Getting advertising wrong

Last week we saw Mark Petticrew and friends make fools of themselves claiming that the booze industry encourages pregnant women to drink. Today they're back with a study proving that they don't know the first thing about advertising. The study has not been covered by the media, which must have come as a disappointment, but you can read it here.

It is obvious from the first line that this is not a serious piece of research:

Alcohol advertising is an important influence on alcohol consumption, and the evidence supports the restriction of advertising and marketing as a cost-effective intervention for reducing alcohol harms.

Having begged the question, they trawl through case studies published by the Institute of Practitioners in Advertising, mostly from between 1981 and 2003, looking for evidence that alcohol advertising is aimed at increasing alcohol consumption.

In a way, of course, it is. The aim of a Stella Artois campaign is to sell more Stella Artois. The aim is not to increase beer sales per se, let alone alcohol sales in general.

It doesn't seem to matter how much evidence shows that advertising does not increase aggregate consumption, 'public health' people still insist that it does - because they think (wrongly) that it would justify a ban.

Overall, 30/39 (approximately 77%) of the case studies explicitly mentioned increasing or maintaining market share as an objective or use this as a measure of campaign effectiveness.

Yes, that's the whole point.

There is consistent evidence from these case studies to support the industry claim that growing or defending market share is an important advertising objective.

Aye, there is.

They then move onto the 'alcohol industry claim' that 'Advertising Does not, and Is not Intended to, Stimulate Consumption'. This is where they come unstuck.

If sales and volume sales can be taken as a proxy for consumption, then all of the 39 case studies present evidence that advertising affects this outcome.

Er, no. Each of the case studies is for a specific brand. A rise in market share for one brand does not imply a rise in overall consumption. It is much more likely to lead to a fall in market share for another brand. Indeed, the case studies offer a reminder of several brands that were once very familiar but are not anymore, eg. Boddington's, Hofmeister, Murphy's.

Here are the authors' first two examples of advertising being used (supposedly) to increase consumption...

“We did not want to just sell Magners in the warm summer months: we wanted it to be drunk throughout the year”.

(Magners Cider, 2008)

The same campaign also: “…amplified Magner’s relaxed timeless brand values and drove winter consumption".

But even Petticrew et al. seem able to grasp that Magner's was trying to get people to drink their cider in winter instead of drinking, say, lager, not in addition to drinking the same quantity of lager...

Arguably, these examples show attempts to increase market share, rather than overall consumption. 

No kidding. But just when you think they have seen the light...

Other examples are more likely to reflect increases in consumption, such as this example where Foster’s lager used a humour-based campaign. The evaluation describes how it

“…gave permission for a ‘drinking occasion’… Blokes were lapping up the ads, but most importantly, they were lapping up Foster’s lager too…as the ads ran, sales rose, and they have continued to do so”.

(Foster’s Lager, 2015)

This campaign did not aim to increase overall consumption, and nothing in the quote suggests otherwise.

As noted above, the fact that all campaigns had increasing sales as an objective and/or main outcome is also an indirect indicator of the intended effects on consumption.

As noted above, it is nothing of the sort. The authors seem unable to distinguish 'consumption of Brand X' with 'consumption of alcohol'.

...one common way of measuring the effectiveness of the campaigns is in terms of willingness to purchase. The Grolsch lager campaign provides one example of this:

“The most concrete measure of increased consumer demand is rate of purchase. Grolsch’s rate has more than trebled since campaign launch…Put simply, consumers are demanding the brand more than ever”.

(Grolsch, 2003)

It would be tedious to repeat all the examples cited by the authors. Suffice to say, they are all like this. First, the authors claim that the company is trying to increase overall demand for alcohol, then they produce a quote that clearly shows the company is only interested in increasing market share.

When they take on the 'industry claim' that 'Any Observed Relationship between Advertising and Consumption is not Causal', they say...

...these case studies often have the stated aim of demonstrating a causal relationship. This is sometimes made explicit, as in these examples from the Campari and Archers Schnapps case studies:

“This case history will show the vital role that an integrated advertising strategy, in terms of planning input, media choice and creative content, played in that growth. It will seek to demonstrate a causal relationship between advertising and Campari sales performance…”

(Campari, 1981)

No doubt there was a causal relationship between the Campari advertising campaign and Campari sales performance, but that has got nothing to do with the question of whether the advertising campaign increased the number of units of alcohol sold. The distinction is blindingly obvious. Are the authors really too thick to see it or are they playing dumb?

...it has been claimed that advertising benefits the consumer by keeping product prices down [47]. However, there is considerable evidence from these data that the opposite is the case, with advertising helping to maintain or increase prices.

Reference 47 is none other than the 2014 reprint of the IEA's Advertising in a Free Society, which I abridged and wrote the foreword to. It does indeed say that advertising tends to lower prices. Advertising is a major lever of competition and competition tends to lower prices. There is plenty of evidence for this (see page 8, for example).

But what is true in general terms is not necessarily true in every case. Premium brands are more expensive because they have a good reputation. That reputation is often built on advertising, as I explain in the foreword...

But what if it could be shown that the more expensive branded good is identical to the unbranded budget good by objective criteria? Would this not be proof that the company is playing on consumers’ vanity and anxiety to exploit them? Harris and Seldon argue that the question is meaningless be-cause a product has no objective value. Even if the branded good is distinguishable from the unbranded good only by the advertising that portrays it as being of higher class, the consumer is justified in buying it. If he values it more highly be-cause he associates it with glamour, good taste or ethical living then he is right to pay more for it. Harris and Seldon argue that ‘if a bath soap, a fountain-pen, or a carpet gives more pleasure when the consumer thinks it is used by a duchess or a television performer, then he is making a logical decision in buying it: he is being more sensible than his critics’

In the whole history of British marketing, one brand has used advertising to justify higher prices more blatantly than any other: Stella Artois. With the slogan 'reassuringly expensive', it turned a bog standard Belgian lager into a premium drink. It is therefore no surprise that Petticrew et al. use it as evidence...

Several case studies also describe the mechanisms by which higher prices are maintained, e.g.,

“The premium price ideally pays for the advertising which tells the consumer the product costs more in the first place”. (Stella Artois, 1993)

Nobody has ever claimed otherwise, but note the contradiction in their logic. If, as they claim, alcohol advertising leads to higher prices in general, they should be in favour of it because higher prices lead to less consumption. 

In reality, the aims of alcohol advertising vary with regards to price. Premium brands use advertising to justify higher prices while retail advertising tends to focus on lower prices. Mid-range brands simply do what most advertising does - remind people that they are still there.

In conclusion:

These data suggest that advertising both aims to change attitudes, expectancies and consumption behaviour, and is effective in doing so.

And so it is. That is the proper purpose of advertising. What is doesn't do is increase per capita alcohol consumption or rates of alcohol-related harm - and nothing in the study shows otherwise.