Advertising management principles are derived mostly from logic and very little from empirical generalizations

Empirical generalizations

Empirical generalizations (EGs) are descriptive ‘if, then’ statements about the observed correlation between two or more marketing concepts (variables). The two most common forms of EG are shown in Figure 1 (reproduced from my 2001 article). One well known empirical generalization in marketing taking the first form, ‘If A, then B’, which I used in the Rossiter and Percy textbooks (1987: 51–4, 1997: 39–40) but dropped in the Rossiter and Bellman (2005) textbook for a reason given shortly, is: ‘The order of entry of brands is proportional to their ultimate market shares’. This observation was offered, apparently independently, in 1976 by the Boston Consulting Group (a correlation of r =.5 in industrial markets) and also in 1976 by the Hendry Corporation (a correlation of r =.43 in consumer markets). The correlation involved here is expressed in the form of a ‘size ratio’ of the market share that the next brand entering can hope to attain as a ratio of the preceding brand’s market share.

OPEN IN VIEWER

Figure 1.

The forms of empirical generalizations (observed ‘if, then’ relationships between two or more marketing concepts)

A second well known EG, also taking the first form, which was used in my second textbook (Rossiter and Percy, 1997: 75–7) but not the third (Rossiter and Bellman, 2005) for reasons discussed later, is Ehrenberg’s ‘double jeopardy’ generalization, which, stated in positive form, is: ‘the larger the brand’s penetration (the number of people who buy the brand), the larger will be its repeat rate (the more often they will buy it)’. This EG may be extended to include a third variable which is: ‘and thus the larger will be its total sales’ (although this last relationship is logical, not empirical, because sales are the product of penetration × repeat rate). This extended EG takes the second form in Figure 1, that is, ‘If J and K, then L’ (if high penetration and high repeat, then higher sales).

Strategic principles

Strategic principles (SPs), in contrast, are prescriptive (as opposed to descriptive) statements of the causal (as opposed to correlational) relationship between a marketing situation (the first marketing concept), a managerial action (the second marketing concept), and an outcome (the third marketing concept). The basic form of strategic principles is shown in Figure 2 (again from my 2001 article). The prescription is ‘if, do’ and the full form of an SP is: ‘If the market situation is X, then do Y, because this will produce the best outcome Z’. An example from my textbooks on advertising is: ‘If product choice is low-risk (X) and brands are chosen mainly at the point-of-purchase (Y), then set as an advertising objective visual brand-pack or logo recognition (which is the correct outcome measure, Z, for this situation)’. This SP is entirely based on logic. No empirical findings are necessary to derive this principle.

OPEN IN VIEWER

Figure 2.

The form of strategic principles (conditional prescriptive ‘if, do’ statements of the best course of action to take in a given situation)

Converting EGs into SPs

To convert an empirical finding or generalization to a strategic principle, you need to figure out the ‘why’ (the cause or causes of the finding). In the case of the first EG above, concerning the correlation between order of entry and market share, the two causes are the ‘marketing experience curve’ (in which the firms earliest to market gain experience earlier in how to successfully sell their product) and ‘pioneering advantage’ (a complex theory in which the first brand to enter defines the ‘prototype’ of the product so that later brands are seen as successively inferior imitators, with fewer and fewer consumers willing to try them because they have diminishing demand capacity having already bought an earlier brand). The first cause, that of marketing experience, would lead to a very different principle than the obvious order-of-entry principle of ‘If you want a large market share, then enter the market early’, which is of no use to the managers of brands who are considering entering late!

A useful SP can be derived from the order-of-entry EG by knowing just the first cause, marketing experience. This SP is for companies that are not the first to enter to ‘buy in’ marketing experience by raiding marketing executives from the pioneering firm (trade paper reports show that PepsiCo, for instance, has raided Coca-Cola’s marketing executives from time to time). An alternative SP for the number-two brand for overcoming the pioneer’s advantage can be derived from a semi-EG, ‘semi’ because it applies only to consumer packed goods, or fast-moving consumer goods (FMCG) products, proposed in a study by Urban et al. (1986). This semi-EG implies that the second brand, in order to surpass the leading brand, must have 36% better (perceived) product quality, or spend 3.4 times the amount on advertising that the leading brand spends, or spend the same but with 3.4 times more effective ads, or some combination that adds up to this total ‘impact’ (see Rossiter and Percy, 1997: 51). Presumably, later brands would have to be much better – or spend much more – than this! On reflection, I thought this empirically derived SP – true as it no doubt is – much too discouraging for managers and accordingly omitted it from my 2005 textbook.

