Reduced Order Models (ROMs) are a useful tool to generate lightweight and portable versions of simulation models. They operate by taking otherwise complex and computationally expensive simulation models and represent the same behaviour in more compact and simple equations or algorithms. This allows it to run at a fraction of the time and computational cost, making it much more versatile. With the proliferation of digital twins in industry, ROMs have the ability to bolster the digital twin of operation by leveraging existing simulation data. Simulation data can be considered an investment in the digital twin, as it allows the models to provide more benefits by not only informing the design of a piece of equipment but also the operation of that equipment in the field. ROMs are an integral part of creating what can be termed the executable digital twin, an approach which sees the ROMs or systems of ROMs taking live operational data from equipment in the field (power plants, factories, mobile stations etc…) and using it to make predictions on equipment behaviour. These executable digital twins can exist as remote applications allowing operators to adjust settings based on predicted future conditions. The different methods for generating ROMs were examined, including the various data sources which can be considered, types of data which are useful for the intended application and the type of ROM in question. The complexity of the data and its suitability for generating an effective ROM is an important consideration. Here, examples of both suitable and unsuitable data were identified. Additionally, the methods for deploying these ROMs - either standalone or as a system of ROMs - in the executable digital twins were examined along with examples of their use in industry. The different technologies available to complement and extend the capabilities of these digital twins were examined and illustrated, their benefits and drawbacks were outlined. Finally, conclusions were drawn as to the different profiles and use cases which can most benefit from executable digital twins. These conclusions are made in terms of the conditions laid out in this analysis from the perspective of data, application, and peripheral technology.