The observation about the buying habits of a particular individual can be integrated with other behaviour patterns, which are found out through questions like:

• Does the customer purchase only when an item is available at a certain price?
• Is the customer more likely to purchase on a weekend or even during the week?
• Does it need to be organic material?

To put it in the simplest terms, if someone purchases only when something is ‘on sale’, but never buys anything in say December, the e-mail sale flier might not be dispatched to that person in December.

An algorithm would weigh various such behaviour traits to decide the likelihood that someone will pay heed to the e-mail message, and once accepted and opened, will decide to click through to the web site and actually buy the product. The more precise and elaborate the data, the better it gets at predicting. Through such data analysis, online marketers may find that 10 per cent of their population might be accounting for 60 per cent of bargain sales. So on the eve of the sale, they can send a full-price ad to everyone else.