by the TerraManta team (Techstars Class 111)
It’s difficult to believe that one year has passed since TerraManta joined Techstars Paris 2017 program.
We did not know what to expect prior to arriving in Paris. What we found was an incredible ecosystem of mentors, advisors, and investors powered by overwhelming desire to help Techstars companies succeed.
Before sharing highlights about TerraManta progress, I have to express countless thanks to the Techstars team – both local and extended, the city of Paris for being our home away from home, and our launch customer – an integrated oil and energy company based in France.
How long would it take for a complete outsider to become an expert in a deep domain?
That’s the question my co-founder, Andrei Klimanov and I, asked prior to starting TerraManta. For example, doctors require at least 10 years of education prior to practicing medicine with more years invested later in advancing their professional competency.
Crude oil is the biggest commodity market in the world. Companies which depend on crude oil and distillates face enormous economic risks. Airlines often hedge fuel prices by purchasing future contracts at a certain price. Crude oil prices go up? Airlines win. Crude oil prices go down? Airlines often have to close fuel hedges – at a price – to take advantage of lower oil prices. The economic risk ranges from hundreds of millions of dollars to billions of dollars annually for these companies.
So how long would it take for a complete outsider to become a crude oil expert capable of forecasting future oil prices? Probably at least 2 decades, while learning from other experts who have biases.
When crude oil prices collapsed in 2015 reaching $28 / barrel by mid January 2016, it became clear that crude oil experts were no longer equipped to analyze the impact of geo politics on crude oil market fundamentals (supply, demand, inventory, distribution, exports, imports) and resulting price actions.
Something was missing.
That “something” is a platform for domain specific analysis of unstructured and structured data with the goal to apply machine learning in deep domains where humans require many years to master and learn with inevitable biases.
The platform is based on a goal centric learning principle, just like human analysts do when they elect to learn something. They set a goal which aligns their knowledge acquisition approach. Acting just like a human crude oil analyst, TerraManta software instead will do the same.
The first goal – can future crude oil prices be forecasted more accurately – became the first product: TerraManta for Crude Oil. Our software analyzes about 4,000 energy relevant news from public data sources every day, obtains crude oil market fundamentals every month, and looks for correlation between topics influencing fundamentals and how changes in fundamentals cause price actions in the market. What human analysts do today, TerraManta does automatically on a mass scale while learning every day.
Today, crude oil prices more than doubled since January 2016. As a result, the interest in TerraManta is growing.
Several dialogs are in progress: US based airline, major bank in Canada, another major bank in France, leading company providing specialized risk analytics for global corporations, and another leading company providing energy analytics.
We are also excited to support a 17-user trial with another global oil company
Does TerraManta software work? We believe so. The software generated a price forecast on 25 September 2019 which was on target until 11 October 2018 (when this blog post was written).
The Techstars Paris program created many memories which will be forever remembered. To name a few ….
Friendships and relationships which transcend borders and continents. Exceptional coffee. Equally exceptional croissants. Grace under pressure redefined by program manager and associates.
Above all, ‘give first’ was transformed from a spoken comment into a real and sustainable personal belief.