Advanced Analytics Blog

A company knows how much it costs to make a product, and the company usually knows how much competitors are charging for competitive products. So, just choose a price that adequately covers all manufacturing costs and is below the prices charged by competitors. Eureka! Problem solved. Consumers will flock to the lower prices, and the manufacturer will live happily for evermore. If only it were so simple.

Prices are highly variable in most product categories, as input costs vary and as market demand and competitive pressures tug and pull prices up and down in response. There is always the conflict between the short-term and the long-term in pricing strategy. Reduce that short-term price too much, and you may never be able to ever return to the higher price. Stick with a higher price too long and you may be out of business.

So how does a company establish an optimal pricing strategy for one of its brands. Let’s explore how we might achieve this magical goal. Where would we start?

Qualitative Explorations

The beginning is good qualitative research. Depending on the product/service category, depth interviews and/or focus groups would likely be a part of the research plan, as would ethnography (or observation) if the purchasing process can be observed or videotaped. The goal of the qualitative research would be to understand the purchasing process, consumer motivations, and consumer behaviors in exhausting detail, including:

• How, when, and where is the product or service purchased? What’s the purchase context or environment? How many competitive products are available at point of purchase?
• How many packages are purchased at a time, and what is the mix by brand if more than one package is bought?
• What roles do price, competitive prices, and pricing differentials play in the brand purchase decision?
• What other factors are involved in the brand purchase decision and what roles do they play?
• What is the role of prices in the product category? How price sensitive are category consumers, and what are their motivations and attitudes related to prices?
• What makes one brand better than another? What is the justification or rationale to pay more for one brand versus another? What adds to the perception of extra value?
• What pricing differentials, brand versus brand, would trigger brand switching behavior, and why?

These questions would vary by product/service category, and there would be many more questions to ask and many more topics to explore during the qualitative investigation. Projective techniques, and subtle, clever questions by the moderator are the keys to understanding pricing at a deeper level. The findings and insights from this qualitative research would identify the most important variables and the ranges of those variables, uncover consumer motivations, and shape the quantitative work to follow.

Sales Data Analysis

The next step would be analysis of actual sales data, assuming such data were available and representative. Scanner data of category sales and individual brands’ sales from hundreds of widely scattered retail stores would be ideal. The sales data would span at least the last 12 months, and ideally the last two to five years. The datafiles would contain unit sales by each brand in the category at various price points by time periods. This data would be organized into acceptable formats, and advanced time-series regression techniques would be used in the analyses. Sales in dollars and units would be calculated using an equivalent market basket for all brands.

After formatting and data transformations, econometric models would be used to determine price sensitivity and elasticity for each brand. Specialized econometric techniques, such as fixed- and random-effects modeling, would produce consistent and reliable dollar and unit sales estimates as a function of pricing and terms.

The equations derived from these analyses would be used to build a DecisionSimulator™. This Simulator would allow researchers and brand managers to explore hundreds of pricing scenarios to optimize pricing strategies and reactions. For example, if competitor A lowered its retail price by 10 cents per package, what should your brand do in response? Stand pat? Cut price by 5 cents? Cut price by 10 cents? How would each of these responses affect brand share, and brand profitability? As is evident, these are short-term, tactical solutions, but knowing how to react to competitive threats is one part of the pricing optimization puzzle. It is important to note that pricing simulations with historical data should not extend beyond the range of prices that appear in the historical data.

Choice Modeling Pricing Optimization

Many competitive actions cannot be modeled based on sales data alone. For example, a competitor might launch an array of consumer promotions that might be invisible in the store scanner data, or a competitor might dramatically increase its advertising spending. These types of competitive actions can often be addressed via choice modeling, a family of tradeoff and conjoint techniques. A simple example will illustrate how choice modeling works.

A large survey of product category shoppers would be conducted. In this survey, a representative sample of category consumers would be asked to make brand choices (hence the term “choice modeling”) when presented with a shelf set of brands, along with pricing, promotion, and advertising information. Each survey participant would be asked to choose which brand he/she would buy, given a set of marketing inputs. This is called a “scenario,” and we can think of it as one shopping trip (or scenario one). Then the brands and the variables and prices all change, and the exercise is repeated. This is scenario two. This process repeats until 6 to 10 scenarios are completed by each participant, following a mathematically derived experimental design. The equations derived from these experiments would be used to build a new DecisionSimulator™, like the one described earlier, or these new equations could be used to build upon and expand the variables in the previous sales-based DecisionSimulator™. Regardless of the approach used, the model would be calibrated (adjusted) based on actual sales data to improve its overall accuracy.

Research analysts and brand managers could then explore hundreds of possible strategies and competitive reactions to improve short-term pricing and promotion decisions. While the findings and the results of the research stages to this point would dramatically improve pricing understanding and pricing strategies, the longer-term pricing issues would need to be addressed separately.

Strategic Positioning and Messaging

The key strategic questions are: what product attributes and qualities, what advertising positionings and messages, would provide the greatest support to your brand’s pricing over the next 20 to 30 years? The primary goal is to increase the pricing differential (vis-à-vis major competitive brands) that consumers are willing to pay for your brand. The qualitative research at the outset of this pricing study would have already identified the major positionings and sets of messages that could form the foundations for a long-term pricing strategy. These alternative positionings/messages would be translated into strategy concepts for rigorous testing, including both qualitative evaluation and quantitative evaluation.

The winning strategic positioning and related messages would become the strategic plan and the marching orders for everything related to the brand (product improvements, packaging communications, advertising positioning and messaging, quality levels versus competitive products, distribution strategy, sales training, etc).

Pricing Power

Pricing power is the net result of good long-term planning and consistent communication of the strategic positioning and messages, plus consistent and correlated execution by the product/brand itself. If you pursue a consistent strategic positioning crafted to build perceived value over time, you can enjoy the great benefits of higher profit margins and a defensive wall to protect your brand from competitive counterattack.

Jerry W. Thomas

Chief Executive Officer

Jerry founded Decision Analyst in September 1978. The firm has grown over the years and is now one of the largest privately held, employee-owned research agencies in North America. The firm prides itself on mastery of advanced analytics, predictive modeling, and choice modeling to optimize marketing decisions, and is deeply involved in the development of leading-edge analytic software. Jerry plays a key role in the development of Decision Analyst’s proprietary research services and related mathematical models.

Jerry graduated from the University of Texas at Arlington, earned his MBA at the University of Texas at Austin, and studied graduate economics at SMU.

• 
•