How Decathlon uses data to optimize in-store operations

Tarush Aggarwal (00:00)

Hey, everyone. Welcome to another episode of People of Data, where we get to highlight the wonderful people of data. These are our data leaders from various industries, geographies, and disciplines, discover how companies are using data, their achievements, and the challenges they face. Today on the show, we have Anandita Mishra, a global leader in AI innovation, digital trust, and data transformation.

‍

With nearly two decades of experience across finance, utilities, retail, and tech, she has a proven track record of turning data into strategic assets. As global director of knowledge activation and trust at Decathlon Digital, she drives AI power initiatives, enhances analytics, and ensures compliance with the EU AI Act, all while fostering a culture of data-driven innovation. Welcome to the show, Anandita.

‍

Anindita (00:53)

Thank you so much and thanks so much for having me. I'm really excited to be here and share some insights around who we are in Decathlon and what we do with data.

‍

Tarush Aggarwal (01:02)

That is awesome. Are you ready to get into it?

‍

Anindita (01:05)

Absolutely! Let's get it going!

‍

Tarush Aggarwal (01:07)

Let's do it. So for those of you who don't know, for those of our audience who doesn't know, what does Decathlon do and how does data play a role in helping Decathlon do what it wants to do?

‍

Anindita (01:18)

Of course, that's a great question. Decathlon is a French sporting goods company and the largest multi-specialist sports retailer globally. What makes us truly unique is that we manage every part of the supply chain. So for value chain, it's in-house from research to design to production, logistics, distribution.

‍

And we also work with global suppliers and directly market our products to consumers through our Decathlon branded stores across around 78 countries. We also have marketplace where we sell other products, other brands like Adidas, Puma, et cetera. So how did it start? We started in 1976 in Lille in France.

‍

And today we have more than 100,000 teammates, coworkers across all countries, over 1700 stores worldwide and offering equipment or sports goods for more than 80 sports. So the core purpose is that our products are for everyone. Whether you are a beginner or an expert, we are always pushing.

‍

to innovate and improve our products and services to serve the many. So right now, we are in a very important phase of the growth and the transformation in Decathlon. So you probably have seen the new brand of Decathlon as well with the new ways of the stores that look like. Our purpose today is move people through the wonders of sports, which

‍

means tackling the current topical societal challenges like what we face today, the rising inactivity, increasing stress and childhood obesity. So our strategic vision for the coming years are based on three main pillars. One is the customer experience, which is the core. The second one is sustainability. And the third one is modernization and digitalization of our business processes.

‍

So as you can imagine that when we talk about digitalization, data is the central offer strategy as well. So we are adopting to a platform thinking and a data mesh approach with governance at the core of it. So treating different areas of the business like think of a player in a well orchestrated game, right? So each player plays a very important and crucial role.

‍

in how we actually use data to create data to use to generate AI applications or actionable insights for the matter. So while building the foundation, we are also leveraging advanced technologies like AI, I mentioned GenAI as well, enterprise knowledge graphs to build the foundation of the future scaling to generate business value.

‍

So for instance, one of our initiative actually is about how do we use context aware search and recommendations to with power of GenAI to improve customer and coworker experience, either in store or online to improve how do we find our products that are relevant for their when they do their sports.

‍

So this is a very tangible example how we are using building foundation while using innovative technology and the data at the core of it to move forward.

‍

Tarush Aggarwal (04:47)

Awesome. What does the company culture look like?

‍

Anindita (04:50)

I would say, yeah, it's very flexible and very responsible and growth is one of the core value in here as well, while keeping the trust and accountability and collaboration, just togetherness is also crucial. I would say it's probably important also not just in Decathlon to everywhere, right? As data...

‍

leader, for example, you always leverage about like, how do we collaborate? How do we bring the value of the data for business throughout everywhere? So while in Decathlon, we want to make sports also accessible to everyone and that it holds in that it was actually reflects in our internal culture too. So our recent transformation, which is

‍

as actually an orbit, a logo, that the new one that represents the shift towards circularity as well and more continuous evolution. So these are the values that we are actually, we all carry forward to whatever work that we do in our daily life.

‍

Tarush Aggarwal (05:54)

That's great. What does your data footprint look like? What's your current stack? What's your current data tech stack?

