I have experience building personas for new software products within the physical retail industry.
These personas were created using data collected from potential customers based on available online information that was filtered using data from our interactions at retail conventions. These new software products plan to define and break into a new market within the physical retail space. This required us to move forward using creative user research methods. It was unfeasible as a small start-up without a large client base to collect formal user data from traditional methods such as contextual inquiries, interviews, or surveys.
These personas were built to be iterative. They were intended to capture both design and marketing persona information and to give the development and sales team a concise, direct picture of our understanding of our consumer/user base with the limited knowledge available at the time. The goals for this persona were to be a starting point for understanding 1) the primary user and 2) how the company will adopt design practices into the development life-cycle as it scales from startup to small company dynamics.
I specifically chose to pare down the persona creation to a single consolidated individual to set the stage for splitting into multiple personas in the future. This was the best choice for our team at the time to avoid the pitfalls of 1) designing for the average of many users and 2) designing with personal bias. Because of the small company size, a formal design process was not yet established within the product development life-cycle. Given these considerations, it made sense to get everyone on board with a single user who could be referenced and included in development and sales discussions which were typically the source of product design decisions.
*note: xtensio was used to create the visuals for the design persona and role details.