How Bilt Crafted a Data Dream Stack—Starting with Sigma

At Bilt, speed isn’t a luxury—it’s a requirement. When you're building a loyalty platform that turns rent into rewards, every product launch, iteration, and insight has to move in sync with the business. 

That demand for velocity shapes how the entire business runs. Bilt is organized into cross-functional “pods”, or small, autonomous teams that operate like startups within the startup. Each pod owns a distinct piece of the product, from loyalty to payments to travel. It’s a setup built to ship fast. But without the right data tools, that doesn't work.

“We pride ourselves on how fast we launch features. But our data tools weren’t keeping up. That had a visible impact on our product strategy.”
_{- James Dorado, VP of Data, Bilt}

That was the challenge facing Bilt’s data leaders. As the company scaled, their existing BI tools couldn’t keep up with the speed of decision-making. We sat down with James Dorado, VP of Data at Bilt, and Ben Kramer, Bilt’s Senior Director of Analytics, to hear how the team rethought their data stack and why Sigma became key to the way Bilt moves. “We pride ourselves on how fast we launch features. But our data tools weren’t keeping up. That had a visible impact on our product strategy,” said James Dorado.

Too fast for the old stack to handle

Before Sigma, Bilt’s data team was juggling speed, scale, and expectations without the right tools to support them. The business was growing fast, and product teams were shipping features rapidly. Pods needed data to make decisions on everything from redemption behavior to payment operations—but the tools meant to enable them were holding them back.

Bilt’s legacy BI platform was built around a rigid semantic layer. That was fine for answering the same 10 questions, but brittle in the face of new ones. “If your business question didn’t fall into the 60% that the static semantic layer could answer,” said Ben Kramer, Bilt’s Senior Director of Analytics, “you were left with custom SQL, ad hoc queries, or Google Sheets. We avoided using the tool because it slowed us down.”

The data team had two choices: spend time hardcoding logic into a model that might never be reused, or respond fast with a manual workaround. That usually meant writing queries in BigQuery, pasting results into a Google Sheet, and sending them off by email or Slack.

“We avoided our legacy BI platform because it slowed us down.”
_{- Ben Kramer, Senior Director of Analytics, Bilt}

“It wasn’t just inefficient,” Ben said. “It discouraged us from investing in self-service at all.” From James Dorado’s perspective, the technical debt was piling up across layers of the stack. “We were doing double work—first in DBT, then replicating logic in BI,” he said. “It created inefficiency and slowed us down.”

Eventually, people stopped using the BI tool altogether. “There was a ripple effect,” James explained. “The data wasn’t all there, so people stopped trusting it. And we stopped maintaining it because it was too slow to be worth it.”

By his own grading scale, James deemed the performance non-optimal. We delivered answers, but they weren’t reusable. People couldn’t explore on their own.” The team knew they couldn’t keep patching around the problem. They needed a tool that could match the speed of their business, without forcing them to compromise on governance, flexibility, or user experience. That kicked off a serious search for something better.

Sigma: The fit that finally made sense

From the start, Bilt’s evaluation process was shaped by one question: how can we move faster without breaking the systems we’ve already invested in? They weren’t looking for just another BI tool. They were looking for a better way to work.

So, the team evaluated a range of tools but quickly narrowed their criteria. The right solution needed to strike a balance: flexible enough to iterate quickly and support complex questions, but structured enough to standardize core metrics and definitions.

“We were really prioritizing flexibility,” James explained. “Not just avoiding rigid semantic layers—but choosing something that allowed us to prototype fast and support real business impact.”

Sigma offered that blend. “What stood out was that Sigma could handle both the 80% of questions you get repeatedly and the 20% that are more intricate and one-off,” said Ben Kramer, “We didn’t want to be stuck building two different systems for those use cases. Sigma let us do both.”

The team was especially excited about Sigma’s data modeling capabilities—defining relationships and metrics that could be reused across pods and workbooks. They loved that they weren’t constrained by the model, and that when questions didn’t fit neatly into pre-built dashboards, analysts and business users could build what they needed, fast. That creative freedom mattered.

“What stood out was that Sigma could handle both the 80% of questions you get repeatedly and the 20% that are more intricate and one-off.”
_{- Ben Kramer, Senior Director of Analytics, Bilt}

“We needed a tool that could do more than just answer standard business questions,” said Ben. “We wanted to be able to run regressions, map out audience behaviors, and dive deep into customer segments. Those aren’t always things you can templatize—but Sigma makes them doable, even at speed.”

A place the data team wants to work

When Bilt rolled out Sigma, the goal was clear: make the BI tool a place where the data team actually wanted to work. “It was very clear that we’ve achieved that,” said Ben Kramer. Each of Bilt’s pods now operates with its own tailored data model. “We’re working toward ensuring there’s a data model for every pod,” he explained. “That way, teams can self-serve answers through a consistent structure—but still have the flexibility to get creative, either themselves or with our help.”

