M/A/R/C Research discusses the best methodologies when pricing merchandise in the retailer industry.

1. Pricing Research in the Retailer Category
Properly setting price is the most direct way to ensure optimal profitability. Improper pricing, either
pricing too low (wherein per-unit revenue decreases faster than demand increases) or pricing too
high (wherein demand decreases faster than per-unit revenue increases), leaves money on the table.
The key to optimizing price is accurately measuring price-demand relationships, the functions that
quantify how much demand will be garnered at each possible price. With accurate price-demand
curves, basic economics 1 can be used to identify pricing that maximizes revenue or profit.
A simple price-demand curve and its associated price-revenue curve are shown below.
Price-Demand Price-Revenue
75%
$93.0
65% $92.5
Revenue (million $)
$92.0
55%
Demand
$91.5 $11.60 =
45%
the price that
$91.0 will maximize
35% revenue
$90.5
25% $90.0
$10.00 $10.50 $11.00 $11.50 $12.00 $12.50 $13.00 $13.50 $14.00 $14.50 $15.00 $10.00 $10.50 $11.00 $11.50 $12.00 $12.50 $13.00 $13.50 $14.00 $14.50 $15.00
Price Price
M/A/R/C believes the most appropriate methodology for pricing analysis—for accurately identifying
price-demand relationships—in the retailer category is alternative-specific discrete choice analysis
(DCA). In DCA, all relevant retailers are studied as separate alternatives in a competitive context that
mimics the real-life options that shoppers face. DCA is highly customizable with the ability to study
many types of effects and/or numerous prices simultaneously. As such, a final DCA model can be
used to set multiple prices (e.g. grocery prices, household products prices, electronics prices, etc.)
simultaneously such that overall store profit is maximized. And so the above example charts for the
retailer optimizing numerous prices simultaneously would be multidimensional and not merely two-
dimensional. For example, below are three-dimensional charts involving the simultaneous
optimization of two prices.
1 Revenue = Price * Demand; Profit = Revenue – Costs.

2. Price-Demand Function (2 prices) Price-Revenue Function (2 prices)
50%
45%
40%
35%
30%
25%
% Demand 20% Revenue
15%
10%
5%
0%
$2
$6
$2
$10
$6
$14
$10
$18
$14
$22
$18
$26
$22
$30
$10
$26
Price B
$34
$20
$30
$10
$38
$34
$30 Price B
$42
$20
$38
$40
$46
$30
$42
$50
$50
Price A $40
$46
$50
$50
Price A
A DCA model culminates in a simulator wherein prices can be set at any price along the entire
continuum of price ranges studied and not simply at the discrete price points studied as test stimuli.
Further, M/A/R/C overlays years of experience in calibrating DCA-model simulators to adjust for any
survey biases, so that predicted demand estimates are as close to true market share as possible. This
precise calibration is accomplished by using DCA in concert with M/A/R/C’s Assessor® product
forecasting methodology.
M/A/R/C’s Assessor, a proprietary forecasting methodology originally developed at MIT and proven
successful in thousands of practical applications, is one of only a few widely recognized, validated,
research-based forecasting models in existence.
Below are some highlights of the DCA methodology in the retailer category.
» The Discrete Choice Exercise for Respondents (Shoppers)
1. Expose survey respondents to a variety of scenarios that include retailer brands that well
represent the category and vary the average prices for relevant categories (e.g. grocery,
clothing, electronics, etc.).
2. Optionally, also vary other attributes of interest besides prices, such as availability of
offerings, product features (sizes, package configuration), store features, etc.
3. Ask survey respondents to allocate purchases across the retailers shown.
» Benefits
1. Experimental design is utilized to select realistic combinations of levels of attributes – but also
ensuring attributes are studied with statistical rigor (balanced and orthogonal) and then
unique blocks of scenarios are shown to each respondent.
2. Respondents are allowed to allocate purchases in categories where alternative options are or
will be available in-market.

