Product support is a key aspect in the marketing of high-technology products, since it strongly influences customer satisfaction and can also be an important source of revenue. Typical forms of support include operator training, equipment maintenance and, if necessary, repair - all of these are normally provided by manufacturers' support organizations. Good support is particularly important in some markets; an example is medical equipment where good operator training and quick repairs are essential because products are used in critical situations. Despite its importance, support has not been extensively researched. This study describes a management investigation of two aspects. Several authors have identified that product support is dependent on product design. Consequently, the same authors emphasize that support should be thoroughly evaluated during product design. This study identifies the range of factors that may be evaluated and shows that most of the companies surveyed do not fully evaluate support during the design stage. These results are not covered by previously published material and have implications for management. As support influences customer satisfaction, it is important to know how customers perceive support. The study investigated the customer attributes of good support, using interviews with medical equipment customers. The results show that a common set of attributes are associated with support, some relating to the product itself and some to the support organization. The characteristics of products which are easier to support were also identified from the interviews. The contribution of the research is that it made an exploratory investigation of the concept product support. It not only gave the first survey data on how companies plan support but also investigated customers' perceptions of product support. Consequently the study provides a foundation from which there is real scope for further management research, into what is becoming recognized as a vital element of high-technology marketing.