Price vs. Advertising: What Moves Sales More?

ConversionSethuraman, R.; Tellis, G. J. · 1991Journal of Marketing Research
Topicsprice elasticity·advertising elasticity·elasticity ratio·meta-analysis·price discounting·pass-through·product life cycle

You have one budget and must choose between a price promotion and a larger advertising push. Across the pooled studies, short-term sales are about 20 times more responsive to price than advertising, based on equal percentage changes. That comparison does not settle the budget choice because sales response is not profit. The profit case also depends on margin, current advertising relative to sales, retailer pass-through (how much of the deal reaches shoppers), and how many loyal buyers take the discount.

Short-term sales respond far more to price, while profit depends on more.

Data chart

Short-term sales respond more to price than advertising

How many times larger the short-term change in sales is after a 1% price cut than after a 1% increase in advertising (median of published estimates); the dashed line is the median for all estimates, not a product type.

All published estimates: 19.5 xNondurable goods25.7 xDurable goods5.5 x
  • This gap is not reliable once other study features are held constant.

Across published estimates, short-term sales respond far more to a price change than to an equal-percentage advertising change, most of all for nondurable goods in raw comparisons; this is sales response, not profit.

Medians of 260 short-term estimates from 16 published studies, 1960 to 1988, mostly of nondurable goods; each estimate divides the sales response to a 1% price change by the response to a 1% advertising change. The durable and nondurable medians differ in raw comparisons, but the gap is not reliable once other differences between the studies are held constant. These are sales responses, not profits.

Data chart

Matching a price cut's profit takes far more advertising for nondurable goods

Percent increase in advertising spending needed to earn the same extra profit as a 1% price cut, in the study's worked example.

50%Nondurable goods, advertising increaseneeded2%Durable goods, advertising increaseneeded
  • These figures come from a worked example, not from estimates.

In the study's worked example, earning the same extra profit as a 1% price cut takes a 50% advertising increase for a nondurable brand and a 2% increase for a durable brand.

A worked example under the paper's assumed values, not an estimate: a margin of half the price, advertising at 5% of sales, sales up 3% for each 1% price cut, and the share of regular demand bought on the deal equal to the share of the deal retailers pass on. Sales respond 25 times more to price than to advertising for nondurable goods and 5 times more for durable goods. The figures move with these assumptions.

Action guide

  1. Require separate sales-response and profit cases.Before approving a price or advertising budget, have the team deliver each case separately. The profit case must include margin, current advertising relative to sales, retailer pass-through, and how many loyal buyers take the discount.
  2. When both levers would add profit, favor discounts for mature or nondurable goods.Discounting appears more profitable for these goods; for products in their growth stage, both probably pay. The study tested maturity and product type separately.
  3. Keep both budget levers open for durable goods.When both would add profit, neither clearly leads. Durable makers' reported advertising looks low against a worked example in the study, though that model's advertising includes technical and sales-force support. Have finance check whether your durable lines' advertising and support fall short.
  4. Commission an advertising-spend review for nondurable brands.Have finance and marketing test current spending against your margin, retailer pass-through, and loyal-buyer data. The review must treat longer-term brand-building and loyalty effects as unmeasured, not established.
  5. Require sensitivity checks across measurement windows.Before approving the budget comparison, have your analytics team deliver results for yearly and finer windows, reading price effects from the most detailed data. Which window suits comparing price with advertising is still open.

Evidence

  • The median price-vs-advertising elasticity ratio is 19.5 across the pooled studies. Short-term sales are about 20 times more responsive to equal percentage price and advertising changes, not equal budgets.
  • In raw comparisons, the median ratio is 25.7 for nondurable goods versus 5.5 for durable goods.
  • Separately, the median ratio is 22.2 for mature products versus 17.7 for products in the growth stage.
  • A trade deal pays only when sales respond to price past a level set by margin, retailer pass-through and how many loyal buyers take the discount. More advertising pays only when sales respond past a level set by margin and current advertising relative to sales. Clearing neither, one or both calls for neither lever, mostly that one, or both.
  • For nondurable goods, the model's illustrative advertising range is about 4% to 8% of sales versus a reported industry range of 7% to 11%. The comparison assumes retailer pass-through and loyal-buyer values that the study did not estimate.
  • The raw median ratio is 17.9 with monthly-or-finer data versus 9.6 with yearly data. It is higher still with bimonthly or quarterly data.

Key takeaway

Compare short-term sales response to price and advertising, then decide with profit data.

Limitations

The studies cover estimates from 1960 to 1988, mostly U.S. and European nondurable goods, and measure short-term sales response to shelf price and advertising, not profit or long-term brand effects. Most studies ignore coupons, and responses hold only for small changes.

Source

Sethuraman, R., & Tellis, G. J. (1991). An analysis of the tradeoff between advertising and price discounting. Journal of Marketing Research, 28(2), 160–174. https://doi.org/10.1177/002224379102800204

Exact values behind the rounded figures: an overall median elasticity ratio of 19.5 (about 20 times); the worked example's ratios of 25 and 5 round the nondurable and durable medians of 25.7 and 5.5.

Read the paper ↗