Node Description
Tune Scenarios Node
The ‘Tune Scenarios’ node generates a single tuned ‘Output WTP Matrix’ based upon observations from a number of Market Conditions.
The ‘Input Market Array’ contains the Competitive Scenarios of different Market Conditions, along with the Products, Prices, and Quantities sold within each Scenario. The different Market Conditions might include historic observations from different sales seasons. Or the different Market Conditions might include Product Sensitivity data observed when Prices of the different Products are altered. Market Conditions vary when there are different competitive Products in the Market selling at Prices that change over time. As a result, the Quantity sold of each Product will typically be different under each Competitive Scenario.
The number of Products in each Competitive Scenario can vary. Products that are not included in one or more Market Conditions:
- may not yet have entered the Market,
- may have exited the Market, or
- may be Out-of-Stock.
This node assumes the level of Customer Demand is constant, with only the nature of the Competition in the Market changing. When Competitors change their Prices or Product Assortment, Customers will make different Purchase decisions even when their underlying preferences remain constant.
As a result, this ‘Tune Scenarios’ node is not designed to interpret seasonality. Only one Output Willingness To Pay (WTP) Matrix is generated, and this must represent the unchanging Willingness to Pay of Customers across all Market Conditions. To simulate seasonality, start with the Output WTP Matrix from this node, then scale the Mean (and SD) of the WTP Matrix with the ‘Scale WTP’ node.
This node is designed to find three values:
- Market Mean: The reference Mean of the WTP Customer Distributions across all Products in the Market.
- Market SD: The reference Standard Deviation (SD) of the WTP Customer Distributions across all Products in the Market.
- Market Correlation: The reference Correlation between all Customer Distributions within the Output WTP Matrix.
But individual Product Mean, SD, and Correlation values can be varied by the user according to the ‘Input Product Variations’ table and the ‘Conformity Boosters’ in the Configuration Dialog.
This Community Node documentation assumes you have already downloaded the open-source KNIME analytics platform and installed the free Market Simulation (Community Edition) plugin. If not, start by returning to Getting Started.
Downloads
#1 Market Array
The ‘Output Market Array’ can be set to output either the ‘Input Market Array’ or the ‘Input Product Attributes’. In either case, the Input is updated to reflect the tuned results from the Market Simulation.
In the first case, when outputting the ‘Input Market Array’, the updated ‘Quantity Error’ column contains the absolute value of the difference between the actual input ‘Quantity’ and the simulated output ‘Quantity’ from each Market Scenario. The Error can be weighed and summed by downstream nodes to determine the ‘Total Error’ has part of the Tuning Loop to tune the Market Model. Note that the Quantity Error for the ‘No Sale’ Product is not calculated unless the user explicitly includes the ‘No Sale’ Product in the Input Product Array and sets a target Quantity to be greater than zero.
Inputs
Market Scenarios
The ‘Input Market Array’ contains a list of Market Scenarios representing historic data from different seasons, or representing different degrees of Product Sensitivity at different Prices. Each Market Scenario contains a set of Products competing in the Market under the different conditions.
In the ‘Input Market Array’ above, there are 12 Market Scenarios containing information about the two Competitive Products as well as the number of ‘No Sale’ Customers.
Node
Configuration
The ‘Tune Scenarios’ node generates the overall ‘Market Mean’, ‘Market SD’ (Standard Deviation), and the ‘Market Correlation’. The ‘Starting’ value for each is set by the user to help the node quickly find the best fitting Tuning Parameters.
Here the user set the Output Market Array to be based upon the Input Market Array and all of the historical Market Scenarios.
Outputs
Port-0 Market Array
The ‘Tune Scenarios’ node creates a different Market Simulation for each of the Input Market Scenarios. In this case, 12 Market Simulations are created. The Output Market Array contains details of the quality-of-fit each Market Simulation has with each Input Market Scenario.
Port-1 WTP Matrix
The ‘Output WTP Matrix’ made up of the Customer Distributions for each Product column in the Market by each Virtual Customer row. The number of rows in the WTP Matrix is equal to the total ‘Number of Virtual Available Customers’ set by the user in the Configuration Dialog.
The Mean of each column will be set to the tuned ‘Market Mean’ and adjusted according to the Product ‘Quality’ variation in the ‘Input Product Attributes’ table.
The Standard Deviation (SD) of each column will be set to the tuned ‘Market SD’ and adjusted according to the Product ‘Niche’ variation in the ‘Input Product Attributes’ table.
The mutual Correlation of each column will be set to the tuned ‘Market Correlation’ and adjusted according to both the Product ‘Conformity’ variation in the ‘Input Product Attributes’ table, as well as the ‘Conformity Booster’ settings for same-Brand, same-Store, etc. Products in the Configuration Dialog.
