Yeti Data

Victor Sherba had spent years running product strategy in the data division at SAP, one of the world's largest enterprise software companies, and before that consulting at McKinsey. He understood intimately the pain point that would become Yeti Data's core thesis: enterprise companies were drowning in customer data scattered across systems, and the traditional solutions (Teradata, IBM, HP) required months or years of expensive consulting and custom development just to integrate and operationalize that data. "This stuff used to take months and years to get the integration right," Sherba recalls. He saw an opportunity to virtualize data connections rather than fully integrate them—let companies describe their data instead of moving it—and reduce setup time from 1-2 years down to 2-3 weeks.

In 2014, Sherba launched Yeti Data with a lean approach. He raised $1.5 million on a convertible note at 4% interest and 10% discount, with a teaser structure (30% discount for early backers, declining over time). For the first 2.5-3 years, the company operated in pure development mode with zero revenue. The core team consisted of just three full-time people, supplemented by consultants brought in for specialized skill sets like security and cloud infrastructure. There was no rush to market—Sherba and his team were building something ambitious: a virtual data warehouse that could unify customer touchpoints and layer AI on top to generate actionable insights about customer behavior and churn risk.

Yeti Data's go-to-market efforts didn't really begin until 2023, four years after launch. Once they did go after customers, the sales motion was pure enterprise: large companies with complex customer data ecosystems and budgets to match. The product's value proposition was powerful enough to command serious ACV. Today, Yeti Data has less than a half-dozen customers, each paying between $250,000 and $500,000 per year. These are companies like Amazon and Walmart (who have built similar internal tools over 4-5 years and tens of millions of dollars)—organizations with massive customer bases and the sophistication to understand the value of accurate churn prediction and behavioral analysis. The all-you-can-eat pricing model with unlimited seats and usage meant customers had no artificial barriers and maximum incentive to feed more data and derive more value.

What worked was patience and a differentiated product. Sherba's insight—that describing data via APIs was faster and cheaper than full integration—proved sticky enough that customers committed to 6-figure annual contracts. The machine learning models that power churn prediction and customer insight generation are genuinely valuable; as Sherba describes it, the system learns and refines itself as new data contradicts its predictions (like discovering a customer takes vacation in August, explaining a gap in purchases). What didn't work, or rather what revealed itself as a gap to be filled by product, was the professional services tail. Customers occasionally asked for custom reporting or additional capabilities; Sherba views this as "a glitch in the machine"—a sign the product isn't yet feature-complete—but the nondilutive revenue is welcome while the team builds those features into the core product.

Yeti Data is at approximately $800K ARR with less than a half-dozen paying customers, and Sherba is confident he'll hit $1M ARR this year. With ARR approaching seven figures, he's preparing for a Series A raise at a $15-20M pre-money valuation, expecting to give up 10-20% of the company on good terms. The company operates at full capacity right now—a constraint that speaks to the strength of demand and the complexity of on-boarding enterprise customers. Sherba's advice to his younger self was simple: "How to be patient." And in Yeti Data's case, that patience—three years of development, a lean team, and a narrow focus on solving one extremely hard problem extremely well—is about to pay off.