5 Years of Obsession.
1 Score That Actually Works.
In 2020, we started building a stock score that thinks like a value investor. Today it's used by investors in 100+ countries, with 7 score categories analyzing 80+ metrics across 100,000+ stocks and a Valuation Model that customizes itself for every company. This is the story of how we got here.
The Origin (2020-2021)
Daniel Pronk runs a YouTube channel teaching fundamental analysis. Back in 2020, every video followed the same process: pull up a stock, run through balance sheet ratios, growth metrics, profitability, valuation. Every stock. Every time.
"I was running the same checklist for every stock. The same process, every single video. At some point I thought: what if this could be automated?"
Daniel Pronk, Co-Founder
Daniel had the investing framework, but turning it into software required engineering he didn't have. Jake Ruth, a software engineer who'd been watching Daniel's videos, saw the same inefficiency Daniel was feeling and thought: "I can build this." He started prototyping on his own, then began sending Daniel weekly progress updates showing what he'd built. Week after week, no response. Jake kept shipping anyway. Eventually Daniel saw enough progress to realize this wasn't just another viewer pitch. That collaboration, Daniel's investing depth plus Jake's engineering, became the foundation of Stock Unlock.
Building the First Score (2021)
Jake built the first version of Stock Scores in 2021, translating Daniel's analytical framework into working Python. From day one, they made choices that set the score apart from every competitor.
At the time, there weren't many stock scoring systems. One competitor's entire "score" was the percentage difference between a stock's current price and its all-time high, marketed as a margin of safety. Others were building momentum-based ratings, where stocks score higher simply because their price is going up. Daniel saw the fundamental problem: as a stock's price runs higher, it actually becomes less attractive. When it drops while fundamentals are improving, it's getting more attractive. Momentum scores have it backwards.
What made it different:
Pure fundamentals only
No momentum. No technical analysis. No price-based signals. A stock's quality is measured by the business itself: profitability, financial health, growth, and whether the price reflects the value.
Industry-specific algorithms
Banks are scored differently than tech companies. REITs use different metrics than retailers. One size doesn't fit all.
Hand-coded Python, not generic templates
Every scoring algorithm was written specifically for our use case. No off-the-shelf solutions.
The first version of the score used industry-level templates. Banks were scored differently from tech companies, which were scored differently from REITs. This was already better than what existed, where most competitors applied a single formula to every stock on the market.
Users immediately started asking for more: "Can I search by score? Can I see how a stock scored last year?" Those requests planted seeds for what would come next.
The Historical Challenge (2024)
Users kept asking: "What did Apple score in 2015? How did Tesla's score change over time?"
We thought historical scores would be straightforward. Calculate today's algorithm against past data. Done.
We were wrong.
The Discovery
Calculating scores historically was like putting the algorithm under a microscope. Dividend scores that looked stable in the present showed unexpected weekly swings across a decade. Companies that changed industries mid-history broke assumptions. Valuation ratios that worked for one era didn't translate to another. Multiply these edge cases by thousands of stocks and 35 years of data, and the scale of refinement needed became clear.
The Big Refactor
By this point, the team had grown. Karan, a software engineer who joined as CTO, had been building the backend infrastructure that made scoring possible at scale. He led the rebuild. Thousands of lines of Python became a config-driven database. Every metric definition, every threshold, every industry rule, all documented and versioned. It was months of painstaking work.
"We're talking nine to twelve months of code refinement, testing, iterating, and really making sure that this system can run without needing an engineer in the loop."
Karan, CTO
Running scores historically revealed edge cases we never would have caught in the present: companies that changed industries, stocks with wildly volatile valuation ratios, metrics that only made sense for certain time periods. Each discovery made the algorithm smarter.
What felt like a setback became the foundation for everything that followed. Clean data. Consistent logic. And a much smarter algorithm. But the historical work also surfaced a deeper question: if two stocks in the same industry could behave so differently over time, why were we still scoring them with the same metrics?
The Breakthrough
Even with industry-specific algorithms, something wasn't right. Daniel was still spending hours manually analyzing individual stocks (Costco, UnitedHealth, Meta) because the score alone wasn't enough.
