WO2007130527A2 - Systems and methods for business to business price modeling using price elasticity optimization - Google Patents

Abstract

The present invention relates to business to business market price control and management systems. More particularly, the present invention relates to systems and methods for generating price modeling and optimization modules in a business to business market setting wherein continuous pricing feedback is used to update and calibrate said optimization. Furthermore, competitive behavior is modeled and incorporated into the optimization process.

Description

SYSTEMS AND METHODS FOR BUSINESS TO BUSINESS PRICE MODELING USING PRICE ELASTICITY OPTIMIZATION

BACKGROUND OF THE INVENTION

[0001] The present invention relates to business to business market price control and management systems. More particularly, the present invention relates to systems and methods for optimizing prices in a business to business market setting wherein continuous pricing feedback is used to calibrate said optimization. [0002] There are major challenges in business to business (hereinafter B2B") markets which hinder the effectiveness of classical approaches to price optimization. [0003] For instance, in B2B markets, a small number of customers represent the lion's share of the business. Managing the prices of these key customers is where most of the pricing opportunity lies. Also, B2B markets are renowned for being data-poor environments. Availability of large sets of accurate and complete historical sales data is scarce.

[0004] Furthermore, B2B markets are characterized by deal negotiations instead of non-negotiated sale prices (prevalent in business to consumer markets). There is no existing literature on optimization of negotiation terms and processes, neither at the product/segment level nor at the customer level.

[0005] Finally, B2B environments suffer from poor customer segmentation. Top- down price segmentation approaches are rarely the answer. Historical sales usually exhibit minor price changes for each customer. Furthermore, price bands within customer segments are often too large and customer behavior within each segment is non-homogeneous. [0006] Product or segment price optimization relies heavily on the quality of the customer segmentation and the availability of accurate and complete sales data. In this context, price optimization makes sense only (i) when price behavior within each customer segment is homogeneous and (ii) in the presence of data-rich environments where companies sales data and their competitors' prices are readily available. These conditions are met almost exclusively in business to consumer (hereinafter "B2C") markets such as retail, and are rarely encountered in B2B markets. [0007] On the other hand, customer price optimization relies heavily on the abundance of data regarding customers' past behavior and experience, including win/loss data and customer price sensitivity. Financial institutions have successfully applied customer price optimization in attributing and setting interest rates for credit lines, mortgages and credit cards. Here again, the aforementioned condition is met almost exclusively in B2C markets.

[0008] There are three major types of price optimization solutions in the B2B marketplace: revenue/yield management, price testing and highly customized optimization solutions.

[0009] Revenue/yield management approaches were initially developed in the airline context, and were later expanded to other applications such as hotel revenue management, car rentals, cruises and some telecom applications (e.g. bandwidth pricing). These approaches are exclusively concerned with perishable products (e.g. airline seats) and are not pricing optimization approaches per se. [0010] Price testing approaches attempt to learn and model customer behavior dynamically by measuring customer reaction to price changes. While this approach has been applied rather successfully in B2C markets, where the benefits of price optimization outweigh the loss of a few customers, its application to B2B markets is questionable. No meaningful customer behavior can be modeled without sizable changes in customer prices (both price increases and decreases). In B2B markets, where a small fraction of customers represent a substantial fraction of the overall business, these sizable price-changing tests can have adverse impact on business. High prices can drive large customers away with potentially a significant loss of volume. Low prices on the other hand, even for short periods of time, can dramatically impact customer behavior, increase customers' price sensitivities and trigger a more strategic approach to purchasing from the customers' side.

[0011] Finally, in B2B markets, highly customized price optimization solutions have been proposed. These solutions have had mixed results. These highly customized price optimization solutions require significant consulting effort in order to address companies' unique situations including cost structure, customer and competitor behavior, and to develop optimization methods that are tailored to the type of pricing data that is available. Most of the suggested price changes from these solutions are not implemented. Even when they are implemented, these price changes tend not to stick. Furthermore, the maintenance of such pricing solutions usually requires a lot of effort. This effort includes substantial and expensive on-going consulting engagements with the pricing companies.

