“I understand data transformation, but I’m no expert. My strength is more in strategic planning and communicating across departments. But I joined the Coalesce POC and together we rebuilt a fact table. About 30 minutes in, I realized how straightforward it was—even I could do it.”
Heavy lifting without a spotter
Challenges
Limited engineering resources, with only a handful of data engineers to manage many complex processes
Connecting new data sources to Snowflake manually was tedious and time-consuming
Hand-coded data transformation using long SQL scripts was inefficient and unsustainable
Single production environment with no separate development or QA environments made changes feel risky
Christopher Rüge is Head of Business Intelligence and Data for RSG Group. When he joined in 2022, there was just one data engineer for the entire global group. “He was focused on the reporting on our membership management software,” Rüge recalls. “So this meant reports on membership adds, cancels, net growth, and revenue. But with just this data, seeing the bigger picture wasn’t possible.”
Today Rüge’s team consists of three data engineers, all of whom wear multiple hats as de facto data analysts and data scientists: “They’re responsible for getting data from our sources into the data warehouse, cleaning that data up, doing exploratory data analysis, building the data model and business logic, building out reports, and often providing ad hoc analysis on those reports.” In the beginning, these data reports were mainly for the CEOs, says Rüge, but over time the audience for such reports broadened across the organization, and the team now provides reports to a number of different departments.
According to Rüge, much of the data they manage comes to them via Magicline, a membership management platform designed for fitness companies that provides data on everything from contracts and advertising to revenue, payment runs, and debt collection. It also tracks member check-ins, data from member apps, and even maintenance data on the fitness equipment and gym facilities. In addition, the team receives marketing data from its customer engagement platform, Braze.
The earlier RSG data stack was very much a Microsoft Azure environment. “Our data ingestion pipelines ran through Azure Data Factory and were orchestrated using SQL Server Integration Services,” says Rüge. “We also had a few legacy PowerShell scripts for some SaaS services, but those were limited in scope and lacked proper validation. Our data model was built in Analysis Services, with Power BI for visualization.” RSG was also using Snowflake as a data warehouse solution, but it was a very basic on-demand setup.
Eventually the team migrated to Snowflake for most production data requirements. However, the team’s biggest challenge was a lack of resources, which made it difficult to ingest new data sources into Snowflake, let alone model the data and transform it into something usable. “We were busy just keeping what we had running and stable,” says Rüge. “So we couldn’t really get new data into our data model. We recognized that manually building pipelines was tedious work, and we were already doing data transformations by writing all the code in very long SQL scripts in Snowflake. We knew we needed to start automating some of these processes.”
Another major challenge the team faced in their Snowflake environment was the lack of separate development and QA environments—everything had to be done directly in production. This created a constant sense of risk, where even small changes could have unintended consequences. “It often felt like we were performing open-heart surgery,” says Rüge, describing the high-stakes nature of making updates in a live environment. With no room for safe testing or iteration, it became clear they needed to rethink their entire setup and invest in a modern, scalable data stack that would support faster development and long-term growth.
Building a leaner, meaner data stack
Solution
Went from a Snowflake on-demand pricing model to an enterprise-level plan
Brought on Fivetran for data ingestion
Adopted Coalesce for data transformation after a successful POC
One of the first changes Rüge made was to uplevel Snowflake from an on-demand pricing model to an enterprise-level plan. He then got input from some solutions engineers and architects to strategize the best way to build a more agile data stack from the ground up. “Because we had very limited resources, we knew we needed to work smarter and automate everything we could,” he says.
Since RSG was already on Snowflake, Rüge decided to bring on Fivetran for data ingestion, and after a successful proof of concept, added Coalesce as their data transformation platform. “From a product standpoint, it was an easy choice to make,” he says of his decision to go with Coalesce. “I understand data transformation, but I’m no expert. My strength is more in strategic planning and communicating across departments. But I joined the Coalesce POC and together we rebuilt a fact table. About 30 minutes in, I realized how straightforward it was—even I could do it.”
Senior Data Engineer Olaf Weik is that lone engineer who had been keeping the lights on before Rüge and the rest of the team came onboard. “For a long time I had been all by myself, which was one of the main reasons we had started thinking about rearchitecting our data infrastructure,” he says. Because of his long history with the company and his core role on the data team, Weik’s buy-in with Coalesce was key. “At first, I wasn’t completely convinced the time, effort, and eventual costs would be worth it even though the POC was solid,” he admits. Weik explains that he’s more of a hands-on learner, and so really needed to use the platform for himself before he could get a true sense of its capabilities.
“The turning point came when I started migrating everything from our current infrastructure into Coalesce,” he recalls. “For example, with our highly complex customer dimension, I could see exactly where data was coming from and what would be impacted downstream. That visibility gave me real confidence. Being able to explore dependencies and understand changes in real time while developing—that was my ‘aha’ moment.”
Pumping up performance
Results
Lowered infrastructure costs by centralizing data management with Snowflake, Coalesce, and Fivetran
80x faster to bring raw data into the staging layer, going from 20 hours of manual work to just minutes
Automated documentation and metadata-driven architecture support GDPR compliance and reduce manual oversight
3x faster to onboard new engineers thanks to Coalesce’s intuitive workflows
10x faster at providing actionable data insights to leadership team
Thanks to support from the Coalesce team during the initial setup, says Rüge, the team was able to quickly and efficiently migrate all their existing code into the Coalesce environment in just a few days: “We broke it down into smaller, more manageable objects, which made it much easier to revise logic and validate the data model.” He notes that one of the best parts of setting up Coalesce was how easily they were able to establish separate environments for development, QA, and production, which means the team can now test and iterate safely without risking the live system. Or as Rüge puts it, “No more open-heart surgery!”
