I recently found an academic paper that suggests advertising raises the willingness to buy a brand at any given price but increases demand more when the brand is offered at lower prices. So, what are the implications?
A timeless topic
The paper by Tulin Erdem, Michael P. Keane and Baohong Sun is titled The Impact of Advertising on Consumer Price Sensitivity in Experience Goods Markets, and dates from 2006. Yes, that is a long time ago, but the subject matter is timeless: how does advertising influence price elasticity?
To answer the question, the authors used Nielsen weekly household level panel data for toothpaste, toothbrushes, detergent, and ketchup to examine how TV advertising and other marketing activities affect a brand's demand curve. The analysis is done at the household level but seeks to identify the effect of marketing on the overall demand curve. The time frame was three years of data for toothpaste and toothbrush but less for detergent and ketchup, and the analysis was limited to households that bought at least three times.
Modeling buyer behavior
Once again, I cannot offer an opinion of the actual model used, although I have it on good authority that it is technically sound. Beyond the math, there are a couple of things that make good sense to me. First, the model is designed to allow for heterogeneity of demand, i.e., buyers are differentially affected by price, advertising, and lagged purchases. Second, that purchase of a brand improves the probability a buyer will buy again, which allows for the formation of habitual purchasing (lagged purchases).
Advertising shifts the demand curve
In brief, for 17 out of 18 brands examined, increasing advertising intensity shifts the demand curve up, but more so at lower price points. Simulations indicate that the demand curve tends to "pivot" around the highest price point (see the accompanying chart). This finding implies that there is a core of buyers willing to pay at any price but a wider group only willing to buy at lower price points. The authors summarize their findings as follows,
"As we see, advertising increases demand at any given price, implying that it increases willingness to pay. At the same time, advertising flattens the demand curve, because WTP increases more among consumers whose WTP was relatively low initially."
It is important to note that while the incremental effect is smaller, advertising does seem to stimulate more demand even at the highest price points. And, given the analysis was limited to households that bought three times, it seems possible that lower prices might generate even more demand from infrequent category purchasers.
Not every brand has the same demand curve
The odd brand out of the 18 is Heinz Ketchup, unique in having a market share of 61% and few direct competitors. A simulation shown in the paper implies that advertising has more effect on people already willing to buy at higher prices (the demand curve is concave, not convex like the other brands). The authors suggest that one reason for this segmented appeal is the brand's association with thickness, a property highlighted for decades in the brand's advertising.
Maybe there is a group of people who really do not appreciate Heinz's evident thickness, but I suspect a simpler explanation. I note that at Heinz's lowest price point the average probability of purchase increases to nearly 90%. In other words, almost everyone is in-market for Heinz when the price is right. Even at a substantially lower price point, Hunt's only has a purchase probability of just over 60%. Years of brand experience and advertising have made Heinz the gold standard for ketchup, and it has become a staple for most households.
Not all buyers value brands
One finding from this paper that coincides with my own experience is that it finds different people respond differentially to price, advertising, and promotion. Table 2 in the paper reports some descriptive statistics, including category level purchasing behavior which indicates that there are strong differences in willingness to pay the given price for any brand. This makes perfect sense to me because not everyone can afford to buy the brand they want, even if they do believe that buying the "right" brand matters. It makes me wonder, if the panel participants had been surveyed, would we have found that those willing to pay the higher prices believed the brands they paid for were meaningfully different from the others?
Direct and indirect effects of advertising
Because the model uses lagged TV ad exposures (adstocks) which allows for the interaction between exposure interval and purchase interval, it takes several weeks for the direct effect of advertising to be realized and then even longer for the indirect repurchase effect to be fully realized. The balance between direct and indirect influences on demand appears to be about one third direct to two thirds indirect, implying that by stimulating an initial purchase advertising has a strong multiplier effect via repeat purchase.
Reconciling findings about price elasticity
The analysis finds that by expanding a brand's customer base disproportionately toward people who value the brand less, advertising increases demand but lowers the average price paid. In other words, advertising seems to raise price elasticity, not lower it.
This dilution effect would appear to fly in the face of studies that find advertising lowers long-term price elasticity, but does it? The model finds that people who buy a brand are more likely to buy it again, and the results suggest people who exhibit stronger "loyalty" formation also exhibit less price sensitivity. In addition, they appear less likely to respond to displays, features, and coupons.