[0012] These solutions have failed primarily because of the lack of reliable price control and management systems. In fact, in B2B markets, reliable price control and management systems may be significantly more complex and more important than price optimization modules.

[0013] There remains a need for effective price control and management systems coupled with straightforward price optimization modules which can perform price changes in an effective manner, and can measure the performance of these price changes across the sales and marketing organizations, and across product and customer segments, both for existing business (repeat business) and new business.

[0014] Furthermore, instead of developing highly customized company-specific price optimization solutions, there remains a need for scalable and customizable price optimization solutions that vary by industry vertical.

[0015] Price control and management systems are employed in business in an effort to gain efficiencies and increase margin. Businesses employ a myriad of enterprise resource planning tools in order to manage and control pricing processes.

[0016] In particular, in the context of business to business markets, effective price modeling and optimization schemes have been elusive given the scarcity of sales data and the relatively small pool of available customers. In this environment, it is important to include all available relevant data, including competitive behavior data, in order to develop robust price modeling and optimization schemes. It is also important to continuously loop back to update and calibrate the price modeling and optimization schemes with new sales data generated from deals consummated with the benefit of the instant price modeling and optimization schemes.

[0017] As such, methods for generating effective business to business price optimization modules, as well as systems and methods for incorporating business to business price modeling and optimization modules into an integrated price control and management system in order to optimize a selected business objective, may be desirable to achieve system-wide price management efficiency.

[0018] In view of the foregoing, Systems and Methods For Business to Business

Price Modeling Using Continuous Learning and Calibration are disclosed. SUMMARY OF THE INVENTION

[0019] The present invention discloses business to business market price control and management systems. More particularly, the present invention teaches systems and methods for optimizing and modeling prices in a business to business market setting wherein continuous pricing feedback is used to update and calibrate said optimization.

[0020] In one embodiment, the method comprises generating a preferred set of prices by selecting a product in a selected market segment, providing sales data corresponding to said product, generating a demand model for said at least one product utilizing said sales data, and generating the preferred set of prices utilizing a price elasticity demand model.

[0021] The method also provides for using deal history data to generate a win probability model, and using said win probability model in conjunction with said price elasticity demand model in generating said preferred set of prices.

[0022] In another embodiment, the method comprises providing competitive behavior data and generating a competitive behavior model utilizing a Nash equilibrium computation, wherein said competitive behavior model is used in generating said preferred set of prices.

[0023] The method further contemplates providing price guidance data for the product, generating a preferred set of guidance prices utilizing the price guidance data, and reconciling the preferred set of prices with the preferred set of guidance prices to generate a reconciled set of prices.

[0024] It is further contemplated by the instant method that market segmentation is optimized and that the selected market segment is selected from a preferred set of market segments. [0025] In another embodiment, the instant method includes providing a set of goals and constraints wherein the preferred set of prices is generated to meet the goals and constraints. The method then provides the set of preferred prices to a price control and management system, and generates a quotation utilizing the price control and management system such that the set of preferred prices is incorporated into the quotation.

[0026] The method of the instant invention also provides that the set of preferred prices may be overridden by user input. The user may then generate a deal utilizing the price control and management system wherein the deal includes a set of final prices. The set of final prices are used to calibrate the demand model, the win probability model, and the competitive behavior model.

[0027] Note that the various features of the present invention described above can be practiced alone or in combination. These and other features of the present invention will be described in more detail below in the detailed description of the invention and in conjunction with the following figures.

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention is illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which like reference numerals refer to similar elements and in which:

[0029] Figure 1 is a high level flowchart illustrating a method for optimizing prices in a business to business environment in accordance with an embodiment of the present invention. [0030] Figure 2 is a more detailed flowchart illustrating a method for cleansing sales history data prior to its use in an optimization scheme in accordance with an embodiment of the instant invention.

