Pricing research is an approach that seeks to determine consumers' willingness to pay for a product or service. The goal is to determine the optimal price point to maximize profit, revenue, and market share. However, we recommend against straight "willingness to pay" questions as consumers tend to underestimate how much they would pay — there are many more sophisticated pricing research methodologies that are relatively easy to set up and interpret.
Here are a few common pricing methodologies, and some of the pros and cons of using each.
Van Westendorp / Van Konan
aytm's Van Konan Price Optimization Model layers revenue and profit views onto the traditional Van Westendorp model, if one can provide assumptions for cost of production, TAM, etc.
In a Van Westendorp exercise, survey respondents are asked four questions:
- At what price do you think the product/service is priced so low that it makes you question its quality?
- At what price do you think the product/service is a bargain?
- At what price do you think the product/service begins to seem expensive?
- At what price do you think the product/service is too expensive?
Responses to each question are plotted as cumulative percent line graphs and, by looking at where specific lines intersect, we can identify an Optimum Price Point as well as a Range of Acceptable Prices.
| Pros | Cons |
|---|---|
| Pretty simple to execute. A good option when you do not have existing intel or a good starting point with regard to potential price points. Works well for established and frequently purchased categories (e.g., food/beverage/other CPGs). Does not require special sample size considerations as all respondents provide data for all questions. | Assumes the category has a price below which it is so cheap you would question quality — not necessarily a fair assumption for every category. For emerging categories, luxury, and tech, VW can yield low-ball estimates; consumers may not have the best sense of what they would actually pay for emerging products and may vastly underprice expensive products. |
Monadic Price Test
In a monadic price test, a range of prices are tested (usually 5 or 6 price points, evenly spaced) and respondents are randomly assigned to evaluate only one, in the form of purchase intent (scaled) or purchase choice (between multiple products — statically priced competitors + the product of interest). By charting differences in purchase intent/purchase choice and estimated revenue per XX (e.g., per 100 gen pop consumers) for each price point tested, we can identify the price point that maximizes market share and/or revenue, and examine price elasticity.
| Pros | Cons |
|---|---|
| Low respondent bias — monadic design eliminates bias towards pricing because respondents only evaluate one price point. They are not aware of the other prices being tested. Results are often clear-cut and provide strategic direction without ambiguity. | Requires a larger sample size than other approaches because of the monadic design that necessitates a minimum number of cases per cell. As a result, monadic studies can be more expensive. |
Price Ladder, Gabor Granger
Similar to a monadic price test conceptually, the primary difference with Price Ladder and Gabor Granger is that respondents don't evaluate just one price point. In both exercises, respondents are asked to indicate their purchase intent (scaled) or purchase choice at various price points. By charting changes in purchase intent/purchase choice and estimated revenue per XX for each price point tested, we can identify the price point that maximizes market share and/or revenue, and examine price elasticity.
- Price Ladder — respondent starts by evaluating the highest price point. If willing to purchase, they are discontinued from the exercise but coded as "YES" for each subsequent lower price point. If not willing, they are shown the next highest price point and asked again (rinse and repeat through the full range).
- Gabor Granger — respondents are randomly assigned to evaluate one price point. If willing to purchase, they are next shown a higher price point; if not willing, a lower one (rinse and repeat through the range).
Both are variations on a monadic price test and are generally considered inferior to the more "pure" monadic test, but require a far smaller sample size as each respondent ends up with a value for each price point tested. If sample costs/feasibility are issues, Price Ladder is recommended over Gabor Granger.
| Pros | Cons |
|---|---|
| Requires a much smaller sample size than monadic price testing, while allowing you to produce the same share, revenue, and price elasticity metrics. | Price Ladder: Respondents may "figure out" the test and eventually indicate they would purchase simply to exit the ladder. To combat this, consider fielding among category purchasers or those who show a baseline interest in the product. Gabor Granger: Evaluating a lower price point may "spoil" a respondent for higher price points. Respondents may rationally be inclined not to endorse a higher price after seeing a lower one, inviting price comparisons that can impact the data. |
Conjoint
The most flexible option for price testing — allows you to assess market share at various price points (any price point in a range), as well as revenue per XX and price elasticity. Conjoint also allows you to assess: A) the importance of price relative to other qualities of a product or service, and B) the actual value of specific features. A range of price points should be tested, and at least one other product variable must be included (conjoint requires a minimum of 2 variables — price is only one).
| Pros | Cons |
|---|---|
| Very robust and can assess many pricing scenarios, including changes in competitor pricing and the impact on the brand's own product share. Efficient use of sample — can be done with as few as 400 completes depending on the number of variables tested. | Can have higher service costs because it requires a custom simulator for analysis. The importance of price can be overestimated, particularly if the number of price points tested is considerably higher than the number of other variables. |