Bridging Data Gaps: An Inside Look at a Major Retailer’s Cloud Migration

Migrating an ERP system is no easy feat, especially when you’re moving from on-prem SAP ECC to cloud-based S/4 HANA. You can think of it as playing 3D chess—there are layers upon layers to consider, and every move impacts future steps in ways you can’t always predict. 

We recently sat down with the Head of Data at a prominent retail conglomerate to hear firsthand about the data challenges she and her team faced and how they tackled them with illumex and its GenAI-based capabilities.

Q: Can you walk us through the challenges your team encountered at the start of this migration?

Head of Data:  Well, from day one, it was clear that this was far from a straightforward data transfer. It wasn’t as simple as moving systems from point A to point B—we were rethinking our whole data strategy. 

With the cloud’s pay-as-you-go model, every piece of data we bring over has a cost implication. On one hand, we wanted to make sure that all our active business workflows could transition smoothly to the new environment. On the other, we didn’t want to bring along any unused modules from our on-prem setup. So, we had to be very careful about what we migrated.

To make things trickier, our legacy system had been built up over decades. Documentation was missing, there were data gaps in our master tables, and many people who’d initially worked on the system had moved on, taking with them crucial context about the data. And since SAP’s coded table names aren’t exactly user-friendly, it was hard to get buy-in from business stakeholders who couldn’t see how the data connected to their daily work.

Q: So, was a simple “lift-and-shift” approach even an option here?

Head of Data: Not at all. If we’d moved all our data workflows as-is, the costs would’ve been astronomical. The cloud operates very differently from our on-premises setup. In a traditional “lift-and-shift” migration, you can move over everything as it is. But with the cloud, you’re billed based on what you store and use. So, dragging unnecessary data along just increases the monthly bill. 

We needed a more strategic, lean approach to migration—one where we’d only migrate the essential workflows and data assets.

But to do that, we first had to understand which data was actually valuable, map out its current usage, and then design an entirely new system architecture based on those insights.

When you’re dealing with old systems where data is labeled in code with no clear semantics or context, that’s easier said than done. We had to know what each piece of data and every data set meant before deciding how to prioritize it.

Q: How did you begin to handle these documentation gaps and data clarity issues?

Head of Data: It became clear we’d need more than traditional tools. We needed something that could give semantic context to our data—essentially translate all those coded table names into meaningful business language that everyone, from analysts to executives, could understand.

That’s when we decided to work with illumex. Using its Generative Semantic Fabric with AI-driven active metadata mapping, illumex helped us map out our workflows, see usage patterns, and pinpoint which data mattered most.

With automated documentation, we could finally auto-generate descriptions for our SAP tables, which was a huge relief. It replaced the documentation we had lost, and on top of that, it also provided business context for each data element. It basically made it searchable in plain language. This made our data accessible and usable. It was crucial because our business stakeholders could now see how data elements fit into their workflows and decision-making.

Surprisingly, the platform filled in gaps we didn’t even know existed. For example, it could identify areas in our master tables where essential fields were unpopulated and then flag them for us so we could address those before moving to the cloud. So, we were actually elevating the data quality while migrating data.

Q: Was there a specific illumex feature that was especially helpful in keeping the project on track?

Head of Data: One feature that really stood out for us was the auto-tagging of data assets. This helped us identify critical fields quickly and consistently. To manage this on our own would have been impossible at our scale. With auto-tagging, it was easy to keep things in line with our new data architecture.

Then, there was the translation of SQL queries into plain language. This might sound small, but for our business stakeholders, it was a revelation. Suddenly, they could understand what a technical data element did and why it actually mattered—without needing a data dictionary to interpret each term. This helped bridge the gap between our business and tech teams so that everyone was working toward the same goals.

And I have to mention the semantic search. With the amount of data we’re dealing with, search performance is critical. At first, response times were slower, but illumex quickly optimized that for us. Now, search works lightning-fast—even with our massive data volumes. Plus, it’s actually fun using the platform. This kind of speed and usability has been a huge win for us.

Q: How did your data strategy shift post-migration? Any major changes in how you manage data now?

Head of Data: Oh yeah, absolutely. Before, we were often guessing—or worse, overlooking important information—because we simply couldn’t access or understand our data well enough. Now, we have a clear and thorough view of our data landscape. This has reduced the risk of errors and misunderstandings.

The advanced search capabilities, in particular, have saved us a massive amount of time. Imagine manually sifting through millions of data entities to find one critical asset. It used to take forever, but now it’s almost instantaneous. 

With this new level of efficiency, we can now build datasets for new products, run in-depth analyses, and easily manage assets. We’re actively using the data to gain business insights as opposed to just maintaining it. This means we spend a lot less time digging and more time focusing on strategy. And that has been a big change for us.

Q: Looking ahead, how does this approach set your team up for future success?

Head of Data: Oh, we’re definitely ready for the future. 

Today, we have a great foundation. This includes clear, understandable documentation (that auto-updates continuously) and quick access to key data. With that, we’re in a great spot to grow and innovate. Plus, our business and technical teams can actually understand each other, so decisions happen faster and are way more aligned with our strategic goals.

Knowing that our data is solid and trustworthy opens the door for us to explore new data products and AI-driven insights. It’s hard to understate the value of that clarity and the freedom of not having to second-guess your data. And we’re always up-to-date, with illumex managing ongoing updates to documentation and metadata—so we don’t risk slipping back into the mess we started with. 

We’re already looking at new opportunities in data monetization, predictive insights, and expanding our business intelligence capabilities. It’s an exciting time!

Q: Finally, what advice would you give to other data teams tackling similar challenges?

Head of Data: My best advice? Start with a strong footing. Invest early in tools that bring semantic context to your data, especially if you’re dealing with legacy systems. Traditional cataloging solutions just don’t cut it when you’re facing the layered complexities of cloud migration.

Automating as much as possible is essential, too. When we tried to handle documentation manually, it took a huge amount of time and was almost impossible to keep up with. Also, having a system that translates technical data into something accessible to all stakeholders is truly valuable. It creates alignment, saves time, and prevents misunderstandings that can be so costly in projects like this. So, bridge that gap!

And finally, it might sound like a cliche, but work smarter, not harder. Set yourself up for long-term success by turning data into a reliable, usable asset rather than just a burden to be managed.