[0031] Figure 3 is a flowchart illustrating a method for generating a demand model for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0032] Figure 4 is a flowchart illustrating a method for optimizing business segmentation for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0033] Figure 5 is a flowchart illustrating a method for providing deal win/loss classification data for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0034] Figure 6 is a flowchart illustrating a method for generating a demand model for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0035] Figure 7 is a flowchart illustrating a method for reconciling optimized prices optimized price guidance for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0036] Figure 8 is a flowchart illustrating a method for generating optimized prices for use in a business to business price optimization system in accordance with an embodiment of the instant invention.

[0037] Figure 9 is a flowchart illustrating a method for using a Nash equilibrium computation in generating optimized prices for use in a business to business price optimization system in accordance with an embodiment of the instant invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0038] The present invention will now be described in detail with reference to selected preferred embodiments thereof as illustrated in the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art, that the present invention may be practiced without some or all of these specific details. In other instances, well known process steps and/or structures have not been described in detail in order to not unnecessarily obscure the present invention. The features and advantages of the present invention may be better understood with reference to the drawings and discussions that follow.

I. OVERALL SYSTEM

[0039] To facilitate discussion, Figure 1 is a high level flowchart illustrating a method for optimizing prices in a business to business environment in accordance with an embodiment of the present invention. A framework for comprehensive price optimization along the operational or product segment level of pricing is presented. The instant price optimization system 100 comprises sales history data at step 110. Demand is modeled at step 120. Prices are optimized at step 130. A deal is negotiated at step 140. The deal is analyzed at step 150. Pertinent aspects of the deal analysis are sent back to the sales history database at step 160. Each of these steps will be discussed in more detail below.

[0040] Historical sales data is used by the demand modeling step 120 to model demand for a selected product or segment. The demand modeling step 120 is followed by the price optimization step 130. The optimization step 130 uses the demand models provided in generating a set of preferred prices for the selected product or segment. The optimization step 130 is followed by the deal negotiation step 140, where the preferred . prices may be used by a sales force in negotiating deals with customers. [0041] A learning and calibration process follows the completion of the deal negotiations. The resulting deals (i.e., quoted prices with customers) may be provided back as deal history data for iterative optimization. The learning and calibration process is carried out in steps 150 and 160. Information from the negotiated deals may be used in the learning and calibration process to update and calibrate the demand modeling and price optimization processes.

. DEMAND MODELING MODULE

[0042] FIG. 2 is a flow chart further illustrating step 110 of FIG. 1. In step 110, raw product/segment sales history data is provided so that product/segment demand models may be generated. FIG. 2 illustrates a method of taking raw product/segment sales data and cleansing the raw data to produce a cleansed sales dataset.

[0043] The process of dataset creation and cleaning begins by inputting raw deal history data at step 210. Raw order history data is input at step 220. The raw data is then subjected to cleansing at steps 230 and 240. Data cleansing includes accounting for missing or incompletes data sets as well as correcting or removing statistical outliers. For example, removing transactional outliers may include removing transaction data indicating sales dollars of zero or of an order of magnitude higher than a calculated average. Data cleansing may also include removing transactions with inconsistent data such as an order quantity of zero. Data cleansing may also include supplementing missing data with derived data. For example, missing region data may be set to a default region. The cleansed order history dataset is then output at step 250. The cleansed dataset is used in generating a demand model at step 120.

[0044] FIG. 3 is a high level flow chart further illustrative of the demand modeling step 120 of FIG. 1. The operation of the demand modeling step 120 will be discussed in general here and in more detail below in the discussion of FIGS. 4, 5, 6 and 7. Preferably, before modeling demand for a particular product/segment, the business segment must first be selected at 310. Sales history data for the selected product/segment is provided at 320. Preferably, win/loss classification data, which defines a deal as a win or a loss based on comparison to the selected industry segment average net margin for the selected product/segment, is provided as well at 320. Both, the sales history data and the win/loss classification data are used to model demand at 330.