As a result, one can envisage the proportion of core buyers – the ones that are likely to buy at any given price – might expand over time, with habitual purchasing reinforced by further advertising. Price sensitive buyers who find the brand acceptable come and go as the price fluctuates but their influence on demand lessens as the core grows. Overall, the brand's price elasticity declines and eventually you end up with a brand that behaves more like Heinz than Hunts.
And scale brings its own rewards
There is another aspect that we need to consider related to brand growth, the advantages of scale can offset the potentially negative effects of any brand dilution. While they might not be a Heinz Ketchup, big brands are the most obvious choice in their category and the easiest to buy because most retailers will stock them. As a result, big brands have the benefit of higher repeat purchase rates than small ones. And big brands – particularly consumer packaged goods brands – benefit from economies of scale, an advantage that is often used to lower prices and further increase volume market share, rather than increasing profits. We should not assume that long-term changes in price elasticity are only the result of consumer attitudes.
Findings consistent with the buyer cycle
If my interpretation is correct, I see a dynamic process by which brands grow over time, consistent with the buyer cycle detailed in this post.
So where does that leave us?
Assuming I have interpreted this paper correctly, and the results can be reconciled with those from other studies, I think these findings provide useful insight into how advertising influences both demand and price paid for consumer packaged goods. Of course, it does raise some interesting questions when it comes to measuring the impact of advertising. For years I have assumed that a major benefit of advertising is to reduce price elasticity, but here we have a study that suggests the opposite. Potentially, more effective advertising will bring in more marginal consumers in the short-term who buy at a low price and then stop buying when prices rise. This short-term effect may be offset as proportion of core buyers expands over the long-term. If so, this makes it even more important to segment people by their willingness to buy at a profitable price point, focus on building the brand's core buyers, and monitor underlying trends over time.
A timeless topic
The paper by Tulin Erdem, Michael P. Keane and Baohong Sun is titled The Impact of Advertising on Consumer Price Sensitivity in Experience Goods Markets, and dates from 2006. Yes, that is a long time ago, but the subject matter is timeless: how does advertising influence price elasticity?
To answer the question, the authors used Nielsen weekly household level panel data for toothpaste, toothbrushes, detergent, and ketchup to examine how TV advertising and other marketing activities affect a brand's demand curve. The analysis is done at the household level but seeks to identify the effect of marketing on the overall demand curve. The time frame was three years of data for toothpaste and toothbrush but less for detergent and ketchup, and the analysis was limited to households that bought at least three times.
Modeling buyer behavior
Once again, I cannot offer an opinion of the actual model used, although I have it on good authority that it is technically sound. Beyond the math, there are a couple of things that make good sense to me. First, the model is designed to allow for heterogeneity of demand, i.e., buyers are differentially affected by price, advertising, and lagged purchases. Second, that purchase of a brand improves the probability a buyer will buy again, which allows for the formation of habitual purchasing (lagged purchases).
Advertising shifts the demand curve
In brief, for 17 out of 18 brands examined, increasing advertising intensity shifts the demand curve up, but more so at lower price points. Simulations indicate that the demand curve tends to "pivot" around the highest price point (see the accompanying chart). This finding implies that there is a core of buyers willing to pay at any price but a wider group only willing to buy at lower price points. The authors summarize their findings as follows,
"As we see, advertising increases demand at any given price, implying that it increases willingness to pay. At the same time, advertising flattens the demand curve, because WTP increases more among consumers whose WTP was relatively low initially."
It is important to note that while the incremental effect is smaller, advertising does seem to stimulate more demand even at the highest price points. And, given the analysis was limited to households that bought three times, it seems possible that lower prices might generate even more demand from infrequent category purchasers.
Not every brand has the same demand curve
The odd brand out of the 18 is Heinz Ketchup, unique in having a market share of 61% and few direct competitors. A simulation shown in the paper implies that advertising has more effect on people already willing to buy at higher prices (the demand curve is concave, not convex like the other brands). The authors suggest that one reason for this segmented appeal is the brand's association with thickness, a property highlighted for decades in the brand's advertising.
