Wolf&Grayson EditorialCompanies collect massive amounts of data, but most struggle to turn it into insights that actually help different departments work together. David Martin Riveros has seen this problem firsthand across his career at General Motors, Uber, and Shopee. Now running his own business intelligence company, Icebergdata, he’s become a Google Cloud partner focused on helping organizations finally make sense of all their scattered information.
Implementing Infrastructure to Unlock Insights
Riveros’s path into business intelligence started somewhat accidentally. Working at General Motors doing financial planning, he found himself essentially doing what we now call business intelligence before it was even a thing. “Financial planning was doing business intelligence,” he recalls. “Business intelligence didn’t exist at that time. I would say in 2016, business intelligence was not a common term.” Those early days taught him something important. Back then, “one of the first software programs at that moment was Power BI and it was among the early solutions.” But having the software was only part of the puzzle. After moving through roles at Uber’s analytics department for UberEats, then Cloud Kitchens, and later Shopee, Riveros discovered the real challenges weren’t just technical.
Identifying Barriers That Block Progress
After years of implementing BI systems, he has watched companies make the same mistakes repeatedly. Here are three major barriers Riveros discovered:
Splitting Definitions Across Departments
The biggest problem? “Different teams work with their own copies of the data and build their own reports on definitions that are different across the company.” This creates chaos that most people don’t even realize is happening. “The operations team may have a different definition of what a customer is than what the sales team does,” Riveros explains. “So when you see a report of sales by customer done by the operations team, it could be based on the amount of products sold for the parent companies. And the sales team could be doing it by the amount of cash collected by the child companies.” When this happens, “reports stay locked into departments instead of being leveraged across the company.” Companies end up with different versions of the truth, and nobody can make decisions that actually stick company-wide.
Struggling to Interpret Available Data
But data silos aren’t the only problem. Even when companies get their data organized, there’s still the human element. “Not all customers, I mean internal customers, know how to interpret data,” Riveros points out. “They probably use a dashboard. But even though they have access to it, they may not ask the right analytical questions.” His solution? “You should have at least one BI analyst for every five BI customers that you have inside a department.” Without this support, people either ignore the data or worse, misinterpret it completely.
Forgetting to Use Critical Insights
The third barrier Riveros sees is perhaps the most frustrating. “Insights do not naturally surface where decisions are made. They live sometimes in dashboards and reports or even slides.” People have to actively remember to check these insights, and frankly, most don’t. “Teams usually by default follow their instincts and go back to old processes rather than incorporating new data-driven insights,” he says. It’s having great advice locked in a filing cabinet that nobody opens.
Building Systems That Actually Work
Riveros’s approach centers on three things that have to happen in order.
- You Need a Single Source of Truth – “We always recommend BigQuery, which is a Google Cloud product. I think it’s the best in the market,” he says. Everything that’s “scattered throughout Excel files, CSVs, systems, databases” gets standardized and put in one place.
- Documentation that People Can Actually Understand – “We need to build another sort of assistive dataset that contains descriptions of the variables,” Riveros explains. Think of it as building a dictionary so people know what they’re looking at.
- Automation Keeps Everything Running Smoothly – Riveros describes this as “a process by which you extract from a database—the raw data source—you transform it, you apply standardization techniques and you load it up into your single data source.”
His company Icebergdata recently made a significant shift. “Icebergdata was a data collection company and now we are pivoting into a data understanding company,” he explains. Instead of just gathering information, they’re helping organizations actually make sense of what they have. This change coincided with becoming a Google Cloud partner. “We work on that, on the infrastructure. We set it up for them. We use Google Cloud for it,” he notes. The partnership allows Icebergdata to handle the entire process for companies that don’t have internal BI teams but need these systems to compete. Looking back at his career path, Riveros sees how each role prepared him for this moment. From those early days at General Motors when BI barely existed, to now helping companies implement systems that actually change how they operate, the core challenge remains the same: getting the right information to the right people at the right time.
Connect with David Martin Riveros on LinkedIn to explore more about building smarter BI systems.
