7 Years, No Revenue: How Sigma Computing Hit $100M ARR

Most founders obsess over getting to $1M ARR fast. Rob Woollen took seven years to get there — then tripled revenue three consecutive years to break $100M. His story reveals what patience, radical product decisions, and one serendipitous Snowflake meeting can do for a startup that refuses to quit.

Key Facts

• Years before first significant revenue: 7 years (Rob Woollen)
• Revenue growth cadence: Tripled three years in a row (roughly $3M → $9M → $27M+) (Rob Woollen)
• Current ARR: Over $100M (Rob Woollen)
• Pre-pivot ARR at time of UI rebuild: Approximately $800K–$900K (Rob Woollen)
• Customers today: Over 1,400 (Rob Woollen)

Why Speed to Product-Market Fit Is the Wrong Obsession

Getting to product-market fit fast doesn't predict ultimate success. Founders who optimize for speed often build copycat products that sell easily early but can't scale. Rob Woollen argues that targeting a big market means a longer, harder road — and that's worth accepting if the destination is a large, independent company.

The startup ecosystem celebrates speed. Zero to $1M in six months. Zero to $10M with AI. But Rob Woollen, co-founder of Sigma Computing, offers a direct counterpoint: he spent seven years generating essentially no revenue before hitting the product-market fit that fueled triple-digit growth.

Host Pablo Srugo framed the tension well: companies like Nike and Shopify also took years to find their footing, yet became dominant. The implication is that early velocity and ultimate success aren't as correlated as the ecosystem suggests. Woollen agrees — and adds a warning for founders tempted to chase quick wins.

The risk of optimizing for speed is building something easy to sell early but impossible to scale later. Copycat products, feature-light tools, and small-market solutions can generate early revenue but rarely become public companies. Woollen and his investors aligned from day one on a single goal: build a large, independent, public company. That alignment gave them the patience to reject shortcuts.

"You can get to one or five quickly and then stall out. And I would not trade for that, given where we've been able to get to." — Rob Woollen

"To get to a large, let's say, public company, you have to go attack a big market, a big problem. It's very hard to do that. It's very hard to find product market fit there." — Rob Woollen

The Founding of Sigma: From Salesforce CTO to EIR with a Whiteboard

Rob Woollen left a CTO role at Salesforce to join Sutter Hill Ventures as an entrepreneur-in-residence, where he met co-founder Jason France. Both were independently circling the same problem: most business users couldn't access the enterprise data sitting inside cloud warehouses without SQL expertise.

Woollen spent nearly seven years at Salesforce as CTO of the platform and head of product management. The insight that launched Sigma came from watching Salesforce itself: a company sitting on massive cloud data, yet largely running on spreadsheet guesses and executive hunches. 'The vast majority of people that were making the important decisions couldn't really leverage all of this data that they had,' Woollen recalled.

Rather than building inside Salesforce, Woollen joined Sutter Hill Ventures as an EIR — a structured but open-ended role where he had a whiteboard, limited pay, and pressure to find a problem worth solving. There, he crossed paths with Jason France, who had spent his career building database infrastructure and was advising Snowflake on query optimization. The two were approaching the same problem from opposite directions: Woollen from applications downward, France from infrastructure upward.

They raised an $8 million Series A before writing a line of production code — a reflection of their track records, not their traction. But their investors were clear: don't spend the money until you know what you're doing.

"I watched, at least 10 years ago when I was there, the company was still largely run on hypotheses, you know, spreadsheets with some guesses in it." — Rob Woollen

"The investors gave us an $8 million Series A out of the gate with roughly what I've told you here. Long track record, big idea, confidence in that they knew us so well." — Rob Woollen

Two Prototypes, Two Years, Zero Product-Market Fit

Sigma's first two major prototypes both failed. The first tried to surface AI-driven insights automatically — users found them irrelevant. The second built a vertically integrated analytics engine with visible code — non-technical users assumed it wasn't for them. Both taught Woollen that end users need empowerment, not automation.

The first prototype aimed to eliminate the analyst entirely: feed it a database, and it would surface the most important questions to ask. The idea was compelling in pitch meetings. In practice, users found the insights either obvious or irrelevant. As Woollen explained, a system flagging Super Bowl Sunday as an anomaly at DraftKings would baffle anyone who works there — but look fascinating to a machine.

The second attempt was more sophisticated: a vertically integrated stack with a custom interface, programming language, and analysis engine. It gave power users fine-grained control. But when non-technical users saw code on screen, they assumed it wasn't meant for them. 'I remember at one point we were showing it to someone, and they were like, Oh, maybe you should show this to our engineering department,' Woollen said. 'And I was like, oh, that is not who we're targeting.'

Two founding employees left after year one. The team shrank from five to three. Woollen admits this was a genuine low point — people who knew the product intimately from the inside were effectively voting no with their feet.

"Every time we would demo it or get people using it, it was sort of, you know, you can sort of tell that polite feedback where people are like, yeah, it's kind of interesting. And then it doesn't turn into them actually being users." — Rob Woollen

"We threw everything away and started over." — Rob Woollen

• Prototype 1: AI-surfaced insights — users found them obvious or irrelevant.
• Prototype 2: Vertically integrated stack with visible code — scared off non-technical users.
• Two early employees departed after 12 months with no breakthrough.
• Key lesson: end users know their domain; tools must empower them, not replace their judgment.