[0045] FIG. 4 is a more detailed flow chart further illustrating the business segment selection step 310 of FIG. 3. The effectiveness of both the demand modeling and price optimization for the selected segment is dependent upon proper segmentation. Segmentation is defined so as to group products and customers which can be expected to have sufficiently similar characteristics. Segments are defined by policy managers based on qualitative segmentation information. For instance, policy managers may define segments based on combinations of product and customer types, geography, markets, etc. At a minimum, attributes differentiating price (e.g. customer type) are used to define segments which may be selected at step 310. Preferably, a list of all relevant segments for key products is available for selection at step 310. [0046] Business segments can be static (non changing) or dynamic (changing over time). Examples of static business segments include: Product segments: Product Family, Product Group, Product Type (e.g., Commodity, Specialty, Competitive), Product Use (e.g. Core Products, Add-on Products, Maintenance Products); Customer segments: Customer Geography, Customer Region, Customer Industry, Customer Size, Customer Relationship (e.g. Primary provider, Spot Purchase, Competitive). Examples of dynamic business segments include: Product segments: Product Lifecycle (New, Growing, Mature, End-of-life), Product Yearly Revenue Contribution (A = Top 30% of total revenue, B = Next 30%, C = Bottom 40%), Product Yearly Profit Contribution, Customer segments: Customer Yearly Revenue Contribution, Customer Yearly Profit Contribution, Customer Product Purchase Compliance (customers who orders less than certain percent of quoted products), Order Compliance (customers who orders less than committed volumes from quote or contract), Payment Compliance (customers who pays their invoices outside of pre-agreed payment terms defined in quote or contract).

[0047] As discussed above, business segments are defined by policy managers at step 410. Preferably, once the initial business segmentation has been accomplished, the segmentation is optimized at 420, giving a preferred business segmentation structure. An optimized business segmentation structure gives the advantage of enabling the generation of more precise product/segment demand models. Segmentation optimization is done by policy managers on an ongoing basis. Updated information is used to continuously calibrate and refine the currently optimized business segmentation scheme.

[0048] Given a currently optimized business segmentation scheme at step 420, the particular business segments relevant to the products or business segments in question for the instant price optimization process may be selected at step 430. The instant price optimization process then continues with the provision of sales history and win/loss classification data at step 320. [0049] FIG. 5 is a flow chart further illustrating step 320 of FIG. 3. In step 32O₅ win/loss classification data is provided. In order to effectively classify deals, raw deal • history data is provided so that product/segment win probability models for the particular product/segment in question may be generated. The raw order and deal history data is cleansed as discussed above in reference to FIG. 2. [0050] The cleansed order and deal history dataset is input at step 510. The data is used to generate deal win/loss parameters at step 520. Deal win/loss data is used to tune the ultimate price optimization process to account for real world results given optimized price sets.

[0051] Deals are classified as wins or losses based upon a comparison between deal transactions (quotes and/or contracts) and order transactions. The matching logic compares things like deal effective date (from and to date), specific product or product group, customer account, and ship-to or billed-to. Deal win/loss classification data is output at step 530 and used to help model demand in step 330. [0052] In a preferred embodiment, demand for a particular product/segment is estimated using the cleansed datasets discussed above to generate a price elasticity demand model and a win probability model. A demand model is selected which fits well statistically with the historical data. For example, any of the commonly used, externally derived, multivariate, parametric, non-separable algorithms may be used to create the price elasticity and win probability models. The model which best fits the historical data is used.

[0053] The price optimization is performed using the optimized business segment scheme discussed above. In order to decide which algorithm to use or give the best fit, the optimization runs all of them and selects the best one, i.e. the one that has the highest statistical significance vis-a-vis the cleansed data set. All of the algorithms provided by the user may be included to find the best fit given the actual data. The user may use any of the commonly used algorithms discussed above and/or the user may provide preferred models based on the particular dataset in question. [0054] Output from the demand model to the optimization model is a set of price elasticity curves and optionally a set of win probability curves. The instant optimization model selects the demand model which best fits the cleansed data as discussed above. Game theory is used to model competitive behavior based on historical data. The instant optimization combines game theory with dynamic non-linear optimization to give optimized prices. The optimization is done subject to optimization goals and constraints provided by policy managers. For instance, the goal may be to optimize pocket margin given a limited change in product volume or product price.

