Media buying company Zenith Media's research director, Deb Perry, who conducted a three-month study of internet site visitation in relation to web advertising, said internet advertisers, publishers and ratings companies were misguided in their use of "big numbers" to measure the popularity of websites.
Perry cited recent statistics given by internet ratings company Media Metrix, which she said listed 500 websites that attracted more than 50,000 Australian visitors per month. According to Media Metrix, the survey sample included 2000 online panellists.
"What does that 50,000 people mean?" Perry asked. "They're just projecting their panel out to the population."
Media Metrix's vice president and director of research for Asia-Pacific, Jonathon Jephcott, admitted all market research, including that surrounding the internet, incurred "sampling error". Moreover, Jephcott agreed internet audiences were "volatile", thus the sampling error for their behaviour would be considerable.
However, he pointed out that the use of a small sample as a means by which to estimate the behaviour of a larger sample formed the basis of all market research, including that of television, which he said had been conclusively proven to be accurate.
"This suggests a misunderstanding of sample survey theory," Jephcott said. "She's (Perry) a distinguished researcher . . . she's being ungenerous."
Similarly, Vivien Feyen, marketing director for internet ratings company IMR Worldwide, rejected Perry's criticism as "not well informed". Feyen agreed that web ratings terms, the validity of which Perry challenged, such as clicks and hits were "vogue terms", but suggested internet market research methodology was "much more sophisticated than Zenith knows".
Perry compared the current standard of internet market research to that used in cinema and outdoor (billboard and public transport) advertising. She said companies spent relatively small percentages of their marketing budgets on cinema and outdoor advertising because accurate demographic figures regarding cinema audiences and exposure to outdoor sites were not available.