DEMAND FORCASTING FOR COMPANY SUSTAINABILITY

         A thorough understanding customer preferences and price sensitivity is important in predicting how they will respond to a pricing policy. By understanding how price influences brand preference and choice, you will be able to more profitably capture a fair share of the value your products and services provide customers. SMD advocates an approach that includes 4 steps:  
1). Segmentation: Generally the factors that influence price sensitivity will vary across different groups called segments. A failure to properly segment the market to account for customer heterogeneity may result in price sensitivity estimates that have poor predictive capabilities. 
2). Preference Analysis: The key to effective preference analysis is to determine the product features or attributes that drive customer choice and then to determine how preference changes as price and other attributes change. There are three broad quantitative approaches to estimating customer price sensitivity.

• Monetary Value to the Customer (MVC) Analysis: estimates price sensitivity based on factors that have a direct economic impact on customers, such as the cost of using or maintaining the product.
• Survey-driven estimates: Typically some form of trade-off or conjoint analysis is used to estimate the importance of variations in a product's price and features. The result is a preference function that assigns a value (either in dollar or utility terms) that can be used to measure the attractiveness of different potential alternatives.
• Revealed preference: When sufficient customer purchase data is available, customer sensitivities to changes in prices, product features, and marketing variables on the likelihood that a customer or market segment will choose a product from a specified set of alternatives based on their actual purchase behavior.When appropriate several of these techniques may be applied jointly. In the absence of survey or economically derived estimates, the judgments of product managers and sales representatives may be used as surrogate measures.

3). Preference-Driven Choice Modeling: The value or utility function that is estimated using the Preference Analysis is the used to predict a customer's choices from a specified set of alternatives.  When market choice data is available, a Logit (discrete choice) procedure can be used to predict choices from a set of different alternatives. By observing how a segment's likelihood of choosing a specific alternative's price, it is possible to estimate a segment's price sensitivity for that alternative as well as their responses to changes in product attributes. Also, when feasible, it is useful to determine whether a customer's perceptions of product features correspond to objective measures of them. When they differ, choice should be based on the customer's perceptions.

4). Demand Modeling: In developing a market demand model the preference driven choice models are enhanced to account for marketing activities, distribution variables, and other factors that influence choice that are not readily measured in a preference analysis. Each respondent or segment is given a weight that reflects their impact on market sales and the choice model is calibrated to the best available market sales data. The resulting model can be used to predict how any firm's sales, market shares, and profits will vary with changes in price or another key attribute.