The Snowflake Meeting That Changed Everything

Three weeks before a scheduled meeting with Snowflake's CEO, Woollen's co-founder realized their existing product was wrong for the pitch. They rebuilt the interface from scratch — a spreadsheet-style grid that translated user actions into SQL and ran queries directly against cloud warehouses. Snowflake's CEO said he wanted to use it himself.

By early 2017, Sigma had spent over three years building products that weren't working. Their investors, who had also backed Snowflake, arranged a meeting with Snowflake's leadership to explore a potential partnership. The problem: Sigma's product at the time had nothing to do with Snowflake — it ran its own analysis engine entirely separately.

Woollen's co-founder flagged the obvious issue: telling Snowflake customers to pull their data out of Snowflake and into a separate system wasn't going to land. So with roughly three weeks until the meeting, they built a completely new interface — a spreadsheet-like grid where every user action translated to SQL, executed live against cloud warehouses like Snowflake. No separate engine. No data extraction. Just a familiar interface on top of the warehouse.

The reaction from Snowflake CEO Bob Muglia was immediate: 'This sounds great. I want to use this myself. When can I use this?' That visceral, personal demand — not polite interest — was the signal Woollen had been waiting for. Within months, Sigma had signed up for a partner account and begun the security certification work needed to connect to real enterprise data.

"We built a brand new interface, one that looks like a tiny part of our product right now. It had a grid. It sort of looked like a spreadsheet. But we said, what if instead of us building all these different pieces, we just gave you something that looks kind of like a spreadsheet?" — Rob Woollen

"I remember Bob said to us, 'This sounds great. I want to use this myself. Like, when can I use this?' And that was a very different reaction than, 'Hey, that looks kind of interesting.'" — Rob Woollen

From $800K to $100M: The Growth That Followed

After rebuilding its UI in 2020 and launching the new version in mid-2021, Sigma tripled revenue three consecutive years — from roughly $800K–$900K ARR to over $100M. The key unlock was a unified 'workbook' interface that let users analyze and present data in the same place.

Even after the Snowflake meeting, Sigma wasn't done iterating. Revenue was stuck around $800K–$900K entering 2020. A salesperson closed three deals in three consecutive days in April 2019 — Woollen thought they'd arrived. The next quarter: zero new customers. The product still had friction. The market for cloud data warehouses was still maturing.

The final UI rebuild happened in 2020 and launched in June 2021. The old product separated analysis from presentation — a workflow Woollen describes as accidentally recreating the painful back-and-forth between business users and analysts they were trying to eliminate. The new 'workbook' unified both in one interface. The response from the market was immediate.

Revenue then tripled three years in a row. Woollen tracks a now-famous Q2 data point internally — a quarter of zero new customers — as a reminder of how close Sigma came to stalling permanently. Today the company has over 1,400 customers and is targeting a doubling of revenue in the current year.

Woollen's personal signal of true product-market fit: 'We were signing customers that were paying us as much or more than all these ones that I had personally spent all these hours on hand-holding. And I'd never even heard of this deal. It just happened.'

"I think three years in a row tripled the company. So, you know, it was like three, nine, twenty-seven, something like that." — Rob Woollen

"What we found was that you spent a huge amount of sort of friction and time going back and forth... it had the feel of, like, I'm making something, and now I need to go to this other place and see what happened." — Rob Woollen

• Pre-PMF ARR: ~$800K–$900K heading into 2020.
• New workbook UI launched June 2021 — unified analysis and presentation.
• Triple growth three consecutive years followed the launch.
• Over 1,400 customers and $100M+ ARR as of mid-2025.
• PMF signal: large deals closing without founder involvement.

Should You Start a Company? Rob Woollen's Honest Answer

Woollen's answer is almost always no — not because startups are bad, but because anyone who can be talked out of starting one probably should be. The drive has to be innate. And founders must genuinely accept that most startups fail, requiring many things to go right including luck.

With seven years of uncertainty, nights questioning whether he should have stayed at Salesforce, and multiple full rebuilds behind him, Woollen has a clear-eyed view on who should start a company. His advice cuts against the celebratory narrative that surrounds entrepreneurship.

His logic is almost a filter: if hearing 'no' from someone like him gives you pause, you're not ready. The conviction has to come from inside — independent of validation, market timing, or investor enthusiasm. For those who have it, he recommends picking a problem you can imagine working on for ten or more years.

He also acknowledges the specific psychological challenge of co-founder dependency: 'I feel like both my co-founder and I, at different points, would be on some high. And often when one of us was low, the other one would be high, or vice versa.' That oscillation — taking turns carrying the company through doubt — is an underrated reason why solo founding is so hard.

"If just talking to me and I tell you no, you're convinced that, oh, maybe I shouldn't do this — you should definitely not do this." — Rob Woollen

"You have to pick something you love. You are truly excited about it. You truly can imagine spending 10 years or more solving this problem." — Rob Woollen

Sigma's Product Iterations: What Failed and What Worked