Maybe there is a group of people who really do not appreciate Heinz's evident thickness, but I suspect a simpler explanation. I note that at Heinz's lowest price point the average probability of purchase increases to nearly 90%. In other words, almost everyone is in-market for Heinz when the price is right. Even at a substantially lower price point, Hunt's only has a purchase probability of just over 60%. Years of brand experience and advertising have made Heinz the gold standard for ketchup, and it has become a staple for most households.
Not all buyers value brands
One finding from this paper that coincides with my own experience is that it finds different people respond differentially to price, advertising, and promotion. Table 2 in the paper reports some descriptive statistics, including category level purchasing behavior which indicates that there are strong differences in willingness to pay the given price for any brand. This makes perfect sense to me because not everyone can afford to buy the brand they want, even if they do believe that buying the "right" brand matters. It makes me wonder, if the panel participants had been surveyed, would we have found that those willing to pay the higher prices believed the brands they paid for were meaningfully different from the others?
Direct and indirect effects of advertising
Because the model uses lagged TV ad exposures (adstocks) which allows for the interaction between exposure interval and purchase interval, it takes several weeks for the direct effect of advertising to be realized and then even longer for the indirect repurchase effect to be fully realized. The balance between direct and indirect influences on demand appears to be about one third direct to two thirds indirect, implying that by stimulating an initial purchase advertising has a strong multiplier effect via repeat purchase.
Reconciling findings about price elasticity
The analysis finds that by expanding a brand's customer base disproportionately toward people who value the brand less, advertising increases demand but lowers the average price paid. In other words, advertising seems to raise price elasticity, not lower it.
This dilution effect would appear to fly in the face of studies that find advertising lowers long-term price elasticity, but does it? The model finds that people who buy a brand are more likely to buy it again, and the results suggest people who exhibit stronger "loyalty" formation also exhibit less price sensitivity. In addition, they appear less likely to respond to displays, features, and coupons.
As a result, one can envisage the proportion of core buyers – the ones that are likely to buy at any given price – might expand over time, with habitual purchasing reinforced by further advertising. Price sensitive buyers who find the brand acceptable come and go as the price fluctuates but their influence on demand lessens as the core grows. Overall, the brand's price elasticity declines and eventually you end up with a brand that behaves more like Heinz than Hunts.
And scale brings its own rewards
There is another aspect that we need to consider related to brand growth, the advantages of scale can offset the potentially negative effects of any brand dilution. While they might not be a Heinz Ketchup, big brands are the most obvious choice in their category and the easiest to buy because most retailers will stock them. As a result, big brands have the benefit of higher repeat purchase rates than small ones. And big brands – particularly consumer packaged goods brands – benefit from economies of scale, an advantage that is often used to lower prices and further increase volume market share, rather than increasing profits. We should not assume that long-term changes in price elasticity are only the result of consumer attitudes.
Findings consistent with the buyer cycle
If my interpretation is correct, I see a dynamic process by which brands grow over time, consistent with the buyer cycle detailed in this post.
- Higher advertising intensity increases brand salience and, as a result, draws buyer attention to the brand's price, displays, features, and coupons.
- Advertising and promotion work together to increase a buyer's willingness to pay a given price, but they have most effect on those who are only willing to pay a lower price. Core buyers willing to buy at high prices will likely buy anyway.
- If people do buy, they become more likely to buy again provided the brand meets their expectations and is easily available, and also more responsive to further advertising and less responsive to price and promotion.
- Over the long-term, the proportion of core buyers expands, and demand becomes less sensitive to changes in price (in part because bigger brands acquire advantages of scale).
So where does that leave us?
Assuming I have interpreted this paper correctly, and the results can be reconciled with those from other studies, I think these findings provide useful insight into how advertising influences both demand and price paid for consumer packaged goods. Of course, it does raise some interesting questions when it comes to measuring the impact of advertising. For years I have assumed that a major benefit of advertising is to reduce price elasticity, but here we have a study that suggests the opposite. Potentially, more effective advertising will bring in more marginal consumers in the short-term who buy at a low price and then stop buying when prices rise. This short-term effect may be offset as proportion of core buyers expands over the long-term. If so, this makes it even more important to segment people by their willingness to buy at a profitable price point, focus on building the brand's core buyers, and monitor underlying trends over time.
But then maybe there are alternative explanations? Oh, and obviously, this analysis only applies to packaged goods. What differences might we expect to see for durables and services? Please share your thoughts.