Shipyard productivity evaluation with key performance indicators

Abstract

Shipbuilding is a unique industry dealing with complex and one-of-a-kind products with ever changing demand trends that force the supply limits. Small and medium sized shipyards are suffering to cope with the challenges to survive and sustain their activities. Due to the nature of this business, shipyards need to pay even more attention to their productivity and performance. This requires combining a deeper engineering knowledge with business goals. Recent development in the information technologies followed with automated processes provide means to use data to create knowledge and thereby a better understanding of the processes and to manage them. This thesis suggests a methodology to evaluate the productivity of a shipyard in all hierarchy levels with key performance indicators (KPI) which are derived with a data-driven approach. It demonstrates that a shipyard could develop its own KPIs from its own historical data and utilise them to evaluate its own performance, identify the shortcomings, increase engineering as well as business knowledge and make more accurate estimations before and after signing a shipbuilding contract. Neither data analysis, nor the KPIs are new to this industry. The value of data collection and analysing has been well appreciated for a long time. However, the use of data is mostly limited to the assessment of only certain business cases and lacks an overall perspective for combining strategic objectives to tactical and operational ones and vice a versa. The proposed methodology offers a step-by-step approach to identify these connections considering the major challenges of shipbuilding. It answers the questions of where to start, how to perform and how to interpret and make use of the analysis both with a top-bottom and bottom-up approaches. The methodology offers a deep analysis starting with the investigation of performance shortcomings through background from internal and external sources. This background information is then used for the identification of strategic, tactical, and operational objectives. The methodology explains how these background and objectives could be connected systematically to success factors and relevant measures. Finally, the methodology provides a detailed guide for discovering the KPIs by means of data collection, data organisation, building a data model, performing statistical analyses and how to select and create KPIs and where to use the results. Application of this methodology is shown in a case study performed with real shipyard data from Uljanik Shipyard / Pula – Crotia, as a part of a research project funded by the European Union, HOLIstic optimisation of SHIP design and operation for life cycle (HOLISHIP). By analysing the shipyard dynamics through background and combining these with the project goals, it was decided to create KPIs for the assessment of new structural ship design alternatives from building cost and producibility perspectives. Relevant success factors and measures were identified before starting data collection and analysis. These analyses were performed in a systematic manner which is proposed in the methodology and given in detail in chapter 5 of this thesis. Part of the data was extracted in a structured form from the computer aided ship design and planning software and material resource planning software of the shipyard. Other required data was collected through unstructured interviews, standalone spreadsheets and documents which were then organised in structured spreadsheets. Due to practical reasons like availability, low cost and familiarity of the shipyard employee, data organisation, modelling and analyses were performed by Microsoft Excel. Mainly six data sources were created based on design, production, and commercial data from twelve previously built Roll-on Roll-off Passenger Ships (RoPAX ships), each in spreadsheet format. Some of the spreadsheets having more than 20 columns and over 20.000 rows. These data-sets were then combined in a data model by use of primary keys that were available for all data-sets, namely yard project number, macro space and yard group. The primary keys have connection to each piece of information in all data sets and reflect the way of job handling and data collection of the shipyard. From the data model, several measures were created and tested by means of regression analyses to investigate the relationship between different parameters. Regression analyses were chosen to ease communication with the shipyard and project teams and performed both in single and multiple linear regression methods. These relations were checked upon discussions with the shipbuilding experts from different departments of the shipyard and demonstrated with visual graphs. The expert opinion already supports a technical relation between parameters and measures and thereby eliminating the doubts of causation shown through regression analyses. By this way, it was possible to create two complex KPIs to evaluate the new design alternatives, one for the building cost and the other for the producibility. Shipyard team was already familiar with the basic use of the selected tool. After a one-day training most of the participants were able to handle the data organisation, data model and also were able to create their own measures and make their own analyses. It was found that better relations could be modelled when more experts were involved with technical background from different departments like production and design. It should be noted that the application of the suggested methodology relies on two major prerequisites. First one is that the data must exist, otherwise the methodology could only guide the shipyard for defining the data which needs to be collected. Second one is the involvement and support of the executive management, otherwise application of the methodology would be another number crunching exercise which would not be used and improved for daily use. Furthermore, application of the methodology requires mass amount of time at the beginning with the involvement of shipyard experts and a dedicated core team to extract data and build up the data model. Also, the results of the study cannot be generalised, but rather could only be used by the shipyard in question. The main contribution of this thesis is to provide a guideline for Small and Medium Sized Shipyards which are in need for a systematic approach for digitalisation and answering questions such as where to start, why data collection is necessary, how to handle data, which data to be collected, what to do with the collected data, how to make analyses, and in which ways these analyses could be used. It was shown that the collected data could be easily turned into engineering knowledge to support the processes improvements as well as business decisions. Finally, the study paves the way for further analyses and studies for improving the productivity with a data-driven approach. Future studies could focus on more complex analyses on productivity of outfitting works and effects of qualitative factors with more advanced tools like machine learning algorithms that are connected to a live data base.