Expanding your locus of control with precision
Challenge
All business owners understand that a company's growth and performance are not just dependent on internal factors. For some organizations however, the role of external factors is much larger and of course difficult or impossible to control. Our client Gerry's' Dnata, one of the biggest air cargo companies of the world, is one such example where the external environment has a huge role on day-to-day business performance. These factors could range from minor daily changes such as weather forecast, cargo weight, dwell time to major ones such as political scenario and unforeseen circumstances in other international airports.
We were faced with the challenging task of developing a product based on advanced statistical and machine learning models that help the company calculate the impact of each factor and analyse risks and calculate revenue accordingly.
Solution
The first thing that comes to a financial controller's mind when asked about sensitivity analysis is Microsoft Excel. However, sensitivity analysis performed through conventional methods in programs like MS Excel generates results that are limited and static, with very limited options, not to mention the tedious process which requires significant amount of effort on data cleaning and data engineering.
We developed a tool that can be used to carry out advanced sensitivity analysis and what-if analysis techniques for the company. The solution allows you complete flexibility w.r.t. statistical variable distributions, variable selections, number of simulations, and various sensitivity analysis techniques. The solution also forecasts future revenues of the company using advanced machine learning techniques and simulation methodologies. It is dynamic in the sense that it requires a database view to be developed only once. The product will then update and provide updated sensitivity analysis on a click. LFD develops the entire ETL process which enables the product to directly fetch latest data from the system.
Result
The users can enter their own thought-of scenarios into the product and evaluate possible outcomes. This decreases the time taken to create financial analysis to almost zero. This also enables quicker reporting for the management teams, and it can be used to evaluate scenarios on the go. The product is highly optimized, and the entire process happens within a few seconds. It will also aid in assessing the risk involved before making a huge policy shift or strategy change. Managers can also use sensitivity analysis to identify the components of a project, when changed even slightly, will most impact the outcome of a projects.
The product also relies on advanced machine learning and simulation techniques. For every prediction made, it develops more than 100,000 possible scenarios before providing an output. The user can also decide how many simulations s/he wants to run for the results. The mathematical model evolves as more data is entered into the system.
