This research report was released to members of the Ehrenberg-Bass Corporate Sponsorship Program in 2018. Following recent publication of the research in the Journal of Advertising Research, and subsequent interest from industry we’ve decided to provide public access to this report.

Our Corporate Sponsors have exclusive access to hundreds of reports like this, as well as videos, marketing commentary, access to Q&A with our researchers and many other resources. If you would like to learn more about the benefits of signing up as a Corporate Sponsor please contact us

What happens to sales when brands stop advertising for long periods?

By Adam Gelzinis, Rachel Kennedy, Virginia Beal, Nicole Hartnett and Byron Sharp

 

Key findings

  • When brands stop advertising for a year or more, we find sales often decline year-on-year following the stop (on average, sales fell 16% after one year, and 25% after two years).
  • The rate of decline is fastest for brands that are already declining before the advertising stop.
  • Brand size also matters. Small brands typically suffer greater declines than bigger brands.
  • Bigger growing brands tend to continue to grow after advertising stops for one to two years, whereas the sales trend quickly reverses for small growing brands.
Note: Further research is underway to document the generalisability of these findings beyond this single study.

Introduction

The sales of a brand are like the height at which an airplane flies. Advertising spend is like its engines: while the engines are running, everything is fine, but, when the engines stop, the descent eventually starts.

— Simon Broadbent

 

Advertising spend can be cut for many reasons. In a recession or other times of financial pressure — e.g., when resources are urgently required in other areas of the business — advertising may seem like a discretionary activity or a dispensable luxury. Because the advertising budget is typically non-fixed, it can be diverted quickly at short notice. In some cases, poor budget planning or extra spending due to seasonality (or certain high reaching events) may run the budget dry for some time until the brand’s next allocation of funds.

Stopping a brand’s advertising and pocketing the budget is also an easy way to improve net profit. Such financial decisions are known as ‘milking’ the brand, and though it may be an effective solution for a business in the short term, it is a gamble with potentially high stakes. It’s a gamble because little is known about the likely consequences.

So regardless of why brands stop advertising, it is important that we understand what happens when they do. How are sales affected? How big might the consequences be?

 

What does past research tell us?

Studies of brands stopping advertising are rare and few of these examine sales effects after stopping advertising for extended periods (i.e. 1+ years).

Split cable experiments — where two matched groups of households are exposed to different amounts of brand advertising, other variables held constant — have been used to look at what happens when brands stop TV advertising for up to one year. When one of those groups receives no advertising and the other group sees normal levels of advertising, we get a relatively clear view of the sales effect of stopping advertising.

More than 150 split cable tests of advertising stops have been conducted from the early 1990s to the mid 2000s (Lodish et al. 1995; Riskey 1997; Hu et al. 2007; Hu et al. 2009). Across all tests, just over half showed no significant sales difference between the two groups. This finding suggests there is a 50/50 chance that stopping TV advertising will have no effect on sales within a year. Of the tests that did report a significant difference, sales were lower on average for the unadvertised group (i.e. a lack of advertising was associated with less sales). Separating tests by brand size, small/new brands more often suffered declines compared to big/established brands.

Other in-market tests in the last forty years have found much the same. Tests for Budweiser beer in the US showed that sales remained stable after one year without advertising, and small declines appeared only after 18 months (Ackoff and Emshoff 1975). Others also found minimal effects after 8 to 24 months of reduced advertising (i.e., not complete cuts) (Aaker and Carman 1982; Eastlack and Rao 1989).

Stability for a year or so after a stop is common, which suggests there may be a ‘safe’ period where brands can cut advertising without severe consequences. While the evidence suggests nothing good comes of it, stopping advertising is not necessarily bad for sales.

But the studies discussed above refer only to cutting TV advertising, all come from the US, and tend to study larger, well established brands (i.e., high mental and physical availability). We need to extend this work to determine how these results apply to a wider set of conditions, or when brands cut all mass media advertising, or stay unadvertised for multiple years.

Longer periods of up to five years have been explored by some advertising budgeting studies. This stream of research looks at consistent under- and over-spending on advertising (relative to benchmarks derived by brand size) and corresponding changes in market share. The Ehrenberg-Bass Institute has shown that brands that remain under-advertised for multiple years are likely to lose share (Danenberg et al. 2016), which is consistent with other research (Binet and Field 2007; Hansen and Bech Christensen 2005; Jones 1990). However, under-spending is not the same as ceasing advertising altogether.

