Percolata

Greg Tanaka started Percolata in 2011 with an ambitious goal: use sensor data to transform retail labor management. He believed that if retailers could accurately forecast shopper traffic and match it with their best sales associates, they could dramatically improve sales without hiring. The thesis was elegant—think "Moneyball for retail associates."

The early years were brutal. Tanaka spent five years (2011-2016) trying different business models, pivoting repeatedly. He initially tried selling monthly subscriptions to sensor data itself, but discovered retailers weren't the real problem: they already had traffic data. The real challenge was making sense of it and translating it into actionable staffing decisions. This pivot moment—from selling data to selling insights—was crucial. By 2016, Tanaka finally cracked product-market fit. He rebuilt the product to become an AI-powered scheduling system that used existing staff and their availability constraints to create optimal shift assignments.

The early sales grind was painful. Tanaka describes sales cycles that stretched across quarters, requiring constant negotiation and persistence. Before PMF, closing deals felt like "pushing a boulder uphill." The company needed integrations with time-and-attendance systems, HR platforms, point-of-sale systems, and workforce management tools—making onboarding complex and enterprise-heavy. Once PMF hit, everything inverted. By 2017, when he closed his Series A with Google Ventures and Draper Menlo Ventures, the sales dynamic shifted entirely. Deals that once took quarters now closed in 1-3 meetings. Inbound inquiries replaced cold outreach.

The pricing model was a masterstroke: $0.85 per scheduled hour. This usage-based pricing directly tied revenue to the customer's most valuable metric—the hours Percolata helped them optimize. By 2017, Percolata had 18.4 million hours under contract annually, translating to $15.6M+ in potential revenue (though not all was delivered immediately due to pilot contracts). The value proposition was undeniable: retailers reported 10-30% revenue lifts by running A/B tests with "twin stores" (high-correlation locations where one ran Percolata schedules and one didn't). With 20-40x ROI, nearly all profit (since retail has fixed labor costs), inbound demand exploded. What didn't work: the sensor data subscription model, complex onboarding requirements, and the five-year slog to PMF. Tanaka notes that rebuilding the product for scalability became the new bottleneck—demand had outpaced delivery capacity.

By 2018, Percolata was working with 40 retail chains across the US, Europe, and Asia, representing stores ranging from 10 to 500 locations. They'd hired 22 people and raised $9.5M. Revenue hadn't yet crossed $10M, but was poised to do so. The company had shifted from a grinding sales-driven business to one driven by referrals and inbound leads. Tanaka reflects on the journey with humility: his biggest lesson was underestimating how hard it is to be a CEO and how much he'd misjudged previous leaders as an engineer. The hard things—capital, product rebuilds, team scaling, and perseverance through PMF desert years—taught him that entrepreneurship looks easier from the outside than it actually is.