Software product-line engineering is a method for improving the efficiency and effectiveness of software development. Introducing such a method into an industrial software development environment is potentially of great benefit, but one cannot afford to stop product development while doing so. Rather, in Avaya we apply an incremental adoption strategy and therefore must identify which part(s) of the product line we will create first. Since we consider a product line to consist of a number of domains, the problem is to identify the right domains to start with. But what does "right" mean? How can we decide if a domain is "right" and worth spending the effort? We identified two driving factors for selecting product-line domains: corporate impact and likelihood of success. Our assessment of candidate domains is driven by these two goals, which we decompose further into a set of domain selection criteria and corresponding questions. The data, i.e., the answers to the questions, are gathered during interview sessions with our domain experts and evaluated according to our goal decomposition formulas. We illustrate the approach by an example application for which we assessed 20 different domains for one of Avaya?s major product lines.