Market Simulation

Pricing Good-Better-Best Products

Nearly 100 years ago, Alfred P. Sloan introduced a “price ladder” to General Motors. The concept was designed to create “a car for every purse and purpose”. Alfred Sloan’s price ladder was designed to distinguish Chevrolets and Buicks from Oldsmobiles and Caillacs. It was an early example of “Good, Better, Best” (G-B-B) Pricing.

Today, G-B-B Pricing can be seen everywhere, from gas stations offering regular, mid-grade, and super-premium gasoline, to hardware stores offering multiple versions of almost everything.

G-B-B Pricing is a powerful, but two-edged sword. On the one hand, it allows companies to place multiple products along the Demand Curve to capture different levels of consumer Willingness To Pay (WTP).

But on the other hand, G-B-B Pricing introduces cannibalization into a company’s product assortment. Cannibalization occurs whenever a customer would have bought the more expensive version of a product, but was dissuaded by the cheaper version also offered by the same company.

Companies with small product assortments often use rudimentary spreadsheets when setting each G-B-B pricing level. These spreadsheets can’t possibly predict how customer behavior will change when prices are fine-tuned. These companies could substantially increase revenue and profitability by using more sophisticated market science when setting each G-B-B pricing level.

Companies with very large product assortments, such as eCommerce companies trying to sell all the way down the “Long Tail”, suffer from cannibalization that is out of control. In fact, chasing the “Long Tail” introduces so much cannibalization that gross revenue and profitability for these eCommerce companies would both rise if they were to cut their assortment. For these companies, Market Science is essential if they want revenue to increase when sales volume increases.

This Market Simulation compares a trial-and-error approach to Good-Better-Best (G-B-B) Pricing, with a Market Science approach based upon an understanding of the Willingness To Pay (WTP) of Customers.

This Case Study provides a high-level overview of the workflow without detailed explanation. It assumes you are already somewhat familiar with KNIME and Market Simulation. If not, start by reviewing the Building Blocks and Community Nodes.

Creating Products

The set of Good-Better-Best (G-B-B) Products used in this Market Simulation are generated using the common Differentiation Workflow Pattern introduced in:

A reference Product is created in the ‘Standard Product Feature’ Table Creator node (#1). Variations are then described in the ‘Good Better Best Product Variations’ Table Creator node (#2).

Quality: The Quality Variation will change the Vertical Differentiation of each Product. The Quality of the ‘Best’ Product (Quality=1.0) is twice as good as that of the reference ‘Good’ Product (Quality=0.0).

Niche: The Niche Variation will change the Strange Differentiation of each Product. The ‘Best’ Product is more appealing to niche customers than the reference ‘Good’ Product.

Expense: The Expense Variation will change the Cost of each Product to the vendor. The vendor’s Cost of the ‘Best’ Product (Expense=1.0) is twice as much as that of the reference ‘Good’ Product (Expense=0.0).

Conformity: The Conformity Variation will change the Horizontal Differentiation of each Product. The ‘Best’ Product (Conformity=0.6) will have a Willingness To Pay (WTP) Customer Distribution that has a correlation of 0.8 compared to reference ‘Good’ Product (Conformity=1.0).

For a more detailed description of Quality, Niche, Expense, and Conformity, see the documentation for:

Good-Better-Best Variations

Vertical Differentiation

Horizontal Differentiation

Product Array

WTP Matrix

Method #1: Trial and Error

The Trial-and-Error method starts with a list of Good, Better, and Best test prices. Each price combination is then tested in turn within the workflow loop. Each loop iteration will calculate the predicted total Quantity sold, Revenue, and Profitability generated across the whole Store. Finally, the results are sorted from ‘Most Profitable’ G-B-B Price levels, to ‘Least Profitable’.

The initial price levels were:

  • Good Product = $100
  • Better Product = $150
  • Best Product = $200

The Trial-and-Error found a combination of price levels that could raise profitability by over 50%.

The improved pricing levels were:

  • Good Product = $100
  • Better Product = $150
  • Best Product = $210

List of Test Prices

Single Iteration Result

All Iteration Results

Method #2: Market Science

While Method #1 ‘Trial-and-Error’ can be effective, it quickly becomes overwhelming to come up with a set of test prices, even when there are only a small number of products to optimize.

Method #2 uses a dedicated ‘Price Maximize’ node to seek a Profit / Revenue / Quantity maximizing set of G-B-B price levels.

The ‘Price Maximize’ node is designed to continuously change the Price of a small set of Focus Products until the Maximization Goal is reached. Both increases in Price as well as decreases in Price are tested. The Maximization Goal can be to maximize Profit, Revenue, or Quantity Sold.

In this Market Simulation, the ‘Price Maximize’ node has been configured to simultaneously change the Prices of all the Products of ‘MyStore’ to maximize total Store Profit.

The initial price levels (again) were:

  • Good Product = $100
  • Better Product = $150
  • Best Product = $200

The Market Science approach found a combination of price levels that could raise profitability by over 60%. The improved pricing levels were:

  • Good Product = $113
  • Better Product = $170
  • Best Product = $226

See also:

Configure Node

Optimized Price Levels

Optimized Totals