain contexts. But for me, understanding and being able to converse and articulate my perspective in different areas was more useful than being able to dive very, very deep into one specific area. (48:16)

Alexey: So you need to know a bit of everything but not too deep. (49:35)

Marco: That doesn't mean that I don't have one specific that in the past I went deep into. But as you progress in this type of role, it's beneficial to get a better understanding of other components. That comes with a consequence of not being able to spend as much time on the thing you're an expert in. Given how fast the field moves, you end up not being much of an expert anymore. Technology advances, tools advance. It's hard to keep up if you have to keep tabs and juggle many balls. (49:39)

Importance of MBA

Alexey: What do you think about MBA? Do CDOs need an MBA to be successful? (50:20)

Marco: I wouldn't say they need one. I don't have one and I'm in this role. Well, we'll see over time about the "successful" part. I think it's likely beneficial. Any type of education that gives you a different perspective — especially the one about understanding your business and managing your business — is likely to be beneficial. In any executive or C-level role, you have to understand that business. You cannot just understand your function and your craft. You have to understand how to use it to drive the business. An MBA prepares you for that. I don't have one, I've never gone through the process of getting one. And I've met many CEOs, CTOs, CPOs without MBAs that are very successful. So it's not a requirement. But it's likely helpful. (50:28)

Alexey: It's maybe like looking at Mark Zuckerberg and decided that I don't need a university education. (51:31)

Marco: He's one of many examples. (51:39)

Alexey: But maybe it's useful, in the end? (51:42)

Marco: Exactly, it can be very useful. It can prepare you well for certain problems. But in the case of Zuck or any of those very successful CEOs, they're just really smart people. They can accumulate a lot of knowledge from the experiences that they go through, the challenges that they overcome, the failures that they have — that's something that often we don't talk about. Those very successful people went through a lot of failures as well. And they're able to learn from it. That experience, especially if you're managing a big business, is equivalent to — if not better than — an academic version of it. (51:44)

The key skills for becoming a CDO

Alexey: Maybe this is one of the main skills — being able to learn from all this experience, to become a CDO. Because the next question is, "What are the main key skills to be a successful CDO? And how to build the skills?" We talked about this a bit at the beginning. You said "You need to be more strategic. You need to have this big picture, vision. You need to always keep the business side in mind. Not just focus on data, but also think how is it helpful for the company in general? How do we work together with other functions like CPOs, CTOs?". Is there something else that is very important to be a CDO? (52:18)

Marco: In terms of skills, soft skills are especially important: being able to communicate, articulate a vision and strategy, being able to inspire the people that will then help you drive that vision forward. It's important to be empathetic to people you work with, and to people you want to serve through what you do. Especially nowadays, there's so much competition anywhere. If you're doing things right, you're starting with your customer. Think about your users and understand that your position and your needs are not necessarily the same as theirs. Be able to put yourself in their shoes, and then drive solutions that are applied to their problems. It's a skill that is important not just for a CDO, but for anyone who's trying to drive a business to success. Those are softer soft skills, but they are as important as some of the hard skills. (53:11)

Biggest challenges within OLX so far

Alexey: There is a question from somebody called "Ioannis Z" — sounds like somebody we both know. What was the biggest challenge within the group so far? Given that it's a pretty diverse group, both geographically and in terms of tech stacks and maturity? (54:16)

Marco: The biggest challenge is that we have a group that is fairly big. It has a set of different companies within it. Each of those companies operates in different regions and geographies, with different domains and business minds — all the way that from jobs and services to car transactions, to classifieds, to goods as well. There's a wide variety of business lines, problems, and user bases to deal with. That generates a lot of complexity. This is even exceeding the organisational complexity of us existing in multiple markets and not being able to travel and meet each other in person. That's been the biggest challenge — how complex it is. (54:41)

Marco: It's also the most fun part — in addition to working with amazing people — is being able to immerse myself in so many different topics. You never get bored. There's so much stuff happening — all the way from online to offline. In some cases, some of our companies have inspection centres, and dealerships, and things like that. Those generate data, even though not collected in the same way. But it's relevant and we can use it in a smart way. There's so much opportunity for data, data science, analytics. All of those things play an important role and solve big problems for our users and for the business. It's an awesome thing to be a part of. But it is challenging. There's a lot to wrap your mind around. (55:31)

Alexey: At the end of each day, your head must be spinning with the amount of information you had to process throughout the day. (56:24)

