You Don’t Need A Better Cube. You Need A Way Out.

Cubes are still everywhere, but not because people love them. They’re stuck with them—buried in old on-prem tech that costs an arm and a leg to maintain, built for a world where you couldn’t explore new questions, only rerun old ones.

Now that data lives in the cloud, those systems aren’t just outdated, they’re slowing teams down. What used to feel like a performance win is now a bottleneck. What once felt powerful now feels stuck.

I’m not here to dunk on cubes—they solved real problems in their time. But I’ve seen what it looks like when teams try to hold those systems together with spreadsheets, exports, and logic only three people still understand. They've kept cubes alive because nothing else came close to the experience or the performance they needed. 

The cube won’t die—but it’s running out of reasons to live

Cubes come up constantly in customer conversations about cloud data warehouses like Snowflake or Databricks. Honestly, I’d bet 90%-95% of large enterprises still use them. Finance and accounting teams still rely on them because they always have. These are users who need consistent answers—month over month, side by side, so they can spot a change.

And historically, cubes delivered that. They were built for it. Back when compute was expensive and storage was cheap, technical teams modeled out every intersection of measure and grain they thought the business might ask for. That’s what a cube is: a pre-modeled, multidimensional structure built to answer predefined questions fast.

But that speed came at a cost. Because if you wanted to ask a new question—or just look at data at a lower level than the cube allowed—you’d need engineers to go back, rebuild the cube, and wait. Sometimes for days. And no one on the business side could explore the data themselves. 

That’s what teams are still running on today.

Even as the warehouse moved to the cloud—into platforms like Databricks and Snowflake—cubes stuck around. Which means many enterprises are still waiting for cubes to rebuild every week or month, just to get updated numbers. In some cases, that rebuild takes two days. That’s two days lost before the business can even begin to ask questions.

No one scales on spreadsheets and hope

A lot of teams already know that cubes are holding them back. But instead of replacing them, they build complex workarounds. I’ve seen entire workflows where people rebuild cubes, pull the data down with Power Query, run logic through Excel macros, export it to a PDF, and send it off to the CFO. That’s not innovation. That’s institutional duct tape.

I recently worked with one of the largest reinsurance companies in the world to fix those workarounds. Their actuaries were modeling billions of rows through an Excel plugin hooked into SSAS cubes—pre-aggregated to the point of collapse. Drilldowns? Off the table. What-if scenarios? Not an option.

Then one division moved to Sigma on top of Snowflake. That’s where everything changed.

They didn’t just replicate the old process, they rewrote it. Same familiar pivot-table interface, but now running live on the warehouse. No rebuilds, no waiting. Teams could drill into the details, define new hierarchies on the fly, even write back forecasts in real time. Instead of spending hours validating a change, they could model it instantly—and move on.

The result? More than a faster system. A smarter one. One that matched what the business needed from cubes—and then blew past it.

Time’s up for yesterday’s tech

Here’s the thing. Cubes were great—twenty years ago. But that tech hasn’t evolved. Companies are still paying to keep the lights on in data centers, patching outdated systems with zero new features. Most are running two or more legacy BI tools, just to support acquisitions or different business units. It’s a mess.

The opportunity now is huge. With Databricks or Snowflake handling scale and performance in the cloud, and Sigma sitting on top, you don’t need the cube anymore. You get the experience you’ve always had—pivot tables, drill-downs, all the structured exploration you’re used to—but it’s live, in the browser, and goes deeper than cubes ever could.

If you're still clinging to cubes, you’ve probably already tried to replace them—maybe more than once. And each time, the business rejected it. Why? Because the experience wasn’t good enough. With Sigma, that changes. You finally get something better. Faster. Familiar, but way more powerful.