Why we can’t give up on the ‘science’ of advertising


There may not be immutable laws governing how marketing and advertising work, but that’s no excuse for failing to take a scientific approach to progressing our knowledge of them.

Advertising has, historically, employed few scientists, but in spite of this it has also been an industry that has often used theories and ideas that have the appearance of science as part of its armoury of tools to impress clients and sell more ads.

Today, every planner worth their salary is expected to have a point of view on Byron Sharp’s How Brands Grow; in my day it was Rosser Reeves’ Reality in Advertising. However, unlike science, which builds on and refines theories over time, advertising has a tendency to consign old theories to the dustbin. Who talks about Reeves these days?

Advertising has a troubled relationship with its own past, and in particular its past conceptions of itself. One of the better accounts of this is Paul Feldwick’s, in The Anatomy of Humbug. “Advertising people,” he says, “have created implicit narratives that give them permission not to think about the past.”

Building on theoretical knowledge

Anyone who has done any serious academic research into the history of advertising will notice that a body of work written in one era rarely makes reference to work done in prior eras. As a result, the idea of progress in our understanding of advertising and how it operates is problematic.

The question of progress isn’t addressed head-on in Feldwick’s book, but he seems sceptical about its prospects. Thinking about the idea of a science of advertising, he says: “The major contribution of science to our understanding of advertising since the late 20th century has been…to demonstrate that any dreams we might have had of reducing advertising to a set of rules or psychological principles that could be successfully applied by trained technicians, were illusory.”

He cites the philosopher of science Stephen Toulmin’s work in support of the claim that over the last 100 years, “the notion of what science itself is has been radically transformed”, from a Cartesian quest for certainty to something more pragmatic. “The more we learn about…the brain, the mind, the human sciences, chaos theory, quantum physics – the less certainty we can pretend to have.”

A lack of certainty is thus presented by way of an excuse for advertising’s apparent failure to improve its understanding of itself. This seems unreasonable. You would be hard pressed to find a serious scientist who would agree that the impossibility of achieving certainty in their findings was an obstacle to increasing their understanding of their field. An absence of certainty is, in itself, no excuse for a lack of progress.

The value of the scientific method

At the heart of the scientific worldview, on which progress in a field of enquiry is predicated, is the scientific method. While there are disagreements about the exact formulation of the scientific method, amongst the preconditions of its application are such principles as: measuring the ‘right’ things, being dispassionate about what you find when you do measure (that is, reporting failure as well as success) and sharing results with the wider community for evaluation and elaboration. These principles are at best patchily applied in advertising.

A dispassionate application of the principles of the scientific method in advertising is a risky business: the incentives in the industry are stacked against it. Consider, for example, the case of campaign evaluation. Outcome measures are often determined by what is convenient to measure (awareness), rather than by what should be measured (contribution to sales).

Then, most clients don’t want to be told that the money have spent on a major campaign has been wasted, even if they learn something along the way. And campaign performance data is seldom shared outside the client and the agency; an exception to this is data submitted to effectiveness awards, but of course this exhibits strong survivorship bias almost by definition.

A lack of certainty is presented as an excuse for advertising’s apparent failure to improve its understanding of itself. This seems unreasonable.

The truth is that being successful in the advertising industry is not dependent on developing a better understanding of how advertising works. People go into advertising in order to make ‘great ads’ for their clients, and all the exciting and engaging activities that entails, not to roll back the frontiers of knowledge about the practice.

Of course, there are academics that make it their job to study advertising, its methods and its effects, but as Feldwick observes, advertising practitioners and advertising academics have relatively little to do with one another. This is as much the fault of the academics as of the industry: after all, they don’t have the same problem with incentives that practitioners have.

The obstacles to progress in our understanding of how advertising works don’t stem from its complexity: this is not an impediment to progress in biology or climate science, for example. Duncan Watts and others have demonstrated the possibility of progress in sociology and the sciences of human interaction, given access to appropriate data and the analytical tools with which to wrangle them.

What holds scientific progress back

Advertising’s failure to improve its understanding of itself is, rather, a consequence of its cultural and economic formation, and the individual needs and incentives that result from it: the incentive to demonstrate successful results to keep the client happy, the need for employees to believe they are engaged in meaningful, well-founded work.

That last need is worth dwelling on. Feldwick has a lot of time for Ehrenberg-Bass, Byron Sharp and their idea of advertising as salience (or ‘fame’, to use the sexier term), saying he thinks “it accounts for an awful lot of what advertising does”. The trouble with this theory, though, is “that it doesn’t make anyone look particularly clever – not the client, certainly not the planners, and not even the creative department – so nobody usually wants to [use] it”.

The Ehrenberg-Bass account of the role of advertising makes extensive use of behavioural data across many markets and categories, with findings published over the years in peer-reviewed journals. As such, it has a better claim to be truly scientific than many of its predecessors, but it doesn’t sit easily with agency culture.

In recent years, management consultancies have started to encroach on the territory of advertising. Consultancies give much more weight to quantitative methods and computational modelling, and to the idea that the activities of marketing can be improved through experimentation and a better understanding of how things work.

Unencumbered by advertising culture, and absent the ingrained incentives of the industry, consultancies can advocate for the application of a more scientific approach to the evaluation and design of advertising and marketing activity. This is a bold move, and one that requires buy-in from the highest levels of the client business: it requires the acknowledgement that some things will work better than others, and that some things might fail.

This, however, is the price of increasing understanding, the price of progress. If clients value progress enough to pay that price, and agencies are unwilling to change, the advertising industry as we know it is at risk.

Nigel Shardlow is director of planning at SandTable.

Loes Mills