It is by now familiar that in the deployment of solar photovoltaic (PV) systems, the cost dynamics of major system component like solar cell/module is subjected to experience curve effects driven by production learning and research and development at the supplier side. What is less clear, however, is the economics of system integration or system deployment that takes place locally close to the user, involving other market players, in the downstream solar PV value chain. Experts have agreed that suppliers of solar PV system must customize their flexible characteristics to address local unique users' and applications requirements and compete on price/performance basis. A lack of understanding of the drivers of the economics of system customization therefore is a deficiency in our understanding of the overall economics of this renewable energy technology option. We studied the non-module BOS cost for grid-connected small PV system using the experience curve framework. Preliminary analysis of PV statistics of the US from IEA seems to suggest that learning in one application type is taking place with respect to the cumulative installation among all types of grid-connected small PV projects. The effectiveness of this learning is also improving over time. A novel aspect is the interpretation of such experience curve effect or learning pattern. We draw upon the notion of product platform in the industrial management literature and consider different types of local small-scale grid-tied PV customization projects as adapting a standard platform to different idiosyncratic and local application requirements. Economics of system customization, which is user-oriented, involves then a refined notion of inter-projeets learning, rather than volume-driven learning by doing. We formalized such inter-projects learning as a dynamic economy of scope, which can potentially be leveraged to manage the local and downstream aspect of PV deployment. This dynamic economy may serve as a focus of energy policy having implications on standardization of design and training for installation, facilitating knowledge reuse among different integration projects and enabling inter-projects learning. (C) 2007 Elsevier Ltd. All rights reserved.