Jake posed the challenge: "This couldn't scale." They needed the algorithm to understand each stock the way an experienced investor would.
The Solution: Per-Stock Customization
Daniel started recording Looms, walking through how he'd analyze each stock manually. What metrics mattered. Why certain ratios were more reliable for certain companies. Karan (our engineer) codified every insight into the algorithm.
Adapts to Volatility
Scores adjust based on how stable each stock's numbers are
Per-Stock Metrics
Finds the most predictive ratio for each company: FCF for Meta, OCF for Amazon
Fair Comparisons
Groups similar companies so a bank isn't scored like a tech startup
Growth-Aware Valuation
Fast growers are valued on growth-adjusted ratios; slow growers shift to yield-based metrics
The system was built from the start to evaluate each stock individually. Even within the same industry, two companies can have very different valuation profiles: Amazon's earnings are volatile but its operating cash flow is stable, Meta is best measured by free cash flow, and JPMorgan is best evaluated on P/E alone. Daniel spent months analyzing stocks one by one, recording which metrics were reliable for each and sending detailed Looms to Karan, who codified those patterns into an algorithm that filters out noisy or unpredictable signals automatically.
This took 9-12 months of iteration to get right. The algorithm now runs across tens of thousands of stocks, doing automatically what Daniel used to do by hand.
Where We Are Now
Today, Stock Scores analyzes 80+ fundamental metrics across over 100,000 stocks from 70+ global exchanges. The system has 7 equally weighted score categories, with the Valuation Model as the standout: it algorithmically identifies the most reliable valuation multiples for each stock, then evaluates whether it appears over or undervalued.
With the launch of Historical Stock Scores, you can now see how any stock scored at any point over up to 35 years of history, with point-in-time accuracy. Every historical score uses only data that was available at that moment, so you can backtest without hindsight bias.
What makes this system work isn't just the engineering or just the investing knowledge. A fully automated approach produces generic templates that every platform ends up offering. A purely manual approach can't scale beyond a handful of stocks. Stock Scores work because Daniel's investing intuition was systematically codified by Karan's engineering, creating something that neither could have built alone.
We're just getting started. But every decision since 2020 has been in service of one goal: a scoring system that respects the complexity of real businesses. Want to see the technical details? Read our methodology.
"The score is wild now. It's not a static score."
Daniel Pronk
"We're not afraid to take on huge technical challenges. And we're not afraid to rethink and be critical of past iterations."
Jake Ruth
See the score in action
Search for any stock and see how it scores across all 7 categories.
Common Questions About Our Story
How Stock Scores came to be
When did Stock Unlock start building scores?
In 2020. Our co-founder Daniel Pronk was spending hours on his YouTube channel analyzing stocks by hand and wanted to automate his process. The first coded version launched in 2021, and we've been refining the system ever since.
What was the biggest breakthrough?
Per-stock valuation model selection. Instead of applying the same ratios to every stock, our algorithm evaluates which valuation multiples are statistically reliable for each individual company. It took about a year of iteration between our co-founder's manual analysis and our CTO's engineering to get it right. You can read more about how it works.
How many stocks do you score today?
Over 100,000 stocks across 70+ global exchanges, using 80+ fundamental metrics across 7 score categories. With the launch of Historical Stock Scores, you can also see how approximately 40,000 of those stocks scored at any point over up to 35 years of history.
Who built Stock Scores?
Stock Unlock is a creator-led company. Our co-founder Daniel Pronk (260K+ YouTube subscribers) designed the scoring methodology based on his own investing research process. Our CTO Karan codified it into scalable algorithms. The entire team is obsessive about getting the data right.
Who is Stock Unlock?
Stock Unlock is a creator-led fintech company used by investors in 100+ countries. We connect to brokerages, track portfolios, and provide fundamental research tools covering 170,000+ stocks and ETFs across 70+ global exchanges. Stock Scores is one piece of a broader platform that includes a stock screener, DCF calculator, dividend analysis, and more. Read the full company story.
Still have questions?