30 Jan A look into the digital future of dry bulk commodities
In the world of commodity trading, data can be your biggest ally. It is what commodity traders and analysts use to make informed decisions, identify patterns, and predict market trends. But as the industry evolves, so does the amount and complexity of data available. Traders and analysts are struggling to keep up with the vast volume of information, let alone make sense of it. In this interview, DBX’s CEO, Alexandre Claude, delves into DBX’s transformative technology and shares his vision for the digital future of the dry bulk commodities market.
Q: How did the idea of DBX come about?
I traded commodities with a focus on agriculture, coal and dry freight for nearly 15 years. I began my career at Louis Dreyfus, a global agricultural merchant and processor, then worked for Vatenfall, a European energy company, before becoming head of a trading desk at Freepoint Commodities.
My experience in the industry allowed me to have a firm understanding of the structure of the commodity trading industry and locate inefficiencies within the existing trading model. I became increasingly interested in the role that technology could play in providing advanced and real-time market intelligence to commodity traders. In particular, I began to explore how remote sensing could be used to provide insights into inventory estimates using satellite imagery and data and how vessel tracking could lead to real-time supply and demand analytics of commodities.
I thought long and hard about what new product could be brought to the commodity industry whilst leveraging the latest technologies, and how to bring the best partners to flawlessly execute the idea. Three years ago, I met with Florian Thaler who had recently founded OilX with the help of The Signal Group. I listened to his experience and was impressed by what OilX had achieved so far with crude oil. I explained my product vision for dry bulk commodities and offered a partnership. DBX was born!
Q: How does DBX cater for different market players?
DBX’s solutions are designed to be used by commodity producers, power utilities, commodity trading houses, banks, and hedge funds. If we step back a moment and think about the structure of the commodity trading industry, I would break this down to two axes:
a. Trading activity: Financial Trading (trades which are cleared on the exchange) and Physical Trading (bilateral OTC trades with logistics involved).
b. Trading model: Asset Light proprietary trading (where there is no help from other businesses within the organisation) and Asset Heavy backed trading (where there is help in the form of trade flows from the organisation such as physical procurement for a steel mill or hedging client requirement for a bank).
(Note: Groups are not mutually exclusive and borders can be blurred, especially for large trading houses that can show different profile types).
Topology of the commodity trading industry
Physical commodity traders can deftly gather intelligence from their network of contacts to fill gaps in the market by crossing bids and offers. They effectively use a series of arbitrage strategies to lock in profits in the physical market. To implement these arbitrage strategies, the traders need to source, blend and deliver the commodities from the producers to the consumers.
Financial traders have different styles: some trade by instinct, some do technical analysis by using charts, and others follow fundamentals by analysing the supply and demand balance of a commodity. The most sophisticated traders follow a hybrid “quantamental” approach: they apply Artificial Intelligence (AI) and Machine Learning (ML) to real-time analytics and combine them with a qualitative expertise. On the one hand, data science removes biases and provides objectivity. On the other hand, fundamental analysis gives context and real-life expertise.
In both cases, access to superior market intelligence lays the foundation for success and superior returns. DBX provides a complete digital twin of the dry bulk commodities supply chain – from upstream production (e,g., a mine) all the way to downstream consumption (e.g., a steel mill) – thereby enabling organisations to make sense of the data and lead to sharper decisions.
Q: Can you explain the technology behind DBX and what specific problems does it address in the commodities market?
Commodity data is unstructured, opaque, and current data sources can lag by weeks or months with large mining companies often publishing quarterly production reports to update the market on the performance of their operations. The amount of data available is growing and unsorted, though increasingly pertinent to decision making. Its current opacity is due to the difficulty of sourcing, aggregating, and cleaning of this data. DBX addresses these difficulties by compiling the data-sets into a form that can be used by traders and analysts, saving time and energy for both to perform other higher value tasks.
As well as this, DBX is able to use AI/ML algorithms to process satellite data imagery to estimate stockpiles in key locations for which data is otherwise absent or unpublished. This allows us to not only provide scientifically based stockpile estimates for well-known sites that complement on-the-ground survey estimates, but also to shed light on sites where surveys are not authorised and quantitative data is not publicly disclosed. This has been observed in mines, non-public ports, steel mills and power stations across the world.
Pleiades VHR PAN image (~ 16 km^2) of Qingdao’s iron ore stockpiles.
Satellite image classification based on a machine learning model.
DBX tracks iron ore exports from the major producers to give a real-time view about their performance. We benchmark our data against official reports and post a ～99% correlation as shown in the chart below.
DBX Producer Iron Ore Tracker, Rio Tinto.
We augment our iron ore tracking technology with nowcasts to estimate how much iron miners will produce over the year and whether they will be in line with their Production Guidance. We have made some bold predictions which have turned out to be accurate. For instance, in May 2022, DBX estimated that Rio Tinto and Vale would revise their Production Guidance lower. On 19 July 2022, Vale cut its Production Guidance, and on 15 October 2022, Rio Tinto followed suit by also cutting their Production Guidance.
Q: How does DBX differentiate itself from existing solutions in the market?
We first iterate with our clients and draw from our domain expertise to clearly define the use cases we want to cover with the DBX platform. We then work our way back to the product and draw on technology to build a best-in-class product.
DBX’s value proposition, a unique multilayered product centred around domain expertise.
A few of our key differentiators include:
• A near real-time Iron Ore supply & demand model for World and China
• Granular data on ports and iron ore / coal stockpiles at hubs in ARA, India, China, UK, USA and much more to come
• Superior flows coverage with split between met/thermal, fines, pellets, lumps, brands, receivers and producers
• Minimum 95% accuracy according to validation from top mining companies and trading houses against their own flows and customs
• Our Platform is updated daily, with nowcasting for the current week and month
• Customisable platform with proprietary data from the client
Q: How do you see DBX’s technology impacting the future of the commodities market?
Commodity companies are increasingly recognising the importance of technology in driving innovation and growth, and embrace its transformative impact. ‘Data is the new oil’ has been a popular expression of late. The expression also strongly resonates with DBX’s mission statement: to shed light on the opacity of the commodity markets by providing the most complete view of the supply chain to empower sharper decision making.
The recent economic and political turmoil have led to high prices, extreme volatility and reshuffling of trade flow for commodities. This created a boom for many commodity players who managed to predict price changes and tap on physical trading arbitrage opportunities. However, the markets are not expected to remain kind to commodity producers, trading houses and hedge funds forever. As the environment becomes more challenging and profitability returns to normal levels, technology will prove to be a key differentiating factor for commodity companies.
DBX processes one million data points every day to provide valuable analytics to traders, analysts, and data scientists. Although we own the algorithms and make the results available via an API/UI, our users ultimately decide how to execute on them.
Below, we will use India to illustrate how some users are already one step ahead of the market. DBX’s technology shows real-time and expected discharge operations in India. For every single cargo, we show the grade, the quantity, the shipper, the receiver and the transaction price.
DBX’s Indian dataset.
This is a unique dataset that opens up various possibilities to:
• Benchmark procurement activities in terms of price and quality
• Get information with regards to Indian trade import margin
• Anticipate where coal will be discharged, who will receive it and what would be the brand
• Demand forecasting and new products
• Business development opportunities
As author William Gibson once wrote, “the future is already here. It’s just not evenly distributed”.
For more information on DBX, visit www.drybulkx.com. To request a demo of the DBX Platform, click here.