These past studies provide clues, but offer little generalisable knowledge of what happens when advertising stops for a long time. It is not clear why sales are stable in some cases but decline in others, nor when decline is typically first noticed, or how severe it may be. To shed some light on these questions, we present the findings of new research from the Ehrenberg-Bass Institute.

 

Our study

The data for this study tracked the media spend and volume sales of 70 competitive brands in an Australian consumer goods category for over 20 years. This dataset allowed us to observe if/when brands stopped and started advertising over the years and how their sales evolved following these changes.

We identified 57 cases where a brand cut all mass media spending for a year or longer¹. The sample of cases includes different sized brands and brands with varying sales trends (e.g., growing, stable or declining). Some brands stopped advertising for many years, whereas others stopped for only one (see Table 1).

Table 1: How long brands stopped advertising
There are 57 cases of a brand stopping advertising for one year. Of those 57 cases, 34 remained unadvertised for a second year, and so on (the other 23 left the sample because they started advertising again or were delisted from the data).

 

For each case, we documented brand sales in unadvertised years and compared this to sales in the last advertised year. We could then see how aggregate sales changed over time when advertising stopped relative to when the brand was advertising².

To compare cases, we created a comparable measure of sales change. For each case, brand sales in the last year with advertising were given an index value of 100, and sales in every other year were benchmarked against this value. An index of 110 shows that sales are 10% greater than the index year, while an index of 75 shows a 25% decrease in sales (see Table 2 for an example).

Table 2: Hypothetical example of indexed sales (illustrating a rather dramatic decline)
Sales in the last year with advertising (the index year) are set at 100. Sales in surrounding years are benchmarked against the index year as the base. Comparing the prior year with the index year tells us that this brand declined in sales before it stopped advertising, as its sales were 10% higher (i.e. index 110 → 100). When the brand stopped advertising, sales declined by 20% after one year (index 80), 40% after two years (index 60) and 45% after three years (index 55).

 

Once this indexing is applied to all cases, the average index can be calculated across all cases in each year.

Sales typically decline gradually after advertising stops

The average trend in sales after advertising stopped was a steady decline that eventually tapered. After one year without advertising, the mean sales index across all cases was 84 (see Table 3). This means that the average change in sales after stopping advertising for one year was a 16% drop. Average change in sales after two unadvertised years was a 25% drop, and after three years a 36% drop.

 

Table 3: Mean indexed sales when brands stop advertising
* Although the average prior index shows decline before stopping advertising, there were near equal numbers of previously growing, stable and declining brands in this sample.

 

Not all brands experienced such steady decline in sales after stopping advertising. Some brands remained stable initially and others even grew (this variation is discussed below). To illustrate this variation, Chart 1 plots all cases as a series of points over time to show how sales evolved across all brands when they stopped advertising. The ‘mean sales index’ in each year is displayed as a solid line.

In the first year without advertising, the mean sales index is 84, though there is wide variation around this average. The spread widens after two years without advertising, yet there is a downward trend in sales for the majority of cases. Despite the variation, decline becomes more common and greater in magnitude as brands go longer without advertising. After four years without advertising, there are no cases with higher-than-base level sales.

Chart 1 also shows a small number of brands that reported higher sales in the first year or two without advertising than previously when they were advertising. Most of these growth cases were brands that were growing already (more on that below), and they were quick to restart advertising.

 

Chart 1: All cases of brands stopping advertising

 

Conditions matter: Brand size and prior sales trends

To clarify why some brands performed better after stopping advertising than others, we split the sample of cases into groups to explore different conditions.

An important condition identified by prior research is brand size. Many of the Institute’s reports demonstrate how the performance of different brands, across a variety of metrics, is related to their market share (e.g. see Corporate Sponsor Reports 26 and 53). Bigger brands are bought by more people and tend to be bought more often. They also enjoy greater mental and physical availability, which means their advertising tends to be more efficient. In this dataset, brands varied quite a lot in sales volume.

To assess whether brand size had some effect on the outcome of stopping advertising, brands were separated into three roughly even groups: big, medium and small brands. Average yearly sales volume was the basis for determining brand size; small brands had <250K units, medium brands had 250K to 1M units, and larger brands more than 1M units.