‍

Anindita (06:01)

Yes, so to that actually, our data landscape is very complex. That actually reflects our scale of our operations as well. what we today we have a multi cloud environment with Google Cloud Platform hosting majority of our operational systems like e commerce or ERP or content management systems.

‍

while AWS is used for our Amazon Web Services is used for our analytics, data analytics platform. So we still have some legacy data warehouse and some systems that are still on premise, but we are currently in the process of moving them also to be more cloud native organization.

‍

And so we are also moving from being a centralized data organization to more data mesh paradigm, as I mentioned, so more distributed data accountability culture. And that also means the platform needs to be really ready to have that as well. So to manage this complexity, we rely on several tools for data platform, which is on AWS for the time being.

‍

Databricks for data processing, Kafka for real-time streaming, dbt for ETL primarily, including Spark also being used for more heavy data processing as well. Tableau is used for primarily analytical dashboarding for BI, while SageMaker, Bedrock, and Vertex AI also being used for various operational or analytical type of analytics.

‍

putting a lot of effort in the data governance part, which means also metadata management is becoming very, very important. So we use AWS Glue combined with Colibra for the metadata management side. Great expectations is used for data quality together with the data governance rules and the policies. So we are talking about competition governance as much as possible as well. And Neo4j is for building the knowledge graph.

‍

Tarush Aggarwal (08:00)

Wow, that is the most, that's just like all of the big names all put together. How big is the data team needed to, know, how big is the data team? How big is the company?

‍

Anindita (08:11)

We have over 100,000 employees overall, which includes, of course, our teammates who are actually serving in the store, so selling basically our products. So that consists of actually 71 % of our employees. They are actually working for stores. Then we have data citizens, where we call them, like the knowledge workers and the teammates. They are heavily working.

‍

within product sports product team or marketing team or commerce team or finance team. That is around 24 % and literally around 300 people there working in the core AI or data part. And I would say that around a little bit more like around 200 are pure engineering that includes both data and software engineering. So this is the scale of

‍

the engineers versus who we are serving.

‍

Tarush Aggarwal (09:10)

What's one achievement which you've had the privilege of leading which you are which which has had an impact on the business which you're really proud of?

‍

Anindita (09:20)

That's a great question and it's very close to my heart as well, of course. So absolutely. In Decathlon, we are currently in the process of scaling Gen.AI-driven, out-powered some of the initiatives, are primarily how do we improve the customer experience, omni-channel, starting from customer service to e-commerce, where product findability is a challenge today. Also for supporting our teammates in store.

‍

to help them to also find the right product so that they can sell it well. But also from my past experience, I think it is very close to also where how I learned to support data to actually generate business value closely was from my basically my experience in IKEA, for example, one of my most transformative achievement was similar, enhancing the omnichannel experience through context aware search and recommendation systems.

‍

For that, we basically used a knowledge graph, similar tech stack, to understand the relationship between the products that we manufacture for IKEA, it was furniture, and bringing the home decoration knowledge to it, connecting with what customers want. So this was together with AI algorithm, we were really able to...

‍

provide business the value of how do we improve the sales, but also how do we get the improved customer engagement for that matter. But beyond the number, it really helped us to foster a unified business understanding throughout the company. And that also helped how we deliver even more value, not just stick to just one.

‍

Tarush Aggarwal (10:59)

When you're delivering such a project like this, what are some of the guardrails you need to set up? And also, what is the success credit like? How do you measure the impact of projects like this?

‍

Anindita (11:11)

Yeah, so it depends on multiple things like it depends on the use cases, for example. And for e-commerce example that I gave you, it was all about how do we improve the conversion? How do we measure the click through rate? For example, there are many matrices that of course we we define together with the business, which is always challenging to coming to the common understanding of matrix.

‍

but majority of them were always linked to a conversion rate and customer satisfaction. So NPS and conversion, these two were really major factor. They're actually intertwined as well. The other part was also that we were looking into how do we intelligently manage our pricing structure, which also said, okay, how do we understand the competitive brands and then bring the market knowledge to

‍

improve our pricing strategy to be relevant, but also give the right service to the right people at the right time, at the right cost. And that's all of them are linked to indeed operational excellence. How do we improve, for example, teammates to support, to provide the right recommended product to customers? That is also operational excellence. How quickly they can do that customer service also has the similar kind of challenges.

‍

and, yeah, it's also how do we serve business people and eventually planet as well. So circularity being one of the crucial criteria in both the companies that I'm talking with data plays a massive role there to trick and trace. Yeah.

