An article I read in the early 2K's summed up several software engineering studies about the importance of people, process, and technology to software projects. The conclusion, overwhelmingly, was that successful projects had a cohesive team of very good developers. In less successful and completely failed projects, developers were not carefully chosen and time was not allocated to develop a sense of team. Software methodology and technology (equipment, tools, etc.) varied widely and was not deemed as critical to success.
The conclusion was that people were 60% of the success of a project while process and technology were 20% each.
I'll try to find a link to the article. It would have been around 2001 in either IEEE's Computer, CACM, Queue, or Dr. Dobb's.