The empirical generalization about ‘double jeopardy’, which was proposed as the ‘most important EG about buyer behavior’ in the Marketing Science special issue article by Uncles et al. (1995), is similarly helpless without a ‘why’ explanation. The most likely explanation of why wider penetration leads to a higher repeat rate is the brand’s retail distribution (see Farris et al., 1989; Fader and Schmittlein, 1993). The more popular brands are available in more retail outlets where people shop whereas smaller brands are not as widely available and thus, many times, if people intend to buy them they cannot, so the brand’s repeat rate suffers. Only by knowing the ‘why’ can the double jeopardy EG be converted into the following SP: ‘If yours is a repeat-purchase brand, then get it distributed as widely as is practically possible’.

1. Buyer behavior objectives and target audience selection

2. Communication objectives to achieve the buyer behavior goals

There are no empirical generalizations that are in any way useful for setting communication objectives for the advertising campaign. Rather, the communication objectives have to be set by logic and introspection, that is, by thinking about what mental responses need to be put into the potential buyer’s mind and then reinforced. Rossiter and Percy (1987, and in their subsequent books) logically identified the best set of communication effects, which have been picked up by many other textbooks (I know this from ‘permission to reproduce’ requests). The necessary communication effects are: category need; brand awareness (two distinct types, brand recognition and brand recall); brand attitude; brand purchase intention; and purchase facilitation. Two of these communication effects – brand awareness and brand attitude – always become communication objectives. Category need becomes an objective only when targeting new category users, as does purchase facilitation. Brand purchase intention becomes a communication objective when the purchase decision is high risk or when using an advertised sales promotion. These SPs, each of them an ‘if, do’ recommendation, are worked out logically, not empirically.

I wish to draw attention to an important logical distinction between the two types of brand awareness, which is still overlooked by academics in forming empirical generalizations (e.g. by Wind and Sharp, 2009, in selecting their 21 ‘laws’ of advertising) and by practitioners (e.g. in the ASSESSOR^TM model). This distinction results in two logical SPs. The first SP is the one exemplified earlier: ‘If the brand choice is made at the point-of-purchase, then the brand awareness communication objective should be visual brand recognition’. The visual brand recognition concept is often wrongly called ‘aided awareness’ (see e.g. Laurent et al.’s article in the 1995 special issue of Marketing Science) and wrongly measured as recognition of the brand name from a verbal list rather than from visual packages or logos as encountered at the point of purchase. The counterpart SP, also logically derived, is: ‘If the brand choice or consideration set of brands is decided prior to the point of purchase, then the brand awareness objective should be category-cued verbal brand-name recall’. The latter concept is often loosely called ‘unaided awareness’ (e.g. by Laurent et al., 1995). Brand recall is measured correctly by providing the product category as the cue, but some researchers wrongly use the advertising as the cue. Again, there are no EGs that can give the manager these essential SPs.

3. Creative strategy to achieve the communication objectives

Creative strategy is the step of advertising management that has the most conspicuous (and unforgivable) lack of EGs. For example not one of the advertising-related empirical generalizations in the 1995 special issue of Marketing Science dealt with creative strategy. Two of the 21 ‘laws’ identified by Wind and Sharp (2009: 249) touched on creative strategy but neither of these was a general finding. One ‘law’, which is to always use a ‘unique selling proposition’, is contradicted by the contingency theory in the Rossiter–Percy grid (see Rossiter et al., 1991). The other ‘law’, which pertains to the repeat showing of the brand visually, obviously can apply only to TV commercials and not to ads in other media. Nor will a search of the academic literature on creative strategy (including the dedicated advertising journals such as the Journal of Advertising and the Journal of Advertising Research) and the practitioner literature (notably the American Marketing Association’s [AMA] Marketing Research journal and the UK journal, Admap) be fruitful.

4. Sales promotion strategy to achieve communication objectives

Sales promotions also must be designed to achieve communication objectives because this is the only means by which sales promotions can cause purchase of the product – contrary to the claim in many textbooks that sales promotions cause purchase directly. In particular, sales promotions increase the brand purchase intention communication effect by first creating a temporarily more favorable brand attitude (see Howard’s [1977] theory of routinized response behavior, also Rossiter, 1987). The logically derived strategic principle here is: ‘If you want to increase brand purchase intention, then design a sales promotion that increases brand attitude’. Promotion offers that increase brand attitude are called consumer franchise-building (CFB) promotions.

Most of the EGs in the 1995 special issue of Marketing Science relating to sales promotion are of no help for deriving strategic principles. Consider, for example, Ehrenberg’s (1995) EG that price elasticity is a constant (namely, –1.5). This EG, like all of Ehrenberg’s EGs, is due to relentless averaging that ignores any deviations. Indeed, the constant price elasticity EG is contradicted in the same volume of Marketing Science by the semi-EG proposed by Blattberg et al. (1995), which is that higher-share brands are less (price-deal) elastic.