‍

seconds.

‍

Tarush Aggarwal (12:46)

Talking about data, what's the culture like? You mentioned in general data is fairly lean, right? We're talking about 300 people in a company of 100,000. You have to be an incredibly efficient, pound for pound, probably one of the most efficient sort of data teams. What's the culture like in the data team and how do you really interact with the business?

‍

Anindita (13:08)

Yeah,

‍

so we have, and that is not just Decathlon also in the past, we are following the four in the box principle, right? So we have the product team, we have engineering team, and the UX UI, which is basically user experience and focusing on how do we become more human centric when we build a product.

‍

So product and the user experience team is really playing a massive role these days, which is, think, a transformative as well, to make sure that the engineers and the data engineers, be data or software engineers, we just don't sit on the ground floor and then build some solution and then hope for the best that it brings the value. But actually making it is that it is really aligned with the product team, the value of the business.

‍

along with that it is really desirable for the users to use. These two factors are really becoming super crucial. So four in the block, it's a game. It's a really a game. And everybody plays a massive important role to make sure that whatever we build, we are actually aligned with the portfolio. And the portfolio is basically aligned with the business strategy.

‍

making sure that at the end, whatever projects or portfolio programs that we are working on is actually associated with one of the business strategies, which are of course associated with our value. So I think that OKR objective alignment is also plays a very important role to make sure that whatever we are building is essentially bringing our value to the company. So we have a portfolio OKR culture.

‍

Tarush Aggarwal (14:43)

Awesome. Talking about portfolios and OKR, what one of the challenges you've faced in a new role like this and being able to drive data strategies at such a large organization? What are one of the challenges you face which you are currently working on solving?

‍

Anindita (15:00)

There are certain fundamental challenges like data quality, but there are also kind of there are certain new types of challenge that is also emerging, especially right now. I'm part of the AI innovation and trust team. While AI innovation talks about the future growth building the foundation for future growth at scale, trust talks about the building the foundation of the governance.

‍

So this is the challenge that is in because we talking about also the upcoming EU AI Act, which talks massively about responsible AI, trustworthy data usage for the company, which cannot happen unless we talk about a governance framework, right? So I would say for now, right now, one of my main challenge for knowledge activation part is

‍

turning raw data into actionable insights that can be leveraged for AI innovation at scale. And that includes generative AI, which is a different level of complexity from technology side that we are talking about. So as we know that better AI isn't about just more data, right? It is about the quality and the connected data. So this is a complex challenge for several reasons, apart from data quality,

‍

which definitely we are focusing on ensuring the content that we are producing, but we are using, we are probably sometimes buying third-party data, but we have a massive amount of first-party data as well that are really consistent, the quality of it really consistent across domains. And we have assigned accountability to make sure that we just don't keep on saying the quality is bad, but keep improving it.

‍

So running that whole loop of continuous improvement is super important and becoming it's challenge at the scale, as we've already pointed out. The second one is data findability and preparation. So our data is scattered across multiple system globally. And of course, which makes us data aware, data centered, but not data driven, right? Because we are not really using data efficiently.

‍

to really make the right decisions that actually generates action. So finding the high quality data, which is really very much challenging. And the third and most important one is the people and the culture. So data transformation is just not about a technical journey, right? It's about the process, it's about the culture. We are empowering our teams to work with data.

‍

in new ways while making sure we don't fall back into the data silos again. That requires considerable data literacy effort. indeed, it's while we are trying to solve a lot of problem, I think the core of it that we are working towards making data more trustworthy and accessible by one of the principles, fair principle, of course, that we are

‍

talking about a lot, which talks about findability, accessibility, interoperability, and reusability. And that means that we also need to understand what we have. So describing data is really becoming more and more important, especially when we talk about also innovation at scale with AI.

‍

And being part of the innovation team actually is also more challenging because for me, it's about the balancing act between innovation at scale with responsibility while keeping the growth. So making sure that we are driving impactful changes while staying true to our mission. And we are building data literacy, creating quality platform and fostering a culture where data is really heart of every decision making. So it's a challenging journey.

‍

but it's a journey and we are on it all together. takes a village.

‍

Tarush Aggarwal (18:48)

Thank you for sharing your wisdom with us, Alindita.

‍

Anindita (18:51)

Thank you.

‍