3. 3. The test provides a realistic competitive context (multiple retailer brands in each choice
scenario; price and other attributes customized to the brand) and influences the respondent
to provide realistic responses to price.
4. Gaming is not likely, due to the even-handed treatment of alternative brands.
5. Any “over focus” on Price can be diffused by presence of other attributes.
» Potential issues
1. DCA is more costly than other forms of pricing analysis.
In the retailer space, in addition to DCA, it can be helpful to conduct as a supplemental component to
the overall pricing assessment what is known as van Westendorp pricing analysis. For a retailer it
would be conducted for each of the category prices – again, such as grocery, clothing, electronics, etc.
– as the average of typical item’s or typical basket of goods price. The van Westendorp technique is
easy to implement, involving four questions wherein respondents directly state their expectations
regarding price levels. As such, being based on simple, direct “stated” measures, the results of van
Westendorp have high face validity; respondents are merely stating where they expect pricing to be.
While some researchers extend the basic van Westendorp method to include the identification of
what they believe to be an optimal price point, we at M/A/R/C do not believe the simple “stated”
data of the van Westendorp method can be used for such and do not use van Westendorp as a
surrogate for rigorous price-demand estimation. Instead we use the van Westendorp method only to
form a range – a window – of prices within which shoppers expect pricing to fall. For identifying
optimal pricing we rely on DCA-based price-demand relationships. However, the results from DCA are
enhanced by comparing them to the information from van Westendorp, where we can answer such
things as:
n Does the optimal price as identified from the DCA fall within the van Westendorp acceptable
range?
n How wide is the range of expected pricing?
n How sharply does acceptance fall off if price goes beyond the acceptable range?
Below is an example chart that represents the final deliverable of a van Westendorp pricing analysis
(e.g. for a given category within the retailer brand) – the green highlight area represents the range of
acceptable pricing, where here “Acceptable” does not have anything to do with maximizing revenue
or profit but is simply referring to being acceptable in the eyes of shoppers.

4. 100%
Range of Too Cheap
90% Acceptable Pricing NOT Bargain
["acceptable" = in NOT Premium
80% minds of purchasers, Too Expensive
not necessarily from
70% a rev or profit
Cumulative %
standpoint]
60%
50%
40%
30%
20%
10%
0%
$6 $7 $8 $9 $10 $11 $12 $13 $14 $15 $16
Price
Low=$10.50
High=$12.25
Beyond the basic technical aspects of estimating a good price-demand relationship, pricing analysis
always also involves knowing and properly modeling all the category-specific, extenuating, and/or
complicating nuances of the market within which the product or service being priced will be sold.
Current competitor prices and positionings should be a strong consideration in setting price.
Knowledge about government regulations regarding pricing in the category or otherwise how prices
are negotiated is important.
Given the high importance of pricing and the many complicating facets that must be considered,
M/A/R/C Research has devoted extensive effort in building deep expertise in pricing analysis. With
many years of experience spanning all industry categories, we’re able to leverage our well-known
Assessor® methodology to ensure an accurate estimation of the price-demand relationship, offer a
wide array of modeling and simulation capabilities to custom fit most any set of market dynamics,
and to deliver a truly optimized pricing strategy for the retailer category.
M/A/R/C® Research (http://marcresearch.com/) is a custom marketing research firm dedicated to helping clients create, evaluate and
strengthen their brands. M/A/R/C’s teams design and execute qualitative and quantitative, traditional and online solutions while
adhering to a client-service ethic built on being easy to work with and delivering what is promised. Our core competency is measuring
attitudes and behaviors to accurately explain and predict market share, revenue and bottom-line impact of a client’s actions. M/A/R/C
has been successfully designing, executing and analyzing studies to help clients across a range of industries since 1965. We deliver
research answers with a business perspective in the language of decision-makers.
For more information about M/A/R/C Research contact your Account Manager, or send inquires to Marketing@MARCresearch.com.
Strong brands start with smart research. 800.884.6272