May 1, 2025 · 41m · mad
Dashboards Are Dead: Sigma’s BI Revolution for Trillion-Row Data
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The MAD Podcast, host Matt Turck interviews Sigma Computing CEO Mike Palmer on how Sigma surpassed $100 million in ARR by replacing static BI dashboards with a cloud-native spreadsheet interface operating directly on trillion-row data warehouses. Palmer breaks down cloud push-down architecture, the consolidation of the modern data stack, and why traditional Text-to-SQL AI tools fail without transparent governance.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 19.7% of the talking time here. How this is scored →
Matt as informed peer 3.5 Guest teaching 4.6 Guest disagreement 3.1 Matt pushing back 1.4
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 36:11 Text-to-SQL as the ultimate hubris of engineering
Mike vehemently rejects the premise of text-to-SQL chatbots in BI, calling them a joke, a terrible idea, and the ultimate hubris of engineering.
Hardest push from Matt ▶ 28:38 Dissecting technical vs governance semantic layersMatt presses Mike on the semantic layer concept, distinguishing between technical query translation and governance to ensure Mike isn't oversimplifying the necessity of a separate product.
Biggest teaching moment ▶ 37:40 Explaining why enterprise AI must assume AI is always wrongMike reframes natural language querying by demonstrating how users ask imprecise questions and why 86% AI accuracy fails enterprise standards without chain-of-thought transparency.
Matt holds his own ▶ 25:56 Citing Looker's LookML acquisition to frame semantic layersMatt demonstrates deep industry knowledge by citing Looker's $2.7 billion acquisition by Google and identifying LookML as the core intermediary layer secret sauce.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Rethinking BI: Why Dashboards and Charts Failed | 2 | 4 | 4 | 1 | Matt opens by quoting Sigma's website about BI being boring and prompts Mike to elaborate on Tableau. Mike forcefully criticizes traditional BI for treating users like idiots with colorful pie charts and compares Tableau to over-engineered Japanese wrapping paper. | |
| Redefining Business Intelligence for Actionable Analytics | 2 | 4 | 2 | 1 | Matt asks baseline educational questions to define business intelligence versus AI and predictive analytics. Mike expands the definition of BI from static historical dashboards to actionable analytics, spreadsheets, and user-written automation. | |
| The Founding History and Incubation of Sigma | 3 | 5 | 1 | 0 | Matt brings up the incubator model of Sutter Hill Ventures and its parallels with Snowflake. Mike details Sigma's origin story, the tragic passing of their early president, and how he stepped in as CEO seven years after founding. | |
| Rapid Acceleration: Growing from $2M to $100M+ ARR | 4 | 3 | 3 | 1 | Matt highlights Sigma's $200M growth round during a dead VC market for traditional BI tools. Mike explains deprecating the old product, driving growth from $2M to $100M ARR, and voices strong opinions against remote work during COVID. | |
| Live Warehouse Queries and Architectural Trade-Offs | 4 | 5 | 1 | 2 | Matt drills into the technical mechanics of building directly on top of cloud data warehouses like Snowflake and Databricks, asking about architectural trade-offs. Mike explains pushdown queries, zero caching, and acknowledges the lack of cross-warehouse federation. | |
| AI Models Moving to the Cloud Warehouse | 3 | 5 | 1 | 1 | Matt asks about spreadsheets-like interfaces and real-time collaboration at scale. Mike explains overcoming database constraints to build a spreadsheet interface without cell references, highlighting how an executive built a 6-billion-row pivot table. | |
| Modern Data Stack Consolidation and the Push for Democratization | 6 | 5 | 6 | 3 | Matt demonstrates expertise by referencing Looker's $2.7B acquisition and LookML, pressing Mike on semantic layers and governance. Mike delivers a scathing critique of Microsoft Fabric as old products in a shiny wrapper and predicts massive consolidation across the modern data stack. | |
| How Cloud Warehouses Will Drive Stack Consolidation | 4 | 6 | 7 | 2 | Matt asks how Ask Sigma addresses text-to-SQL translation in natural language BI queries. Mike forcefully rejects text-to-SQL as the ultimate hubris of engineering and explains how Ask Sigma exposes chain-of-thought logic to handle flawed AI responses. |
Statements from this episode (18)
Mike Palmer: Tableau acts as superficial wrapping paper for data
“And I think that Tableau is wrapping paper in Japan. You really wanted the toothbrush, but somehow what you really got was wrapping paper.”
Palmer: "Download to Excel" was the most common feature in all BI tools
“The most common feature in all BI products was the button that said download to Excel.”
Palmer: Sutter Hill Ventures incubates all its portfolio companies in-house
“And where Sutter Hill's a little bit unique is that Sutter Hill doesn't invest in pre-existing companies. All the companies are incubated directly out of Sutter Hill.”
Sigma had under $800K ARR and 50 customers at year seven
“So I joined, you know, almost seven years into the company. They had less, you know, about 800,000 dollars in ARR, 50 customers and about 50 employees.”
Sutter Hill's Mike Spicer funds companies despite overwhelming failure evidence
“Which is, by the way, a testament to Mike Spicer at Sutter Hill because he is, he has so much conviction in his companies that despite overwhelming evidence that the company is a total failure, he keeps investing in it.”
Mike Palmer requires all Sigma Computing employees to work in-office
“Even in face masks, I required everybody to come back to the office and I still require that today. Which is its own conversation, but I'm not a believer in remote work”
Sigma Computing grew from $2M to $100M ARR in 3.5 years
“We went from a couple of million dollars in June of 21 to a hundred million dollars three and a half years later”
Sigma CEO: Sigma does not cache data, running all queries live
“We do not do any caching. So everything you do in Sigma is live query on the warehouse.”
Palmer: Sigma's query transaction overhead is under one second
“Our customers have up to trillions of records. And our overhead on that transaction is less than one second.”
Palmer: AI models will live directly inside cloud data warehouses
“I think that the models are going to live next to the data in these warehouses, and we're already seeing that.”
JPMorgan built a six-billion-row pivot table using Sigma Computing
“But the fun part is like a reference JP Morgan Who uses our product, and when they connected the first time, the guys, I think, is sort of like, I can swear in here. Holy , holy , I just built a six billion row pivot table, right?”
Palmer: Financial services is fundamentally a data arbitrage business
“If you work in financial services, you're in a data arbitrage business. So the faster you can get to unique data, the better you are at your job.”
Internal study: Enterprise clients touch only 0.1% of warehouse data tables
“So we looked at all of our enterprise customers Table utilization over a 60 day period, and .1% of the tables in their warehouses were ever touched.”
Palmer: Data stack tools will consolidate into fewer categories in 3 years
“I think if you were to fast forward three years from now, you'll see all these categories, many of these categories rolling into smaller numbers of categories.”
Mike Palmer: Microsoft Fabric is a marketing rebrand of legacy software
“And I'm directly accusing Microsoft here. I think their application products are not good in their cloud products. I think fabric is a run it back strategy on office through 65, where they just provide a marketing name to some longstanding old products.”
Mike Palmer: Google's Looker acquisition failed by locking out non-experts
“Google's experiment with Looker has failed, and with Looker, they built a LookML, which by the way, I would say back in 2017, it was probably the best product on the market, but they built a product for experts, and they ensured that it was going to be a produ…”