What happens when a corporate office of a big Italian airport specialized on predictive demand analytics, decides to find a different method to approach its day-to-day work? In order to answer this question, we had the possibility to apply the Dynamic Risk Management (DRM), an innovative AFS methodology to forecast and assess risks that we have developed with the help of academia specialists.
To get a grasp of what we are talking about, it is fundamental to understand the importance of the passengers’ curve (PAX curve). For an airport, the PAX curve drives the business most importantly as the business uses it to forecast passenger flow, on which the business bases its pricing system – the principal revenue driver in the industry – as well as the estimation of secondary revenues – i.e. parking, shops, etc.
A few years ago, the officials at an Italian airport asked for help. Specifically, they wanted a new method to improve how they estimated passenger numbers. The traditional approach needed updating. Their goal was more accurate revenue forecasts. That would help reduce gaps between budgets and actual results. For my company, this posed a new challenge. We were tasked with advancing the traditional methodology since they wanted us to develop a more sophisticated replacement. One that was able to predict, with more accuracy, the outcomes of future years.
We tested different models in order to select the one with the lowest statistical error. We compared advanced quantitative models – both econometric and machine learning – able to make statistical analysis on the possible evolution of the PAX curve, to select the best one. At the end, we combined econometric models with an innovative model of deep learning.
Differently to the traditional Enterprise Risk Management (ERM) models, we needed a different approach. The traditional models are simple, mainly qualitative and focused on past results. We needed a predictive model that could identify possible risks that may affect the PAX curve in the future.
We reviewed academic literature for help. This let us identify an ERM 4.0 model as well as recognize significant future risks. The model improves the EBITDA estimate by about 20%. We tested the model and compared the results to the traditional method. We also compared to real data from an Italian airports group. The new model showed an improvement of the results by on average 2% monthly compared to the old method.
We reached our goal.
Our software simplified the implementation of this complicated methodology by employing Artificial Intelligence in a tailor made way, using the IA-RN model. We moved forward by connecting the quantitative model, strongly based on risks, with the financial and budget planning. In order to obtain this result, we chose EBITDA as the driver KPI, observing and calculating the effects of risks and the PAX curve on the indicator.
The software, differently from the traditional ERM system, provides real time data and risks levels.
Overall, the main advantages of this new methodology are:
- Improving the EBITDA forecast
- Direct link between risks and financial planning
- Direct link between risks and investments (fundamental part in the pricing system of the airport sector)
- Quantitative forecast of risks with Sensitivity Tests
- Optimization of the risks portfolio and insurance policies with efficient frontiers.
The sofisticated quantitative methodology we implement has been proved efficient in more than one occasion. It is a tailor made, easy efficient tool that reduces the statistical error and simplifies the day-to-day work of the ones that use predictive demand analytics.
At AFS, we are passionate about fostering innovation and empowering ambitious minds to flourish. Our mission is to provide best-in-class financial services for traditional and crypto deals, exploit European grants, and use quantitative methods to improve clients’ performance. We aim to help our customers unlock their full business potential.
Are you looking to fund your next venture or unlock grant opportunities?
Contact us at [email protected]!