To support that autonomy, Bilt put clear governance in place from day one. Sigma’s roles and permissions gave them the structure to scale exploration without opening the floodgates. “The data team are the key editors, but anyone can dive in,” said Ben. “People save their own versions, share them internally, and then ask us to review and publish when needed. It’s the right balance of flexibility and oversight.”

“Sigma’s permission model has been a huge improvement. We can grant more custom roles—enough access to be productive, but with guardrails.”
_{- James Dorado, VP of Data, Bilt}

James echoed that sentiment, pointing out how Sigma helped Bilt move past the rigid access controls of legacy tools. “With previous solutions, people who needed access were constantly blocked. We had to rely on break-glass permissions just to get things done,” he said. “Sigma’s permission model has been a huge improvement. We can grant more custom roles—enough access to be productive, but with guardrails. Those people are now way more self-serve. We hear from them a lot less, which, honestly, is a great sign.”

Governance isn’t just about control—it’s about clearing paths for the right people to move fast, while keeping sensitive data protected. “You want to show the right data to the right people without getting bogged down in red tape,” James said. “At the same time, you don’t want the summer intern seeing everything in the warehouse.”

And it’s working. Teams are building their own dashboards and workflows. The Loyalty pod, for example, uses Sigma to monitor redemption patterns in real time. According to Ben, the team can now isolate specific behavior, like who redeemed a new offer, what they’d done before, and whether we changed their habits. That level of detail used to be painful to analyze. With Sigma, the team describes it as fast, flexible, and intuitive.

That spreadsheet muscle memory and the fast track to adoption

But what made Sigma stick so quickly? Familiarity.

“The spreadsheet interface was huge for us,” said James. “Everyone at Bilt has touched Excel or Google Sheets at some point. That muscle memory made it click instantly.”

“People saw Sigma and just said, ‘Wait—I can do that?"
_{- James Dorado, VP of Data, Bilt}

Instead of exporting dashboards to do real work elsewhere, business users could work directly in Sigma—with full flexibility to pivot, filter, and model without SQL.

That shift changed behavior across the org. “People saw it and just said, ‘Wait—I can do that?’ It was immediate,” James said. Finance moved off brittle Google Sheets workflows thanks to Sigma’s input tables and writeback features. And alerts now flow straight into Slack, keeping teams looped in on what matters without constantly refreshing dashboards. “It’s been powerful,” Ben noted. “We’ve cut manual steps and made insights part of the flow of work.”

“With Sigma, we can fine-tune access with custom rules. Power users get what they need to move fast, but we still have control.”
_{- James Dorado, VP of Data, Bilt}

Even governance got an upgrade. “In our old system, getting someone access meant breaking glass,” explained James. “With Sigma, we can fine-tune access with custom rules. Power users get what they need to move fast, but we still have control. And honestly—we hear from them less now. That’s a win.”

Next stop: Full-speed AI, in Sigma

As for what’s next, Bilt’s already deep into Ask Sigma and other AI-powered workflows.

“Sigma’s approach to AI is the right one.”
_{- James Dorado, VP of Data, Bilt}

“We’re early adopters,” said James. “We’ve demoed Ask Sigma at all-hands meetings, encouraged teams to try it, and it’s already delivering. Especially in meetings—someone forgot a stat, needs a quick answer, and instead of pinging the data team, they just ask. And it works. That confidence builds quickly.”

Data apps are on the horizon too. One of the first use cases? Financial modeling—packaging what used to live in local Excel files into live, interactive workflows powered by real-time data. “Our finance team is really excited about what we can build here,” James said. “It’s a big step forward.”

And longer-term, Bilt sees Sigma as more than just a BI tool. Sigma provides support for how they operate—and iterate. “Sigma’s approach to AI is the right one,” James said. “It’s not about replacing analysis. It’s about speeding up the time to insight—and keeping people in the loop. That’s what will stand the test of time.”

This is what modern looks like

At Bilt, momentum matters. The business moves fast, and their data tools need to keep pace—without slowing teams down or adding layers of complexity.

“We think Sigma’s in the best position to push analytics forward with AI."
_{- Ben Kramer, Senior Director of Analytics, Bilt}

Sigma is helping them move faster: fewer blockers, more experimentation, and a clear path from insight to action.

“There’s so much nuance in how we use data,” said Ben. “The models, the logic, the way teams interact with it—it all lives in Sigma now. That context is powerful. And it’s why we think Sigma’s in the best position to push analytics forward with AI.”

That future’s already taking shape. Bilt’s embedding AI into day-to-day decision-making, exploring new ways to productionalize workflows, and turning more of their manual processes into dynamic, real-time systems.

And they’re doing it with a partner who works the way they do. “We don’t want a tool that tells us how to work,” James said. “We want a tool that works how we think. Sigma gets that. It’s flexible, it’s fast, and it’s built by people who think like us. That’s rare.”

Read more about Sigma in financial services here.