Skip down to main content

e-Commerce as a Micro-Economic Microscope

Date & Time:
17:00:00 - 18:30:00,
Thursday 10 June, 2004


The fundamental goal of e-Commerce is to conduct business using digital electronic communications as the medium for sales and marketing. For retail businesses, this reduces to the problem of how to sell things on the web. In the large, this problem has two basic sub-problems, that of getting customers to visit a website and that of getting customers who do visit to buy products.

The first problem is analogous to what is called marketing in a traditional retail business and the second is analogous to sales. The sales aspect of e-Commerce, in particular, is susceptible to optimization and is a focus of one aspect of my research. Another and perhaps more significant aspect of e-Commerce, however, is that it provides the micro-economic equivalent of a particle accelerator which can provide new insights into the smallest units of economic activity.

By providing an idealized environment for experimentation, e-Commerce websites can provide information on how individual consumers make purchase decisions. Much as the interaction of sub-atomic particles informs us successively about quantum mechanics, chemistry and ultimately biology, understanding how people make micro-economic decisions teaches lessons about the entire range of economic activity.

Of particular significance is that the traditional micro-economic assumptions of behavior do not appear to be particularly accurate. Instead, much simpler basic laws of behaviour appear to be at work. My talk will cover the evidence for this behaviour, provide a sound framework for the examination of questions like this and describe some of the consequences of the observations so far. At the end, I will offer sweeping conclusions that are not yet supported by evidence, but which provide good fodder for discussion.

Data Dump to delete


  • Name: Professor Ted Dunning
  • Affiliation: Consultant
  • Role:
  • URL:
  • Bio: Ted Dunning has researched method for pattern discovery in diverse systems over the last 20 years first in an academic setting at the Computing Research Laboratory at New Mexico State University and later in emerging companies in southern California including HNC Software (now Fair Isaac Company), MusicMatch Software and ID Analytics. His work has been concerned with finding common threads in the analysis of symbolic sequences in language, genetic sequences, human web-browsing behaviour, musical preferences, purchasing behaviour and financial transactions. His early work on log-likelihood ratio tests for multi-nomial distributions is widely cited in the computational linguistics literature and he has authored or contributed to a number of patents. Most recently, he is an independent consultant focusing on data-mining and e-Commerce optimization.