Goldcorp published its mine data and let the world find its gold

In 2000, Goldcorp's Red Lake mine in Ontario produced just 50,000 ounces a year and the company was struggling; its geologists believed far more gold lay deeper, but drilling to locate it was slow and costly. In mining, secrecy was sacred.

Open the vault: publish the mine's 400MB of proprietary geological data and pay strangers to find what the company's own geologists couldn't.

CEO Rob McEwen, a former Merrill Lynch financier haunted by a 1999 MIT talk on Linux, posted 400 megabytes of exploration data (some of it from 1948) on the internet and offered $575,000 in prizes — $105,000 for the best entry — to anyone who could say where to drill. Over 1,400 submissions arrived from 50 countries; an Australian duo, Fractal Graphics and Taylor Wall & Associates, won with a 3-D model built entirely from the public data.

The winning map surfaced 110 drill targets, half of them previously unknown to Goldcorp and 80% holding substantial gold; deposits worth over $6 billion were identified, two to three years were shaved off exploration, and Goldcorp went from struggling to the industry's most profitable miner.

In 2000 Goldcorp, a Vancouver gold miner, was in trouble: its Red Lake mine in Ontario produced only 50,000 ounces a year, exploration had stalled, and every traditional avenue — blind drilling, more geologists, more secrecy — was expensive. Founder and CEO Rob McEwen had heard at a 1999 MIT conference how Linus Torvalds built Linux with strangers over the internet, and wondered why his industry could not do the same.

So McEwen did what no miner had dared: he put 400 megabytes of Red Lake's geological data, some of it dating back to 1948, on the company website and offered $575,000 in prizes — $105,000 for the best entry — to anyone who could pinpoint where the next gold was hiding. More than 1,400 scientists, engineers, geologists, students and even military officers from 50 countries took part.

Fractal Graphics and Taylor Wall & Associates, an Australian team, combined the public data into a 3-D model of the deposit and won. Their map surfaced 110 drill targets, half of which Goldcorp's own geologists had never considered and 80% of which held substantial gold. The new deposits were worth over $6 billion and cut two to three years off the exploration timetable.

The challenge turned a struggling mid-tier producer into the most profitable gold miner in the business. Goldcorp kept growing for nearly two decades, until January 2019, when Newmont — the world's largest gold miner — bought it for about $10 billion US.

The crowd had no dogma: outsiders looked at 400MB of data with fresh eyes — mathematicians, students and soldiers — and found what the company's own geologists could not.

Prize money bought results the payroll could not: $575,000 targeted one hard question, mobilizing 1,400+ solvers from 50 countries without any of the overhead of hiring them.

Openness removed the fear of being wrong: submitting a hypothesis cost a solver nothing, so wild ideas were tried; the best ones were not filtered by internal politics.

The publicity became part of the payoff: a struggling miner that admits its geologists are stuck generates attention no ad campaign could buy.

When insiders are blind, the answer may be outside: a $575,000 prize for open data returned deposits worth over $6 billion — a ratio of ~10,000x the industry had never seen.

Red Lake was transformed: the mapped sites were drilled with an 80% hit rate, Goldcorp's production and profitability climbed to the top of the industry, and the challenge became a template for prize-based open innovation — a decade and a half later, Integra Gold ran its own $1-million CAD 'Gold Rush' contest on the same model. Goldcorp kept growing until January 2019, when Newmont — the world's largest gold miner — acquired it for about $10 billion US.