This paper examines the question of whether geostatistics really holds much promise for geotechnical engineering, and attempts to identify conditions that if satisfied could convert geostatistics into a real design aid for geotechnical engineers. The paper briefly looks at the origins of geostatistics, and examines its successful application to a variety of fields. It also describes the potential benefits geostatistics can bring to geotechnical engineering, and explores possible reasons underlying geotechnical engineering's difficulties in fully harnessing the power of the method.