Cross-skilled Workforce Allocation

Context & Aim

Boston Scientific in Clonmel is experiencing a time of transition towards a leaner data management system. Manual transformation of available data into visual reports with business insights has now been replaced by automatically generated data visualisation dashboards. Having appreciated the potentials and benefits of analytics from a product perspective, managers and practitioners are interested in expanding the use of analytics to different business units and process areas. In particular, production managers are interested in exploiting the data collected to optimise operators’ coordination policies in their labour-intensive production environment.  


In Boston Scientific Clonmel, operators are cross-trained following a plan that does not necessarily respond to strategic staffing requirements. Training is often used to occupy idle operators with obvious consequences on utilisation and costs.


Once trained, operators are not always given the time to build their proficiency at the new station and end up back working at a previous station with obvious consequences on utilisation and costs.


Ideally, staffing plans should guarantee a mix of proficiency levels and versatility so that productivity inefficiencies caused by absenteeism, planned holidays or transfers of operators to other lines are minimised. However, this is not managed in a strategic way. Rotation by operators happens on a ‘must’ basis rather than a planned ‘proficiency’ basis. Likewise, operators’ skills certifications should be maintained through ad hoc rotation policies but there is currently very little visibility around this which leads to significant certifications losses.


The application developed for Boston Scientific focusses on the allocation of cross-skilled operators to workstations in a labour-intensive assembly line characterised by capacity constraints and quality-related rotation rules. Considering the complex nature of the system operations and the presence of the human factor, which introduces considerable control challenges, the allocation approach has been designed to aid line supervisors and operators in the allocation decision process without subverting current allocation practices. Optimal allocation plans are generated using a mathematical programming approach. A graphical user interface has been developed to control input parameters and visualise allocation results.


The output allocation plan suggests how operators should distribute their working hours across different workstations so that throughput targets are achieved (or the gaps to these targets is minimised) and skills certifications are maintained while rotation rules are satisfied. The operationalisation of the allocation plans is left to the line supervisor/operators and is based on self-management practices that are currently in use on the production floor.


The allocation application can be used to make optimal allocation decisions and assess production plans feasibility based on the workforce available. A trial implementation of the application has already shown that significant potential benefits can derive from its daily use. A 70% reduction in expired certifications has been observed within three months of the tool usage, which obviously translates into significant training cost savings. The allocation application has also been used to assess the ability to satisfy a surge in products demand; production capacity issues have been highlighted and training requirements have been identified.

  • Ability to assess production plan feasibility based on available workforce versatility
  • Identification of production bottlenecks
  • Better visibility into operators’ rotation requirements
  • Reduction in training certification losses
  • Identification of training requirements
  • Possibility of expanding the allocation application to other production areas and, potentially, to other production sites in the Boston Scientific network

Many thanks to IMR personnel for their co-operation and knowledge in developing a Resource Optimization tool. Shortly after implementation, we had an increased customer request. By using the tool, there was a clear understanding by management of the issues faced by production as they did not have enough resources cross trained. Decision was made to reduce the plan for one month to allow for necessary training to take place. There was no impact to the customer. There has been a reduction of 70% in expired certs for last three months. We plan to continue this partnership to further enhance the use of the tool.”


Fergus Quilligan
Director of Analytics