Looking for the sweet spot

Editor's Note: Tom Rigby is founding partner at Callosum Marketing Inc., a Montreal research firm.

Among all of the stages in the product development process, setting the price is arguably one of the most important. This decision has implications across departments and business functions and typically exerts a significant influence over consumers’purchase intentions.

Ultimately, the powers that be need to select a price that is high enough to protect their margins but not so high that it scuttles demand and renders any marketing campaigns futile. This challenge is true for products already on the market but it becomes a particular point of contention when new products are being launched and for which there is no historical pricing or in some cases, no competitive brands to use as benchmarks. 

Fortunately, a variety of quantitative research techniques exist that can help clarify a product’s optimal price point. In this article, I will elaborate on three such techniques, explain how to apply each one and point out their respective advantages and disadvantages. 

The Gabor-Granger technique

Developed in the 1960s by economists André Gabor and Clive Granger, this technique gives researchers a simple way to test various price points and to calculate the price elasticity of demand.

The process begins by showing respondents a description of a product as well as a proposed price and then asking them how likely they would be to buy the product at that price. Deciding which price point to show first is up to the researcher but options include:the middle price, the current retail price or selecting at random. For the purposes of this example, I have decided to begin with the middle price of five options (Figure 1).

If, at this step, the respondent indicates a positive response (“definitely would” or “probably would”), they are shown the product again but now with a higher price and then asked for their purchase intent once more. This higher price can be chosen at random or pre-set.

If, however, the respondent indicates a neutral or negative response at the first price point (“might or might not,” “probably would not” or “definitely would not”) the process then works in reverse. That is to say, the respondent is shown a lower price and asked again for their purchase intention. 

This pattern is repeated multiple times until the highest price point a respondent is willing to pay has been established. Once the data has been compiled, it is graphically represented as the demand curve (i.e., the percentage of respondents who indicated a positive purchase intention at each price point). These results are then used to also calculate the revenue curve (i.e., the percentage of respondents who would purchase the product at each price point multiplied by that price). Overlaying these two graphs highlights the price/demand combination that will lead to the highest average revenue per respondent. 

For example, in Figure 2, we see that 40 percent of respondents are likely to buy the product at a price of$8.00. This translates to an average revenue per respondent of $3.20 (40 percent times $8.00; see Figure 3). After performing this calculation at each of the other price points and plotting the revenue curve, $3.20 proves to be the highest average revenue per respondent and, therefore, we have a clear indication that retailing the product at $8.00 is the best option of the ones tested (see Figure 4).  

At this point, a researcher might feel satisfied with their conclusions and provide their recommendations.However, the Gabor-Granger technique has another important application that often gets overlooked: the ability to calculate the category’s price elasticity of demand. This metric measures the responsiveness of demand to changes in price and demonstrates the implications of lowering or raising prices in the category. To determine the category’s overall price elasticity of demand, we must first calculate the elasticity between each price point (Figure 5).

Going back to our previous example, we can see that decreasing the price by 20 percent (from $10.00 to$8.00) increased demand by 30 points (from 10 percent to 40 percent). This yields an elasticity of -1.5 (30/-20). By repeating this calculation between each of the other price points and then taking the average of all four results,we obtain the overall category elasticity of -0.7 (Figure 6). This suggests that the category is relatively inelastic and that for every percentage-point drop in price, demand could be expected to increase by 0.7 of a percentage point. With these results, the decision makers can make more informed sales projections when adjusting prices, like during periods of sales and promotions.

As is the case with most research techniques, the Gabor-Granger approach has both advantages and disadvantages. On the positive side, it is most celebrated for its simplicity and convenience. With just five questions, this technique provides compelling evidence of which price point is best and also clarifies the relationship between price and demand in any given category. 

However, the disadvantages of this technique are twofold. Firstly, it does not ask the respondent to trade-off price for other product attributes and, in doing so, ignores the many variables that could influence purchase intention (for example, available budget, competitive pricing, brand affinity, etc.). The other main criticism of this technique is that it uses a sequential ordering of questions. If the respondent is quick to catch on that a positive purchase intention leads to a higher price and a negative purchase intention leads to a lower price, there is some potential for bias in their response options. Consumers are, after all, typically motivated to pay the lowest price. 

Due to these advantages and disadvantages, researchers often use an alternative technique that retains the best of the Gabor-Granger approach while mitigating its potential for bias: monadic price testing. 

Monadic price testing

From the respondent’s perspective, taking part in a monadic price testing exercise seems very simple and straightforward. They are shown a product description and a price and are asked how likely they would be to buy the product at that price. However, unbeknownst to the respondent, behind the scenes of this technique there is a more sophisticated analysis at work. 

When employing a monadic price-testing approach, the overall sample of respondents is in fact divided into various groups of an equal size, with respondents being randomly assigned to each one. For a new product, there will be as many groups as there are prices to test. For a product already on the market, there will be one group that is exposed to the current price and then one additional group for each new price point that is to be tested. 

