Digitalization of energy trading

In Europe, the amount of energy from renewable sources as a proportion of the total energy supply will increase by a further 30% over the next five years. Forecasts for the future weather-dependent infeed of electricity from wind power and photovoltaic systems very much determine pricing on the energy trading market: Prices are becoming more volatile, and extreme price scenarios are becoming more frequent. The volumes traded on the spot market shortly before the delivery date are continuing to grow; purchases and sales are becoming ever-faster; and large trading organizations are transacting up to 30,000 deals a day.

These developments mean that energy trading companies can only be successful in the future if they automate their processes. The majority of trading organizations already use partially or fully automated algorithms today. New technologies in the areas of Big Data and data analytics can also unlock great potential, which is why energy trading organizations are developing the methods they need as well as introducing new processes. There is a need for standard processes to be fully automated and for the organization to be further developed into an “Energy Trading Factory”.

The Energy Trading Digitalization Framework developed by Horváth is the framework for an overarching digitalization strategy. It summarizes the individual starting points and development areas in the transformation to becoming an Energy Trading Factory

• Sources: The traditional systems for energy data management (EDM) and energy trading and risk management (ETRM) as well as the exchange interfaces will continue to exist in the future. However, in order to save costs and benefit from economies of scale, these systems will be operated in the cloud. New processes are aligned with the standard of the software in use, meaning that no expensive customization is necessary.
   
• Data: The data in the various source systems are merged into a data layer. Advanced database technologies allow quick access to large amounts of data. Traditional data warehouse solutions are supplemented or replaced by data lakes in the cloud. These guarantee easy set-up of the infrastructure and faster scaling. Robotic Process Automation (RPA) is used as a bridge technology for interfaces that would be too costly to implement.
   
• AI models: Artificial intelligence offers new ways of evaluating large quantities of unstructured and complex data. It recognizes patterns, and uses this to support decision-making or to automate it entirely. Data science and machine learning platforms also make access to this technology possible for those users who do not have sufficient programming knowledge. Data scientists – introduced as an entirely new role in many companies – use the more complicated machine learning methods for analysis.
   
• Applications: At application level, there are trading tools as well as reporting and analytics tools. The interface to short-term trading is often already automated today, by means of simple algorithms. Based on the large volumes of data and data analytics models, in future not only pure trading execution but also the creation or adaptation of trading models will be automated. In addition to standardized reports for the back office, the “Reporting and Analytics” area also offers employees comprehensive business intelligence functions in real time. They enable drill-down data evaluations and also analyses of different scenarios, for example in financial planning.
   
• IT & infrastructure: With increasing automation and digitalization, a functioning IT infrastructure becomes more important than ever. As such, probabilities of failure must be determined and necessary redundancies defined. In addition, the topic of access management is increasingly gaining in focus.
   
• Process & organization: Digitalization and automation also place new demands on processes and the organization’s functional model. Automating previously manual tasks enables the use of existing capacity for more value-adding activities. Areas of activity change, disappear or are newly created; the tasks of some functionaries therefore need to be re-described, and new roles, such as data scientist, or new tasks in the area of business intelligence emerge. Governance processes must also be readjusted: Responsibility for the automated trading systems must be assigned to individuals, risk strategies must be adapted, and access rights and responsibilities for the data must be defined.