Port-2 Correlation Matrix
The output set of correlations that define the relationship between each Product’s Customer Distribution column within the ‘Output WTP Matrix’. The ‘Output Correlation Matrix’ will be symmetrical such that the number of data rows match the number of columns. The Product correlation values depends upon: (a) the tuned ‘Market Correlation’, (b) the Product ‘Conformity’ variations in the ‘Input Product Attributes’ table, and (c) the ‘Conformity Booster’ settings for same-Brand, same-Store, etc. Products in the Configuration Dialog.
Port-3 KPI Indicators
The ‘Output KPI Indicators’ contain select information about the tuning process and the quality of the final results, including:
● Market Mean: the final Mean of all Product Distributions in the Market WTP Matrix (before the Product Means are individually modified by their ‘Quality’ variations).
● Market SD: the final Standard Deviation (SD) of all Product Distributions in the Market WTP Matrix (before the Product SD are individually modified by their ‘Niche’ variations).
● Market Correlation: the final Correlation between all Product Distributions in the Market WTP Matrix (before the mutual Product Correlations are individually modified by their ‘Boosters’ and ‘Conformity’ variations).
#2 Product Attributes
The results sent to the ‘Output Market Array’ can be very detailed or can only contain summary results. Selecting the ‘Input Market Array’ option will generate the more detailed results by causing the ‘Output Market Array’ to extend the original ‘Input Market Array’ table by appending the Quantity Error, Mean, and SD results to each Market Scenario.
In this case, the user has selecting the second option: ‘Input Product Attributes’. This will generate a less detailed summary ‘Output Product Array’ which includes only simulation results from the Prices and Products found in the ‘Input Product Attributes’ table. This ‘Output Product Array’ can be directly used with the ‘Output WTP Matrix’ by downstream Market Simulation nodes.
Inputs
Market Scenarios
These Input Market Scenarios are the same as above.
Product Attributes
Additional detail concerning each Product found within the ‘Input Market Array’. In most cases, this additional detail remains constant throughout all Market Scenarios, but in some cases this additional detail represents the default value used whenever the ‘Input Market Array’ contains missing values.
Product Variations
The ‘Input Product Attributes’ can contain ‘Quality’, ‘Niche’, and ‘Conformity’ variations.
Quality: represents the relative Quality the Product (between -2.0 and +2.0) has from the Market Mean. The Quality variation modifies the ‘Mean’ of the Product. Quality = 0.0 (default) means that the Product offers what is expected from the tuned Market (no change to the reference Market Mean). Quality = +1.0 means that the Product is a vast improvement from the norm (200% x Market Mean). Quality = -1.00 means that the Product is vastly inferior to the norm (50% x Market Mean).
Niche: represents whether the Product is relatively more appealing to a Customer Niche or to the mass market. The Niche variation (between -2.0 and +2.0) modifies the Standard Deviation (SD) of the Product. Niche = 0.0 (default) means that the Product offers what is expected from the tuned Market (no change to the reference Market SD). Niche = +1.0 means that the Product’s Customer Distribution has a wider variance than the Market norm (200% x SD). Niche = -1.00 means that the Product has a tighter variance than the Market norm (50% x SD).
Conformity: represents the degree of Conformity (between +1.0 and 0.0) the Product has from the Market norm, and the relative difference in Correlation the Product has with respect other Products. Conformity = 1.0 means that the Product precisely offers what is expected from the tuned Market (no change to the reference Market Correlation). Conformity = 0.0 means that the Product is vastly different from the Market norm.
Node
Configuration
The user has selected to include the ‘Input Product Attributes’ in the ‘Output Market Array’. This option allows the user to predict the Market Shares of Products from the tuned WTP Matrix given new Prices for each.
Boosters
The Conformity Booster inputs will increase the Correlation between Products having the same Brand, Store, Location, etc. Boosting is often required as same-Brand Products are perceived by Consumers as being similar, so the Correlation between same-Brand Products needs to be increased. Conformity Boost = 0.0 means that the Product Correlation will be unchanged and equal the tuned Market Correlation. Conformity Boost = 1.0 means that the tuned Market Correlation will be ignored and same-Brand Products will all be 100% correlated with one-another. Conformity Boost = 0.20 is typical, and will boost the Correlation between same-Brand Products by 20%. Hence if the final tuned Market Correlation = 0.5 then same-Brand Products will be boosted to have a mutual Correlation of 0.6.
Outputs
Port-0 Product Array
The ‘Tune Scenarios’ node can predict the Market Share of the Products found in the Input Product Attributes using the tuned WTP Matrix and the new Input Prices for the Products.
The Mean and Standard Deviation (SD) of the Willingness To Pay (WTP) Customer Distribution for each Product is different because of the ‘Quality’ and ‘Niche’ variations found in the ‘Input Product Attributes’.
Port-3 KPI Indicators
The Output ‘Market Mean’, ‘Market SD’, and ‘Market Correlation’ will vary as the Market Simulations for each of the 12 Market Scenarios will be tuned taking into account the ‘Quality’, ‘Niche’, and ‘Conformity’ Product Variations.