Another condition that may explain the variation is prior trend in sales. The decision to stop advertising is unlikely to be made lightly and it will probably factor in the past sales performance of the brand. In this dataset, there were brands that were on vastly different trends in sales prior to stopping advertising.

To assess whether preceding trends explain sales changes after stopping advertising, we separated cases into three groups: previously growing, stable and declining brands. Sales trends were identified by the sales index for the year before the final advertised year; indexes below 90 are considered growing, indexes from 90 to 110 are considered stable, and indexes above 110 are considered declining.

The sample of cases can also be organised by both brand size and prior sales trend. These two-factor groups make some patterns clearer, however, for most combinations, there are too few cases to show clear differences.

 

Conditions matter: Bigger brands decline slower on average

Chart 2 plots the average sales indexes over time for the three brand size groups. Smaller brands (n=23) declined faster than the medium brands (n=17), and the medium brands declined faster than the big brands (n=17) on average. Big brands appeared somewhat stable after one to two years without advertising, and no big brands had a sales index less than 20 after stopping advertising (i.e. none were delisted, even after nine years without advertising). The medium and small groups both include brands that declined to near zero sales and were ultimately delisted.

 

Chart 2: Results split by brand size

This ‘size advantage’ for bigger brands is consistent with the Institute’s research that shows lower ad elasticities for larger brands, and lower ad intensiveness (Danenberg et al 2016), and fits with the evidence regarding mental (and physical) availability. A brand’s sales rest largely on its relative availability amongst consumers, and bigger brands have greater mental availability than small brands. Stopping advertising means brands cannot build or refresh mental networks through mass communication, but other nudges come from buying or using the brand, seeing other people buy or use the brand, or seeing in-store displays and activations (which typically also favour bigger brands). The upshot is that bigger brands’ greater mental and physical availability will likely better insulate sales from decline after stopping advertising compared to smaller brands.

 

Conditions matter: Sales trends before stopping relate to trends after

Chart 3 plots the average sales indexes over time for three prior sales trend groups. Every brand that was declining before stopping (n=20) continued declining in the following years without advertising (irrespective of brand size). The average rate of decline was fastest for already declining brands. If a declining brand goes unsupported, these patterns speak to dire consequences; sales halved on average in a period of two years.

 

Chart 3: Results split by prior sales trend

Previously stable brands (n=19) also held a reasonably consistent trajectory (stability on average) in the first two years following the stop. Stable brands that stayed unadvertised for more than two years did then start to experience substantial decline from that point.

The initial stability may be due to investment in marketing activities other than advertising, but it is also likely because of the largely habitual purchasing of consumers. In repeat purchase situations, when consumers choose between alternative options, habitual repetition of behaviour manifesting as a bias or preference (e.g. brand loyalty) is the norm. For stable brands that stopped advertising, this ‘inertia’ in purchase propensity may contribute to the continued stability.

Brands that were growing before advertising stopped (n=18) experienced very different sales changes after stopping if we consider their size. All big and medium sized brands that were previously growing (n=8) continued to grow for 1-2 years after stopping. In stark contrast, all previously growing small brands (n=10) stopped growing and declined below their base-level sales. This suggests an advantage for bigger brands akin to greater “momentum” in sales growth. An already-large and growing brand may continue to grow without advertising support because of scale advantages in other marketing functions and market-based mental and physical availability. While smaller brands appear to rely more heavily on their advertising to drive mental availability and continue growing.

 

“What if we stop for one year, then come back on-air in the following year?”

There were 14 cases where a brand stopped advertising for one year, then resumed in the following year. Table 4 reports how year-to-year sales trends changed for these brands, as they stopped then restarted advertising.

Before the ‘pause’ in advertising, there were near equal numbers of growing (n=5), stable (n=5), and declining brands (n=4). After a year of no advertising fewer of our 14 brands were growing and more were declining. Resuming advertising in the third year did not halt this trend. The number of cases growing (by 10% or more in year-to-year sales) continued to fall, and even more cases moved into decline (in sales by 10% or more). Growing brands tended to grow slower (or stop growing altogether and become stable), and stable brands started seeing decline.

 

Table 4: Sales trends; stopping then restarting advertising (14 brands)

Only 2 of the five growing brands and 3 of the stable brands managed to keep their sales trend after resuming advertising for a year. After going silent for one year then returning to air, the other half of the initially growing or stable brands were unable to return to their previous sales trends. This suggests that it may take longer than a year of spend to make up for a hiatus of a year — possibly because the pause leads to physical availability losses.

