Predict Leads

Rox Seever and his co-founders were running a previous startup that sold to sports clubs when they noticed something powerful: personalizing outreach messages dramatically improved conversion rates. As engineers, they became obsessed with automating and scaling this insight. They built machine learning systems to extract business intelligence signals—partnerships, hiring patterns, new clients—from billions of data points across the web. Predict Leads was born not from a grand vision, but from solving a real problem they encountered themselves.

Launched in 2015, the founders faced a brutal reality: building a big data company requires enormous amounts of training data and machine learning expertise. It took them a full year just to get their first paying customer. During that gap, they survived through a combination of EU innovation grants and cash flow from their previous lifestyle business (which generated $100-200K annually for the two of them). This runway allowed them to obsess over data quality without the pressure to sell prematurely. "It looked very dark at the end of 2017," Rox recalls, but their persistence paid off—by early 2018, their error rates dropped significantly and their accuracy improved dramatically.

Their go-to-market was intentionally lean: cold emails and warm introductions from mentors they met at the Startup Bootcamp accelerator in Amsterdam (which gave them €15K for 8% equity). The turning point came when they stopped trying to perfect the product for skeptical customers and instead focused on delivering exceptional results to those willing to try them. Once the data quality improved, customer satisfaction soared, and they moved from struggling to close deals to having customers naturally "decide to take the next step and integrate with us."

The founding team learned that their initial strategy—testing locally in Slovenia before expanding globally—was a mistake. Rox's advice to his younger self: "Go big from the beginning." They also discovered their customers didn't want leads; they wanted raw company intelligence data comparable to clearbit or FullContact, but differentiated by advanced signals (partnerships from news crawls, hiring intent from 12+ million websites, product launches). This pivot from lead generation to intelligence data was crucial. Their customer acquisition cost remained low (cold email + mentors) because the product's value for VCs and corporate strategy teams was self-evident once the data quality improved.

By mid-2018, Predict Leads had grown to 30 customers paying an average of $12K per year, putting them at roughly $300-350K ARR (up from $40-50K just a year prior). With zero customer churn on their core API business and a lean team of six in Slovenia, they were profitably bootstrapped and had no plans to sell. When Nathan asked if they'd accept 2X revenue in cash (roughly $600K), Rox firmly declined: "We just spent so many hours. The potential we have is longer. We're not playing this for the short term." His goal was ambitious but measured—doubling ARR by year-end while keeping existing customers delighted.