Beyond the horizon of any sample lies the shadow of the population from which it is drawn. In reality any sample findings immediately beg a number of key questions:

» Is this sample representative of the population e.g. are the characteristics of banks indicated in the sample shared by all UK banks? If not, how is the sample different?

» How can the population be sized on the basis of the sample e.g. how many racks, how much space, how much power does the market require?