Electricity generated by distributed energy resources (DER) located close to end-use loads has the potential to meet consumer requirements more efficiently than the existing centralized grid. Installation of DER allows consumers to circumvent the costs associated with transmission congestion and other non-energy costs of electricity delivery and potentially to take advantage of market opportunities to purchase energy when attractive. On-site thermal power generation is typically less efficient than central station generation, but by avoiding non-fuel costs of grid power and utilizing combined heat and power (CHP) applications, i.e., recovering heat from small-scale on-site generation to displace fuel purchases, then DER can become attractive to a strictly cost-minimizing consumer. In previous efforts, the decisions facing typical commercial consumers have been addressed using a mixed-integer linear programme, the DER Customer Adoption Model (DER-CAM). Given the site's energy loads, utility tariff structure, and information (both technical and financial) on candidate DER technologies, DER-CAM minimizes the overall energy cost for a test year by selecting the units to install and determining their hourly operating schedules. In this paper, the capabilities of DER-CAM are enhanced by the inclusion of the option to store recovered low-grade heat. By being able to keep an inventory of heat for use in subsequent periods, sites are able to lower costs even further by reducing off-peak generation and relying on storage. This and other effects of storages are demonstrated by analysis of five typical commercial buildings in San Francisco, California, and an estimate of the cost per unit capacity of heat storage is calculated.