This is a significant example of a popular theory, based supposedly on empirical evidence, which was found to have nothing to support it when looked at from another angle.
Now there has been much insistence in certain quarters that advertising on the internet cannot build brands in the same way as TV. Would it be a huge surprise if this turned out to be another phlogiston theory?
Throughout most of this generation of media expertise, some very persistent theories have reigned about the roles that specific media have performed. During my time involved in media planning, I have done my best to insist that almost any media can perform any role if it is used in the right way.
The right kind of advertising in a glossy magazine can add glamour to a brand's image. But so can the right kind of advertising on a poster, or even a direct-mail shot.
TV can build a brand, it can drive people to a cut-price sale (which can, if exploited too much, have the opposite effect on the brand long term), and it can allow direct selling in a very short-term, accountable manner.
So why would the internet not be as flexible? I have now seen conclusive proof that the internet does indeed have a positive and identifiable brand effect, but that - as I have suspected for some time - it is largely not used for this purpose because it is too expensive as a display brand medium, compared to other options.
As the internet is taking such a large proportion of overall advertising budgets, this may be of no interest whatsoever to media owners selling display online. However, it is certainly interesting to planners and buyers when considering the roles of each medium for a campaign and as the opening for a negotiation.
We must not hold on to old theories about how media work, no matter how reassuring. There is significant potential in adding value for clients by trying different ways of achieving the old objectives.
Sue Unerman is chief strategy officer at MediaCom