Before the advent of cloud computing, when a system was being designed, the team would have to estimate the performance requirements and then procure hardware resources to match those estimations, an expensive and slow process often undertaken with limited data. Not only is this a bad pricing decision, but it also means that unused capacity is sitting idle in data centers. Furthermore, performing these estimations could be impossible if the idea is to design a completely new business channel. Therefore, starting out with the smallest investment possible and aiming for exponential growth is a potential path to success for a company trying to innovate using the cloud. Buying or leasing hardware from a longstanding vendor is not cloud computing. Cloud computing is being able to provision resources on demand and remove those resources when they're no longer required.