Based on a paper by Jean-Bernard Kazmierczak

For marketers, understanding the effects of advertising and trade promotion is like the search for the Holy Grail. This is where the greatest part of marketing expenditure takes place, the impact on both short term and long term brand performance can be critical and yet the way that advertising works, particularly in the long term, is still as much of a mystery as it ever was.

Until now, academic research and commercial services to address the issue have all concentrated on the short term effects of advertising, i.e. the increase in sales during a campaign and in the few weeks immediately following.

While these analyses provide valuable insights, the benefits of advertising are always perceived to be mainly long term, i.e. brand building, rather than simply short lived sales uplifts. Indeed, in the short term, advertising can be a costly and rather inefficient way of generating sales. What is needed is a way of quantifying the long-term effects of advertising and promotion. A methodology for doing this has now been developed at Information Resources.

Evaluating the long term effects of advertising - a new approach
The new approach is based on the premise that if advertising produces a strengthening effect on a brand, this will be reflected in increased core demand (everyday (base) sales) and reduced price sensitivity (sales will be less affected by price increases), which cannot be attributed to any other cause.

In order to isolate the changes that are attributable to advertising rather than any other influences (market trends, competitive activity, seasonality) these other effects must be quantified and removed. Any increase in core demand or decrease in price sensitivity can then be said to be due solely to the effect of advertising or promotion.

To do this requires highly detailed information. It is not enough to know how many products are sold nationally or even how many sold through one particular store group in a given period, these figures may hide an enormous variation in activities in individual stores. For example, if 5% of stores are running promotional displays over a month, this could mean a different group of stores each week or 10% of stores for half the time, etc. In order to be able to understand all the price reductions, special offers, etc. that occur, the best source of information is therefore weekly individual store data. This provides thousands of observations and allows the figures to be analysed very deeply both in terms of the number of variables and also of geography.

Methodology
The formula employed to measure the impact of advertising (or any marketing variable) on sales is based on a model that, in simple terms, looks like this:

Total sales = Function of: base sales; market trend and seasonality; price elasticity (including base price and temporary price reductions) and marketing activity (direct and competitive; promotions and advertising - short term).

As advertising is assumed to have an impact that continues after the actual campaign has finished, this is accounted for by a calculation known as an Ad-Stock. The Ad-Stock builds in a carry-over effect that diminishes with time. Sometimes two ad stocks are used to account for short-term effects (a few weeks) and longer-term effects (up to a year).

The new element that has just been developed is a factor that represents the long-term pressure of advertising and promotion. When this is applied to the calculations it shows whether there is any change in the core demand and/or the price elasticity.

Case Study
The example below is based on a highly promoted and advertised consumer goods product. Store-level weekly scanning data from a five-year period (1995-1999) were used, along with advertising inputs for an additional year (1994-1999). A short-term model was run over the five years, with a long-term model over the four last years only.

The results of the study of short-term effects were very typical: 67% of total sales were base sales, 28% were due to promotions and 5% were attributed to advertising (equally split between TV and other media). If we look at the ratio between incremental sales due to advertising and the corresponding investments (Figure 1) the results are also typical. The short-term efficiency of advertising is poor.

Fig. 1 - Short-term efficiency ($ incremental sales divided by $ investment in advertising) in three regions over 5 years.

The long-term part of the model can be reported according its two aspects: core demand and price elasticity.

Figure 2 shows the evolution of the core demand across time. In this case, we found that core demand increased with an increase in long-term advertising and also when long-term promotions were increased - not unexpectedly as the brand is a big leader in the category.

Fig. 2 - How core demand (line) varies across time, quarter by quarter. The decreasing demand at the very end corresponds to reduced long-term advertising pressure.

With respect to price elasticity, price sensitivity is reduced when advertising is increasing and promotions are decreasing. See example in Figure 3. This implies that, in the long term, advertising makes the product less price sensitive while promotions make it more sensitive.

Fig. 3 - The same product is becoming more sensitive to price when long term advertising is decreasing and promotions increasing.

Having looked at both the long term and the short-term effects, it was then possible to make a comparison of total "incremental" sales due to the full effect of advertising and also of promotions. For advertising, the total effects were much greater than what was observed in the short term; the increase factor was in the range of 2 to 4. In contrast, the total effect of promotions was slightly smaller than what was observed in the short term. The decrease of efficiency was about 20 percent.

Next steps
Further work is now in progress where promotions can be divided according to type and the long-term effects of each type examined. This will enable us to determine whether promotions such as displays and in-store leaflets have a 'brand building' effect similar to advertising (e.g. increased core demand and reduced price sensitivity). It will also demonstrate any reducing effect on core demand and increase in price elasticity that may result from promotions such as price reductions and multi-buys in the long term. Click here to request more information on IRI's study Understanding the Long-Term Effects of Advertising and Promotions.