The key to this technique is that while each group is asked the same question, they are each exposed to a different price point (Figure 7). At no point in the study are respondents aware that other prices are being considered and it is this feature that helps mitigate any concerns of influence or bias.

Once the survey has been launched and the data is compiled, the researcher compares results between the groups in order to determine how a change in price affected demand. The same process that was used to plot the demand and revenue curves in the Gabor-Granger technique can be applied to the outputs of the monadic test and the price elasticity of demand is once again easily determined. 

As mentioned previously, the main advantage of this approach is that it produces the same results as the Gabor-Granger technique but without the potential for respondents to be biased by the fact that their answers dictate whether they see a higher or lower price option. Another advantage is that this technique is even more convenient as it requires each respondent to only complete one question.

There is, however, one downside to this approach and that is its necessity for larger sample sizes. For this approach to be properly executed, each price point needs to be tested by a separate group of respondents. So if, for example, 10 price points were going to be tested, a sample of 1,000 would likely be needed in order to result in a sufficiently robust sample of 100 respondents per group. Requiring this larger sample size means that this technique is almost always more expensive to run than Gabor-Granger.

Another point to consider with this technique, as well as Gabor-Granger, is that both only test consumers’reactions to prices that have been provided. As with any aided question this requires certain assumptions on the part of the researcher and leaves no room to understand consumers’ spontaneous perceptions of value and price. For this, we must employ an alternative approach: the Van Westendorp price sensitivity meter. 

The Van Westendorp technique

Introduced in 1976, Dutch economist Peter Van Westendorp developed his eponymous technique in order to determine a product’s optimal price point without forcing respondents to react to predetermined options. More specifically, the Van Westendorp approach works by asking respondents four open-ended questions:

• At what price would you think this product is too expensive to consider?
• At what price would you think this product is so inexpensive that you would question its quality and not consider it?
• At what price would you think this product is getting expensive but you would still consider it?
• At what price would you think this product is a bargain – good value for the amount being charged?

The results are then combined to form a graph with four curves, one for each question. A variety of conclusions can be drawn from the graph but the two most important are the “optimal price point” and the “optimal price range” (Figure 8). 

The optimal price point is the point of intersection between the “too expensive” and the “too cheap” curves. At this point, an equal number of respondents think that the price is too high has think it is too cheap. Based on this equal trade-off, this price is typically considered to be the best option to move forward with. In our example, taken from a study on a brand of computers, that optimal price point would be about $950. However, rather than providing one specific number, the researcher can also advise on an appropriate range. This range extends on the lower end from the intersection of the “getting expensive” curve and the “too cheap” curve. The upper end would be the point of intersection between the “too expensive” curve and the “bargain” curve. In theory, these lines set the upper and lower limits for optimal pricing and if the client chooses not to go with the optimal price point (either because they think it is too high or not high enough) they can instead select another price within the optimal range.

The main advantage of the Van Westendorp technique is that all of the responses are obtained via consumers’own input and thus this approach clarifies existing perceptions. There is,however, some debate as to whether this technique is appropriate in all categories. For example, one major tenet of the Van Westendorp technique is that there is a point at which the price is so low that it would affect perceptions of quality: “too cheap.” While this may be true in high-involvement categories like automobiles or medical procedures, some question whether it is also true in low-involvement categories (like paper towels or toothbrushes). If there is in fact no price that is “too cheap” for consumers, the approach loses its relevance. This being said, on an overall basis, most practitioners tend to think of this technique as being a mainstay of pricing research.

Difficult to know which one is best

With the above options at every researcher’s disposal, it can sometimes be difficult to know which is the best one to recommend to clients. In my practice, when budget permits, I advise my clients to begin the survey with a Van Westendorp analysis and follow it up with a monadic test. This ensures that we obtain a clear understanding of consumers’ spontaneous perceptions on price and quality in the category but that we also have the opportunity to directly test purchase intentions at the prices that my client is considering. This combined approach allows us to coverall the bases and also demonstrates the size of the gap that exists between the highest price that my client wants to charge and what consumers spontaneously demonstrate to be the optimal price. The larger the gap, the more my client will need to reconsider their decisions.

If, however, budget does not permit the use of both techniques, the decision usually comes down to whether the client has a clear idea of potential prices or whether they are more flexible and curious to see how respondents feel spontaneously. In the former case, a monadic test should be used. In the latter, the Van Westendorp approach should be used.

Leaving money on the table

For all of the time and effort spent crafting ad campaigns designed to increase demand, there often seems to be a lack of market research behind the pricing strategy. As shown in this article, with just a few short survey questions, many brand managers may find that they are in fact leaving considerable amounts of money on the table or that by making just a small reduction in price, they could generate a big gain in demand. With all of these potential benefits, the question may not be whether you should incorporate pricing research but whether you can afford not to.