With an increasing interest for extra heavy oil (EHO) extraction, a number of heavy oil fields will initially be produced using a cold production scheme, while at a later stage thermal production in the form of steam injection can be introduced to enhance the recovery. Such a fullfield conversion from cold to hot production has so far only been performed to a limited extent, hence the experience and understanding of the various challenges with such a task is limited. While much work has been carried out on both cold and hot production of extra heavy oil as stand alone processes, little has been published around the procedure of converting from cold to hot production. This conversion contains a number of pitfalls detrimental to a successful field development. In an EHO cold to hot conversion scheme, several choices will have to be made with respect to well positioning, completion options, timing and scale of the conversion. Investment decisions in surface equipment, handling and cost of water and gas for steam generation are also important considerations. All of these designs and decisions affect the success and profitability of the conversion scheme. One of the basic challenges with modeling cold to hot conversion is the difference in scale required. The thermal simulation of a hot production scheme requires a very fine grid to capture the fine scale processes, while for full field simulation one needs a much coarser grid to be able to obtain a manageable model both in size and simulation time. The high number of producers and injectors required on a fullfield scale adds to the complexity of the problem, making recovery, production profile and optimal number of wells highly uncertain. A simplified translation to fullfield conditions of these parameters from sector models or spreadsheets, can introduce erroneous results and conclusions. To assure the best field performance over its lifetime, these issues need to be considered already in the initial field design. Based on the work performed, several interesting observations have been made regarding how to optimize field performance.