| September 12, 2007
I recently read a new paper
by Len Lodish and Carl Mela called "If brands are built over years, why are they managed over quarters?"
It brought to mind the quote attributed to Einstein, "Not everything that can be counted counts and not everything that counts can be counted." Except, in this case, I believe that much of what counts can
be counted–you just need to use the right tools to do so.
In their paper, Lodish and Mela argue that many brands are in trouble because management has taken a short-term approach, relying far too heavily on scanned sales data. Fifteen years ago, Lodish was an early advocate of the new technology, but, according to this write-up
on Knowledge@Wharton, at that time he had also warned that "while these tools could be an effective way to measure the impact of discounting … there are other long-term measures that may not be as easy to collect, but are just as important."
Well, it seems that marketers did not heed his warning, because in their paper, Lodish and Mela focus on the misuse of scanned sales data to assess the profitability of price promotions. Fairly early on, they state,
While immediate, large, observable short-term increases in sales arising from discounts are readily apparent, the long-term effects of discounts and the effects of other components of the marketing mix, including advertising, new products and distribution, are more difficult to measure.
They suggest that the short-term benefits of promotions are, in the long term, largely offset by three factors:
- Consumers learn to expect price discounts and buy only on deal, often stockpiling purchases, which brings sales forward in time but does not increase the total amount bought.
- The use of price as an incentive encourages consumers to focus on price as a differentiator rather than product quality or other brand attributes.
- Competitive response which increases the number of promotions but results in the same amount of sales being made at lower margins.
Implicit in their assessment is that baseline sales models can be misleading. Baseline sales models provide estimates
of the likely
level of sales in the absence of a discount or other short-term trade promotion. They are used to assess the short-term profitability of a promotion. But as Lodish and Mela explain, the more people buy on deal, the more profitable the promotion will appear, unless allowance is made for any consequent fall in baseline sales resulting from such activities. If marketers ignore the erosion of baseline sales and continue to fund supposedly "profitable" price promotions, they risk turning loyal brand buyers into price shoppers who may come to find store brands an even better deal.
Contrasting the short-term with the long-term, Lodish and Mela offer several studies that demonstrate the long-term effect of advertising, new products and distribution. These include a study conducted over the course of five years by Ataman, Van Heerde and Meta on 25 packaged goods brands in France. They conclude:
Findings from this research suggest that advertising and product play the greatest long-term role in increasing the baseline sales of a brand. Discounting activity had a smaller but negative long-term effect on baseline sales.
However, as this project demonstrates, you can only assess long-term effects over a long period of time. After taking a swipe at the short-term mentality of Wall Street, which, they suggest, often dictates a marketing strategy that works against long-term success, Lodish and Mela bemoan the paucity of longitudinal data collected by many of the firms that they visited. They conclude, "If brands are built over years or decades, it is hard to see how firms can attain any insights into brand building with just 52 weeks of data."
More than once, the paper refers to the "myopia" which results from focusing too closely on short-term effects. In my estimation, Lodish and Mela point up two basic forms of metrics myopia in their paper:
- Managing brands by the data you have, not the data you need
- Taking business decisions based only on short-term sales data
Their recommended panacea is to augment the short-term metrics with others designed to assess the long-term health of the brand, namely trends in baseline sales and price elasticity measured over months, quarters and years. If baselines are increasing and promotional sales are decreasing, they suggest, "then brand equity is improving." If, however, price elasticity is increasing and the baseline is deceasing then "there is a problem with the firms' brand building strategies."
I agree with the basic assessment made by Lodish and Mela in this paper. All too often, managers do get lured into making decisions based on the data that's available, not the data that is really called for. Then they justify those decisions as being based on "the numbers" rather than careful thought about what the numbers might really mean. But are Lodish and Mela really saying something that's new? Will their recommendation give managers all the tools they need to manage their brands? I think not.
At an Advertising Research Foundation workshop in 1997, I had this to say about the use of base sales identified from market mix modeling (a more accurate and sophisticated version of "baseline sales"):
These core sales represent around 60% of the total, varying from as low as 34% to a high of 79%. Of course, this is a very useful measure of brand equity in its own right, and the size and trend of this core component of sales can tell us something of the fundamental health of the brand.
My recommendation for curing metrics myopia included not just measures of long-term sales but also understanding of the attitudinal drivers of sales, for the following reason:
The key problem that I see with sales response modeling is that…the analysis treats all consumers as the same, and they are not. To scanners, a sale, is a sale, is a sale, but a sale to a loyal buyer must be worth more than a sale to a non-loyal buyer. So we need to understand what aspects of the brand lead people to adopt it as their own.
I recommended conducting brand equity research to understand the drivers of purchase consideration and to anticipate trends in brand health. Marketers do need to track the long-term sales equity of a brand, but they also need to measure the consumer mindset behind those sales.
In not recommending the integration of attitudinal data, Lodish and Mela seem to be suffering from their own metrics myopia. Throughout the paper, I was expecting them to suggest integrating survey metrics with modeled scanned sales data to provide a more holistic picture of brand health and to diagnose why the long-term trends are what they are. After all, most people buy brands for a reason – even if it is because they think it is the cheapest or they just like it more than others – and attitudinal data can provide a complementary, and sometimes anticipatory, picture of brand health. No such recommendation was forthcoming.
Reliance on scanned sales data has led to a tectonic shift from brand-building advertising to brand-destroying price promotions. If marketers and modelers had paid more attention to "soft" measures like brand consideration and brand attitudes instead of simply relying on the "hard" sales numbers, maybe we would not be in the situation we are today, with so many brands diluted and under threat from store brands.
What do you think? Please let us know and while you are at it, please rate the post. Feedback is a good thing even if it is sometimes disappointing.