“Migrating to Coalesce was far easier than I expected,” says Weik. “For example, adding a new source like Magicline, which provides over 100 tables, used to take us hours of manual work—selecting each table, creating views, and ensuring everything was mapped correctly. With Coalesce, it was done in under five minutes. The tool automatically created the views, mapped the column lineage, and streamlined the entire process. In the past, this would have taken at least a full day, and likely more, with a much higher risk of human error. Coalesce made it effortless, and that was the first moment I realized just how much time we’d saved.”
“Migrating to Coalesce was far easier than I expected…. With our highly complex customer dimension, I could see exactly where data was coming from and what would be impacted downstream. That visibility gave me real confidence. Being able to explore dependencies and understand changes in real time while developing—that was my ‘aha’ moment.”
—Olaf Weik, Senior Data Engineer, RSG Group
RSG Group operates in regions with strict data privacy regulations, such as GDPR in Europe. As Rüge explains, “When we were first rebuilding our data stack, one of the most urgent needs was improving data security and data governance.” He explains that using Coalesce to split their workflows into separate environments—and being able to use that same data without replicating it—was a big step forward. “Our Legal team was happy to learn that Coalesce works on top of Snowflake and only with metadata, so we don’t have any GDPR compliance issues. And in terms of data governance, our former CFO—one of our CEOs today—was excited by Coalesce’s automated documentation and column lineage features, which allows us to validate data and ensure accuracy across the company.”
One of the most dramatic improvements the team has seen was the speed with which they can integrate new data sources, using Coalesce to bring raw data into the staging layer. What once took 80 hours now takes just 30 minutes thanks to the combined power of Fivetran and Coalesce. “We initially only had a few data sources integrated,” says Rüge. “My data engineers tracked their time in Jira, so we were able to get a really good comparison between the two approaches.” Coalesce specifically reduced what was previously 20 hours of work just getting data into the staging layer down to a few minutes, he says. “It’s definitely a huge time-saver—it’s probably close to 80x faster with Coalesce.”
Rüge says the ease with which he can now onboard new engineers has been a game-changer. “Before we had Fivetran and Coalesce in place, onboarding a new engineer took roughly half a year. But the latest engineer we brought in was able to gain a full understanding of our data model and data flows in just two months. Imagine trying to explain a 500-line DML script you didn’t write yourself—that’s a lot harder than simply using Coalesce’s intuitive graphic interface, which offers column-level lineage and breaks transformations into smaller, manageable steps.” Adds Weik, “I’m really glad we now have this kind of solution in place. I no longer worry about onboarding new developers or external consultants—even for large, multi-month projects.”
“Before we had Fivetran and Coalesce in place, onboarding a new engineer took roughly half a year. But the latest engineer we brought in was able to gain a full understanding of our data model and data flows in just two months.”
—Christopher Rüge, Head of Business Intelligence and Data, RSG Group
With their newfound ability to integrate all their data sources, the team has finally gained a 360-degree view of the customer and the business. “We now understand the full customer lifecycle,” says Rüge. “Before Fivetran and Coalesce, that wasn’t something we could even consider. We’re now able to do such things as incorporate geospatial marketing to help our expansion team decide where to look for possible new gym locations.”
Rüge says that while the RSG leadership team may not know every technical detail, they’re fully aware of the infrastructure shift—and the impact is clear. “We’ve presented the data strategy to them along with a glimpse of what’s coming next,” he says. “They’ve seen a huge improvement in data availability and the insights we can deliver. Our time to insight has drastically reduced—we’re at least 10 times faster at delivering insights if not more.”
In many cases, teams are seeing meaningful data for the first time. “We used to be limited to basic metrics: check-ins, revenue, cancellations,” he explains. “Now, departments like marketing can finally target the right customers through the right channels. We’ve introduced churn prediction and are building a Cortex AI agent with Snowflake that lets users interact with data directly in Microsoft Teams. It only works because of the structured, well-documented data model we built with Coalesce.”
“We now understand the full customer lifecycle. Before Fivetran and Coalesce, that wasn’t something we could even consider. We’re now able to do such things as incorporate geospatial marketing to help our expansion team decide where to look for possible new gym locations.”
—Christopher Rüge, Head of Business Intelligence and Data, RSG Group
Weik says that the team has recently brought on Coalesce Catalog as well, with the hope that the AI-powered data catalog will save the team time just like Coalesce and Fivetran have. “With such a small team, we don’t want to spend hours on documentation, governance, or answering ad hoc questions,” he says. “Ideally, people will just ask something in Slack or Teams—like how much revenue we made yesterday—and get a human-readable answer from the chatbot. Long-term, we want an infrastructure that doesn’t rely solely on the developers. It should be automated, documented, and easy to maintain, even if someone leaves or takes a vacation.”
Where before the team lacked a sustainable, scalable platform, Coalesce has helped them build real muscle behind their new data strategy going forward. With well-documented, metadata-infused code as the core, they’ve laid down a rock-solid foundation—one that’s built to flex, scale, and go the distance.
“I no longer worry about onboarding new developers or external consultants—even for large, multi-month projects.”
—Olaf Weik, Senior Data Engineer, RSG Group