Moreover, all the empirical studies of price elasticity have overlooked Moran’s (1978) important conceptual distinction between ‘upside’ price elasticity and ‘downside’ price elasticity (see Rossiter and Percy, 1987: 332–4, 1997: 28–9, 45–6, 513–15, and Rossiter and Bellman, 2005: 19–20, 273, 355). This distinction forms the basis for two SPs. One is: ‘If you want to increase the brand’s upside price elasticity, then advertise the overall value delivered by its benefits’. The other is: ‘If you want to reduce its downside elasticity, then advertise the uniqueness of its benefits’. Wind and Sharp’s (2009) readers may note that Moran’s two joint principles make it contingent, not general, to use a ‘unique selling proposition’.

Moran also proposed an entirely logic-based strategic principle which he called the ‘ratchet effect’, and which is important for the timing of the brand’s sales promotions. This SP is: ‘Price promotions should immediately follow bursts of advertising in which the brand’s perceived value (benefit delivery) is “racheted up”’. In this manner, even a relatively small price cut will be perceived as good value and will lift sales more sharply than it would otherwise. Other advertising textbook writers do not do their homework and neither, of course, do most marketing managers, so they are not familiar with Moran’s brilliant ideas for advertising and sales promotion strategy. I have used Moran’s dual price-elasticity theory in numerous consulting applications – for instance, for Cussons Pty Ltd in the UK some years ago when they first started price-cutting their flagship brand, Imperial Leather soap – and it was always I who introduced Moran’s ideas to the managers. The ‘finance guys’ at Cussons persisted, by the way, and in my opinion fatally cheapened the brand.

In my textbooks, there are, however, some structural tactics for the execution of in-ad and point-of-purchase (p-o-p) sales promotions that do rely on empirical findings. As with the structural tactics for ads, these only make sense to apply, however, after you’ve worked out the sales promotion strategy for the brand from logical principles first.

5. Media strategy to achieve the communication objectives

The next task in advertising management is to formulate the media strategy (for the ads and for the advertised sales promotions). Two of the best known empirical generalizations about advertising relate to media strategy. These are Lodish et al.’s (1995; see also Lodish and Mehla, 2007) generalization about the short- and long-term effects of TV advertising ‘weight increases’ and Leone’s (1995) related generalization about the duration of advertising ‘carryover’. The negative version of Lodish et al.’s EG is that if a ‘weight’ increase doesn’t work in its first year then it will it not suddenly start to work in the next year or two after that. It is hard to believe that any manager would expect the risen-from-the-dead phenomenon, and certainly by a year later the manager would have stopped using the ineffective campaign anyway. The positive version of Lodish et al.’s EG is that ‘weight’ increases that do work in the first year have a carryover effect on sales in the following two years that is equal to the sales effect of the advertising in the first year.

However, this EG seems to be contradicted by the EG proposed by Leone (1995) which is that the average carryover duration of advertising on sales is only six to nine months. This, I believe, exposes weaknesses in the marketing concepts that are used to measure the data that go into these empirical generalizations. In the first place, as pointed out by Rossiter and Percy (1987 and subsequent) and Rossiter and Danaher (1998), the concept of ‘weight’ and the operational measure of weight as ‘gross rating points (GRPs)’ are inadequate for media strategy (and media planning) because they confound the concept of reach and the concept of average frequency. For example Lodish et al. (1995) found that the 2-year sales carryover was mainly due to an increase in the average buyer’s repeat rate of buying the brand, rather than to penetration, that is, to an increase in the number of buyers. If so, this would dictate the strategic principle for media strategy of emphasizing frequency (to increase the repeat rate) rather than reach; it would not simply mean increasing the ‘weight’. Leone’s (1995) EG, like that of Lodish et al., left unanswered the two key questions of ‘what’ carries over (communication effects) and what has to carry over (which ones) in order to cause the subsequent sales. In other words, Leone failed to provide the ‘why’ for the EG.

There is a media-related debate that has been going on for years (e.g. most noticeably in the JAR) about whether it is necessary for the brand’s advertising to be recalled – that is, to ‘carry over’ – for the advertising to influence sales. This in turn depends on how advertising recall is conceptualized (and measured). For example Millward Brown’s brand-prompted claimed ad recall measure shows a positive relationship with sales, whereas the original Burke Marketing Research day-after advertising recall measure, which is program-prompted and then ad-prompted and requires proof of recall, does not. In any event, the debate is superseded by theory (see Rossiter and Percy, 1987, and subsequent). It can be logically worked out that if brand attitude maintenance is largely dependent on the brand’s advertising (transformational advertising) then brand-cued advertising recall will be a cause of attitude and thus sales. In contrast, if the advertising mainly imparts information about the brand’s functional benefits (informational advertising, which includes price-promotion advertising) then recall of the ad will not be necessary since the information is already absorbed in the increased brand attitude communication effect.

6. Budget determination