We introduced the data product approach three years ago, inspired by the data mesh framework. While we didn’t adopt the full set of data mesh principles, we embraced the idea of treating data as a product. For us, this means ensuring our data products are self-sufficient, consumable, autonomously discoverable, and exploitable by end users to address business questions.

They can take different forms — dashboards, well-documented and accessible datasets with the right security measures and SLAs, or even AI solutions that leverage data to provide services.

There’s a lot to unpack, but let me highlight a key example. Our customer knowledge data product consolidates everything we know about our customers while ensuring compliance with GDPR.

It is particularly valuable because, unlike OTAs, we don’t just capture data at the booking or reservation stage — we also collect insights from the guest’s stay. This allows us to track the full customer journey and enables a much higher level of personalization.

We’ve packaged this data into a customer data platform (CDP), empowering our marketing teams to segment audiences and autonomously work with fresh and live customer data.

One of the key challenges is capturing data across our numerous brands, hotels, and countries. To address this, we have structured our organization around one of the core pillars of data mesh — domain-oriented data. This means that our central data team is organized by business domain.

For example, we have teams dedicated to loyalty and customer data, corporate data, B2B, and so on. This allows each team to not only be experts in data management but also develop deep expertise in their respective data domains. For instance, the team specializing in loyalty data fully understands what loyalty data entails, what defines a guest versus a member, and how to build KPIs specific to this business function.

Our central team plays a crucial role but is also supported by smaller teams across our hubs and geographies. These localized teams ensure that we remain close to operations, acting as key relays for the central organization. Their role is to make sure that everyone who needs access to data can understand it, use it effectively, and flag any issues or gaps.

Looking ahead, one of our biggest challenges is democratizing the data products. To support this, we are launching a data portal — an internal data marketplace that will allow users to browse and understand different KPIs, explore available data, and assess data quality.

Since not all data products need to be perfect to be valuable — some can still be useful for discovery — we aim to provide transparency regarding data quality and scalability. This way, users can access data with a clear understanding of its readiness for production and its overall reliability.

Our first challenge is data quality. The majority of our data comes from what we call “the stay” — a guest’s experience at the hotel. Collecting this data from the PMS software in each hotel can be very challenging.

In the past, Accor followed a multi-PMS strategy, meaning we have different software systems with varying configurations. To extract data seamlessly, we need to ensure that we have the right connectors in place, accurately retrieve the data without missing any stays or days, and properly map each hotel’s specific configuration to our central warehouse.

This is a complex challenge, especially given the dynamic nature of the hotel business. Hotels frequently change configurations, which requires us to continuously adapt our data extraction and quality maintenance processes. Ensuring that we consistently extract and process high-quality data has been one of our biggest challenges.

The second challenge is data governance. Accor is not primarily a technology-driven business; we are a human-centric business. Hospitality is, first and foremost, about human experiences — the interactions at the front desk, the quality of a room, and the overall stay.

As a result, we are not naturally data-driven with fully established processes and governance structures. Many large companies struggle with data governance, but for us, it’s an even greater challenge. Ensuring clear data ownership, well-defined processes, a single source of truth for critical data, and clear business rules is something we are actively working on. Strengthening these governance practices remains a key focus area for us.

Another challenge is culture. Making data-driven decisions is becoming more common at Accor, and it is increasingly seen as a standard practice. We now have better visibility into our operations, make decisions based on insights, and track the impact of those decisions using KPIs.

However, we still have room to grow when it comes to using data and AI for future-oriented thinking — anticipating what new services, products, or innovations we can introduce. This is an area where we need to continue our efforts.

We are trying to better understand our guests — who they are, their brand preferences, and their behaviors. However, we still lack insights into those who are not choosing to stay at our hotels.

If we want to be highly effective in the future — differentiating our brands and offers in the market — we need to understand this as well. That’s an area we aim to explore further. Not only gaining deeper knowledge about our existing guests but also understanding who isn’t staying with us and why.

The next step is about reflecting on past experiences — both successes and failures — and identifying key lessons.