Although the general concepts are the same, the functionality of self-checkout lanes seems to vary widely from chain to chain; at one large grocery chain where they have been in use for eight or nine years they seem to generally work well and the stores frequently have just one employee for eight (frequently in use) lanes. Many people seem to prefer these to a cashier. At another, smaller grocery chain the self-checkout lanes are virtually unusable and demand employee support after almost every item, resulting in no advantage for the store and frustrated customers. In this case people wait in line for several minutes for a cashier even when a self-scan lane is open.
What is the difference between sucessful and unsucessful self-scan lanes? I think it comes down to an interaction between two factors: the useability of the techonology and people's willingness to give up the human interaction of a cashier.
Some people are eager to give up dealing with a cashier. They find the interaction time-consuming and irritating, and would rather just get out of the store. Others (myself included) prefer that human interaction. The former group will tend to self-scan even when the technology is slightly flawed, and even when self-scanning slows them down because of bagging, while the latter will self-scan only if the technology is good and they percieve another advantage, such as speed for a small purchase.