Marco: Yeah. I won't deny that. That's very true. (56:31)

Demonstrating the CDO skills on a job interview

Alexey: We have a question from Carmine. Let's say somebody wants to interview and get hired for a CDO position. How can you demonstrate the skills we talked about, like looking across the organisation, creating synergies, and so, on that interview? (56:36)

Marco: That's a tricky question. It depends on a lot of things. One, it depends on the skills that are necessary for that specific CDO role. We talked in the beginning, CDO roles can have different flavours and different organisations. The second one, it depends on the process. Some processes have a presentation where you have a case study. You need to demonstrate your communication skills by sharing what you thought. And your strategic skill by looking at the problem and thinking, "Is this the real problem? Are there different problems associated with this one?" Is this the right problem for us to be looking at?". And then thinking how you can leverage different components — the resources, the teams and so on that you have to push that forward. In some cases, that will happen in one-on-one conversations. (56:56)

Marco: But in general, it's just demonstrating that you care deeply about the problems that are being presented to you, that you'd go and gather as much context and information as possible. Think about what other data points would be necessary to provide a solid answer to that. And be honest, when you don't have an immediate answer, and you need more data. That's part of our role is identifying when we need more data. Make sure that you can communicate that clearly, leverage the knowledge of others and scale yourself up through the teams that you're able to build and hire in order to drive those visions. (57:52)

Marco: It's hard to answer that question specifically because different people have different ways of demonstrating those skills. In most cases, just answer the questions that are asked of you. And show if those questions aren't being asked directly, show that you can go beyond what's just being asked. (58:36)

Alexey: Some of these things are useful for any interview — doing this homework and trying to show that you care about the problem, understanding what the company is actually doing, what kind of problems they have. That is useful for anyone from a junior data scientist to a CDO. (58:54)

Marco: Absolutely. Also, being honest about what you don't know, not being afraid to ask questions, and fail and say the wrong things in some situations — because no one has the answer for everything. (59:16)

Overcoming resistance

Alexey: Ricky says that he agrees 100% with the emphasis on strategic vision, and understanding the needs of the business. But how to overcome resistance to the details of your vision when somebody disagrees with that? (59:40)

Marco: The first thing is to understand why they disagree. In some cases, they might have had a better idea. You should be open to taking that feedback and adjusting your idea. Also, often, convincing someone is about breaking down the problem and providing evidence that your idea is good. It might be a hypothesis: you have a hypothesis that this will be a good solution for your problem. Have a concrete plan on how to prove that hypothesis. What is the data, or information, or evidence, or context, or input that leads you to believe that it's a strong hypothesis? It should be a hypothesis that you can validate, measure, and provide this proof. (59:55)

Marco: So, the best way to approach it is to break down the problem and then demonstrating, articulating in a clear way, why you want to do it. But always be open to being disproven, to getting feedback. I like to surround myself with people that know about certain problems much more than I do. They're likely to come up with better solutions. So if someone disagrees with something, I always try to understand why. Maybe together we can come up with a better solution. Or maybe I can explain my solution or perspective better. And that's feedback in itself as well. (1:00:55)

Alexey: That was a summary of persuasion 101 course, right? (1:01:37)

Marco: I don't know how this is always applicable. But in some situations, you also have to say, "It has to be this way because of constraints that we cannot discuss" or something like that. And there's not a lot of persuasion other than "We have to make a decision. This decision". But often, it's a factor of having that discussion, together aligning on whether that's the right solution or not. In some cases, you have constraints that force you to make decisions very quickly. You have to do those. And be ready to fail, but also be ready to learn from it. And take responsibility and accountability if it doesn't go well. (1:01:43)

Wrapping up

Alexey: Thanks, Marco. I hope it was a nice break from your usual routine. Thanks a lot for joining us today for answering all these questions, for sharing your experience with us. And thanks, everyone for being active and asking so many questions. I'm sorry that we didn't cover many questions — there are still eight questions. This topic must be quite interesting for everyone. Thanks, everyone, and thanks, Marco. (1:02:24)

Marco: Thanks for having me. Thank you, everyone, who dialed in and listen. I hope there was some use out of this. For me, it was fantastic. And an awesome break from the usual day-to-day. Thank you for that. (1:02:51)

Alexey: Have a great evening. (1:03:05)

Marco: Cheers, everybody. (1:03:09)

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.

DataTalks.Club. Hosted on GitHub Pages. We use cookies.