 

Summary & Implications

It seems that Broadbent was largely right to liken an unadvertised brand to an airplane without engines. Sales declined for most brands that stopped advertising, and decline became more common and greater in magnitude as brands went longer without advertising. A key exception was brands that were both large and growing before they switched off advertising; these large growing brands tended to maintain an upward trajectory without mass media advertising (even after two years in some cases). Noting that all these brands turned advertising on again after their pause of a year or two, so we weren’t able to study what might have happened if they had stopped for longer.

While it may be tempting to withdraw the advertising budget for a boost in profits, the evidence suggests that doing so risks putting the brand on a downward sales trajectory. Without refreshment, mental availability erodes. Silent periods extend the gap between consumers making a category purchase and them last seeing brand advertising. In this gap they may be nudged by a competitor’s advertising, or your critical memory-based brand linkages will erode, and so they are less likely to think about you, especially if they are a light buyer. And since light brand buyers are the biggest group of any brand’s customer base and are the most important group for brand growth (see Report 73), the time off-air could prove costly, particularly for smaller brands.

Maintaining a brand’s market share requires a sufficient share of voice for its size. Negative deviations from this benchmark (i.e. underspending) can lead to losses in market share. This study does, however, present evidence that big growing brands may escape serious negative consequences in the short term when advertising is paused, though they might have grown even faster with advertising support.


¹ A full year without advertising ensured we had identified a genuine stop in advertising and not just a brief gap in a schedule.
² The analysis was also repeated with brand market share as the target metric, to control for changes in category size. The results align with the sales analysis, which suggests that category expansion/shrinking did not markedly effect brand sales.

Media Resources

Further questions for our research team?

References

Aaker, D & Carman, J (1982), ‘Are You Overadvertising? A Review of Advertising-Sales Studies’, Journal of Advertising Research, vol. 22, no. 4, pp. 57-70.

Ackoff, R & Emshoff, J (1975), ‘Advertising Research at Anheuser-Busch, Inc. (1963-68)’, Sloan Management Review, vol. 16, no. 2, pp. 1-15.

Binet, L & Field, P (2007), Marketing in the Era of Accountability, World Advertising Research Centre, Oxfordshire, United Kingdom.

Danenberg, N, Kennedy, R, Beal, V & Sharp, B (2016), ‘Advertising Budgeting: A Re-Investigation of the Evidence on Brand Size and Spend’, Journal of Advertising, vol. 45, no. 1, pp. 139-146.

Eastlack, J & Rao, A (1989), ‘Advertising Experiments at the Campbell Soup Company’, Marketing Science, vol. 8, no. 1, pp. 57-71.

Hansen, F & Bech Christensen, L (2005), ‘Share of Voice/Share of Market and Long-Term Advertising Effects’, International Journal of Advertising, vol. 24, no. 3, pp. 297-320.

Hu, Y, Lodish, L & Krieger, A (2007), ‘An Analysis of Real World TV Advertising Tests: A 15-Year Update’, Journal of Advertising Research, vol. 47, no. 3, pp. 341-353.

Hu, Y, Lodish, L, Krieger, A & Hayati, B (2009), ‘An Analysis of Real World TV Advertising Tests: A Recent Update’, Journal of Advertising Research, vol. 49, no. 2, pp. 201-206.

Jones, J 1990, ‘Ad Spending: Maintaining Market Share’, Harvard Business Review, vol. 68, no. 1, pp. 38-43.

Lodish, L, Abraham, M, Kalmenson, S, Livelsberger, J, Lubetkin, B, Richardson, B & Stevens, M (1995), ‘How TV Advertising Works: A Meta-Analysis of 389 Real World Split Cable TV Advertising Experiments’, Journal of Marketing Research, vol. 32, no. 2, pp. 125-139.

Riskey, D (1997), ‘How T.V. Advertising Works: An Industry Response’, Journal of Marketing Research, vol. 34, pp. 292-293.

image-description

Now available as an eBook exclusively to Apple iBooks

image-description

The Ehrenberg-Bass Institute for Marketing Science is the world’s largest centre for research into marketing. Our team of market research experts can help you grow your brand and develop a culture of evidence-based marketing.