Orlando Machado was a data scientist long before the role became one of the hottest in technology. From academia to the BBC, dunnhumby, and Aviva, Machado has spent the past 20 years exploring how statistical techniques can turn data into predictions and create better customer experiences. In 2019, DataIQ ranked him number one on its list of the most influential people in data-driven businesses.
As chief data officer for the Lego Group, a role he’s held since 2021, Machado is bringing his data savvy to the world of play. In an interview with McKinsey’s David DeLallo, he explains how he’s helping to strengthen the data and AI foundations of this famous brick-building enterprise. An edited version of their conversation follows.
David DeLallo: Lego is a physical toy, so what is the data and AI connection?
Orlando Machado: Well, as you say, we are a physical toy first and foremost. We always have been. It’s our 90th anniversary this year, so we’ve been making physical toys for a long time. But what we do know is that consumers want physical experiences to be supplemented by digital experiences. That could be about shopping, so creating great online shopping experiences. But equally it could be about play. So we have a number of different digital products aimed at allowing people to build Lego products together. We’re increasingly trying to create digital products that supplement the physical experience. That also goes for the experiences we’re trying to create for colleagues internally. We want to have a seamless contemporary workplace where people’s lives could be made much easier using technology.
David DeLallo: What role do you see AI playing in this?
Orlando Machado: The Lego Group does everything from product design, new-product development, manufacturing, retail, wholesale, supply chain, logistics, brand building, and consumer engagement. And all those processes are consuming and generating data. So we see AI as an opportunity to make any of those processes better. It could be about helping our molding machines work more effectively, or it could be about more effective customer engagement, or it could be about just creating fantastic online building experiences to help kids play together when they’re using the physical product.
David DeLallo: Can you provide an example of how data dramatically improved a process?
Orlando Machado: We have an app called Lego Life. Kids can upload pictures onto that app. We’ve always used human moderators to ensure that people upload pictures that are safe. But we discovered that we could quite easily train an AI agent to reject some content. For example, kids will upload a picture of something they built, and they might capture their own face in the picture. We can train an AI agent to reject that content immediately, which means there’s less work for the human moderator, because they only have to do the second round of moderation. That enables much greater efficiency. It also creates a much better experience for the child because if you’re going to have some content rejected, it’s much better if it’s rejected instantly rather than waiting for a human moderator to step in, by which time you’ve probably broken up the thing you’ve built and posted lovingly onto the Lego Life app. So we try to find this sweet spot for efficiency, process improvement, and great customer experience. I think when we do that, people have these “aha” moments. They suddenly understand that this technology isn’t something to be afraid of. It’s something we can use responsibly to deliver for our key consumers, who are kids.
David DeLallo: Because Lego produces children’s products, concerns about AI and privacy must be magnified. How are you helping to instill trust?
Orlando Machado: Two things. One is that we take a very conservative view when it comes to the use of data at scale when it’s about kids. It’s probably surprising to a lot of people, but there are many experiences we create for kids where we don’t collect data because we’ve actively decided to err on the side of caution. The second thing is we launched a data ethics framework last year. It’s high level. It talks about being positive, clear, fair, and responsible, which are very high-level principles. But they at least allow us to have a debate about new use cases using a framework, which is a bit less blunt than we had before. If people were worried before, they would opt out. Now they have a framework within which they can debate about whether this is right or wrong according to the Lego Group’s values. And that’s really been a big step forward. People can understand that there can be sides to a debate. We can reach an agreement, and we can help use data in a positive way if we think that it’s of benefit to our consumers.
David DeLallo: You were a chief data scientist before you were a chief data officer. What interested you in making the switch? And how has that data science background helped in the data officer role?
Orlando Machado: I think data science is still an untapped opportunity for most companies. Part of the reason for that is that the foundation isn’t always there. The flow of data, the governance, the quality, the ethical frameworks, and the education are not there in many companies. I was always interested in data science. I’ve been doing it for 25 years. I wanted to be able to make more of an impact at a bigger scale, and to do that, I wanted to be responsible for putting all of these foundational elements in place.
David DeLallo: You’ve been chief data officer at Lego for about 18 months. What do you see as your key responsibilities?
Orlando Machado: I think chief data officers come into the role via a number of different paths. Personally, I started my career in academia. I worked on statistical modeling, machine learning, and data science so I’m trying to help unlock the value from data across the Lego Group and help the whole company spot new opportunities for using advanced machine learning, advanced data science, and AI. But in order to do that properly, people need access to data. So my role includes sorting out our foundations and our data platforms, and enabling the seamless flow of data across the organization. It also includes things like governance—making sure that we keep an eye on data quality. We have a good data ownership model. And it also includes data ethics. So we’re trying to provide frameworks to help people understand how they can do the right thing even when they’re using extremely contemporary technology that they might not understand themselves. We’re trying to help people navigate that complex and fast-changing landscape.
David DeLallo: What are the top priorities for you today, and how are you tackling them?
Orlando Machado: One of the first things we’re trying to do is put the technology in place that will allow data to flow effectively across the organization, essentially connecting data producers with data consumers in a marketplace model rather than the traditional model of having a central team that will try to do everything. That’s a major change for the organization, and it’s a major rollout.
Alongside that, we’re working to develop the cultural adoption and new habits that are required, as well as the data education and upskilling that are part of any change effort. These parts will always take a bit longer to achieve because we’re trying to get people to understand the opportunities that arise from technologies they haven’t been used to working with before. But increasingly, we see that the benefits speak for themselves. As soon as we start to lift the lid on the opportunities and people start to understand, we see a huge amount of enthusiasm for using data-driven technologies.
David DeLallo: How do you go about prioritizing where to focus your efforts, given the many ways to use data?
Orlando Machado: It’s tough. In a central function, you can never hope to make prioritization calls for everyone. That’s the reason we’ve taken this enablement approach, because if we can enable as many people as possible across the Lego Group to get better and better at using data, then they can make priority calls for themselves on a local basis, which I think is much closer to the business challenges we are facing. Where we do have to prioritize centrally is where there are new ideas that need to be tackled by data scientists or specialist teams. So we tend to think about three things really. One is whether there is a short-term tactical benefit. Sometimes there are applications for data science that drive a very short-term, very measurable benefit.
The second one is about strategy: Does it have a contribution to our stated long-term strategy? We have stated strategic goals about things like sustainability. So if we have a data science use case that helps with a strategic goal, that’s another way we prioritize.
The third one is a little bit more esoteric, but we think about things that are iconic. Because sometimes there are use cases for data science and AI that might not deliver a short-term benefit and might not be perfectly aligned with our strategy, but the education benefit is so high—they will help provide the message of the opportunity to such a wide audience—that we want to do it anyway. If we can get an improvement in the way we count Lego bricks, for example, something that people can relate to, people will remember it and make the connection that data science and AI did that. So when they come to the next problem, they think about data science and AI being a potential solution. We always want to have a few of these use cases that we prioritize purely for the education benefit.
David DeLallo: How are you helping to train the wider workforce on the use of data and analytics?
Orlando Machado: We aim to provide data education at all levels, from a data scientist who might be a subject-matter expert already but wants to get to the next level, all the way to somebody who maybe has a passing interest in data—they’ve heard about it and want to know what it is. We’re at the start of this journey. We’re helping a number of our business intelligence and analytics teams work closely with the central data office to establish these new ways of working, whether that’s about using our technology or our services or about upskilling and education. We are starting with some slightly more data-savvy audiences, typically people who are in business intelligence or analytics functions. But the aim is to reach everyone.