Are developers really the ones making this decision in most organizations? For us, it's not just developers, but anyone involved in the software development lifecycle for these websites, including separate QA teams that perfor testing and have to factor in the staffing models in relation to value of more extensive browser support. Another consideration is the type of website and audience using it. If its used by the general public and is a primary source of business revenue or support, loss of market share may be a valid concern if browser support is limited & may easily justify the need to spend more. With B2B sites the user base is not 100% controlled but may be more controlled and the value of supporting a larger variety of browsers may be difficult to justify, although usage stats help with evalutating this as well.