Cost estimation and tendering optimization through data science for a multinational construction company

Abstract

BESIX, a large multinational contractor, is one of the few companies that is able to build the most extraordinary skyscrapers, hotels, maritime works and other infrastructure works in the world. The examples are countless and range from the Burj Khalifa (i.e., the tallest building in the world) and the Tour Triangle (i.e., the newest state-of-the-art building in Paris) to the Princess Elisabeth Base on Antarctica. Also, Legoland Dubai, Ferrari World Dubai, the Warner Bros Theme park and two of the new stadia for the 2022 Qatar World Cup are constructed by BESIX. Even the very own new campus building of the Vlerick Business School in Brussels is constructed under BESIX’ guidance. Unfortunately, due to a cloudy economic environment with rising material prices and the nature of the industry (i.e., low margin, high volume), the financial performance of the construction industry is lagging. As such, construction companies must find new and alternative ways to help them make a better selection of projects and a better cost estimation. Hence, they will be able to continue constructing such mesmerizing construction works all over the world. The construction industry has arrived late to the digitalization revolution. This means that lots of untouched potential in data-driven decision-making and Machine Learning based optimization have yet to be discovered. Luckily, driven by its mission to create sustainable solutions, BESIX is already quite developed in its data management. The next step is to use this lead and start using data analytics to increase its profitability. The goal of this In-Company Project is to help BESIX choose better tenders and make a better cost estimation by using data science and exploiting the Master Data Management. The project is structured around three business use cases that help keep focus and attain the objectives of increasing BESIX’ profitability and efficiency. The first business use case calculates a probability of winning a tender by using historical tenders and their win rate. By only having seven input variables, the algorithm can generate a probability of winning a new tender with an accuracy of 82%. Even more, when the algorithm is used in the right context, it can improve the accuracy of the decision by approximately 20,59%. A few recommendations follow up on this algorithm and eventually it proves the advantages of a data science approach to the goal. The second business use case makes use of the probability of winning a tender in combination with an order book correction, a strategic profitability correction and a profitability margin estimate to pinpoint the interestingness of a new tender in a ranking of historical tenders. This grasps a part of the intuition and experience of the commercial department. It protects BESIX from losing its know-how due to retirements and resignations. Finally, the third business use case is advantageous after a positive tender decision is made. It tries to optimize the cost estimation for the bid. The conceptual solution is to find patterns in historic project data and to forecast a cost for specific packages in new tenders. This business use case will need to be developed in a strong closed loop, where there is feedback from the tender department, which will enable a better quality of the algorithm. The final chapter, a roadmap for implementation, describes the different steps of how to implement the data science solutions. Furthermore, the second and third business use cases were discussed with several data science companies. Consequentially, in the final chapter, a market study is made of the different partners that would fit BESIX to develop the final two business use cases. By implementing these suggestions, BESIX will be ready for the future of construction, of which, however challenging, data science will be a part. That will enable BESIX to continue astonishing the world with its construction works and to excel in creating sustainable solutions for a better world.