[0055] FIG. 6 is a flowchart illustrating a process for generating the price elasticity and win probability models. Cleansed order history data and win/loss classification data is provided at step 610. By using the algorithms discussed above, first a win probability model is generated at step 620. Next, a price elasticity model is generated at step 630. The combined models are used to generate a demand model at step 640. The models are output to the price optimization step 650.

III. PRICE OPTIMIZATION MODULE

[0056] FIG. 7 is a high level flowchart further illustrating the optimization step 130 of FIG. 1, in accordance with an embodiment of the present invention. The optimization process will be discussed generally here with a more detailed discussion of the various components to follow.

[0057] Demand model information is provided from step 650. Preferably, in order to effectively optimize prices in a data-poor B2B setting, competitive behavior data is incorporated into the price optimization scheme. Competitive behavior is provided at step 710.

[0058] It is also important to provide optimization goals and constraints in any optimization scheme. The user may decide to optimize for profit, sales or volume maximization. Once the optimization goal is selected, optimization constraints may be set. The user may set the constraints in conformance with the particular business objectives as discussed above.

[0059] The user may choose to constrain the following factors: maximum price increase, maximum price decrease for a business segment (e.g. Product Yearly Revenue Segment A) or intersection of business segments (e.g. Product Yearly Revenue Segment A and Biotech Industry Customers).

[0060] Optimization goals and constraints are provided at step 720. Competitive behavior data along with selected optimization goals and constraints are used to optimize prices at step 730. Previously generated and optimized pricing guidance is provided at step 740. The optimized prices are reconciled with the optimized pricing guidance at step 750. Reconciliation data is provided to the deal negotiation step 140. [0061] FIG. 8 is a more detailed flowchart illustrating the price optimization step 730 of FIG. 7, in accordance with an embodiment of the present invention. Demand model data is provided from the demand modeling step 120 at step 810. Competitive behavior data and optimization goals and constraints are provided at steps 820 and 830, respectively. Prices are optimized to meet the selected goals and constraints at step 840 (this will be discussed in more detail, below). Finally, optimized prices are output for reconciliation at step 850.

[0062] The resulting optimized, reconciled prices are used in deal negotiations. The resulting deals, (i.e. quoted prices with customers) may be provided back as deal history data for iterative optimization. This continuous learning and calibration is done in order to fine tune the instant optimization process with real world data reflecting the actual results of incorporating the optimized prices into the deal negotiation process. [0063] FIG. 9 is a more detailed flowchart further illustrating the optimization step 840 of FIG. 8. In one embodiment, once competitive behavior data and optimization goals and constraints are provided, prices may be optimized for the particular product/segment in question. First, competitive behavior is modeled at step 910 using fictitious play and Nash equilibrium computation. Accurate prediction of competitive behavior is especially important in a B2B environment given the relatively small number of major customers.

[0064] Next, at step 920, a dynamic, non-linear optimization is conducted using an iterative relaxation algorithm. The Nash equilibrium computation is combined with the selected non-linear optimization model to give optimized prices subject to optimization goals and constraints. Optimized prices are output at step 930. [0065] While this invention has been described in terms of several preferred embodiments, there are alterations, permutations, modifications and various substitute equivalents, which fall within the scope of this invention. It should also be noted that there are many alternative ways of implementing the methods and systems of the present invention. It is therefore intended that the following appended claims be interpreted as including all such alterations, permutations, modifications, and various substitute equivalents as fall within the true spirit and scope of the present invention. In addition, the use of subtitles in this application is for clarity only and should not be construed as limiting in any way.

Claims

CLAIMSWhat is claimed is:

1. A method of generating a preferred set of prices in a business to business market environment, said method comprising: selecting at least one product in a selected market segment; providing sales data corresponding to said at least one product; generating a demand model for said at least one product utilizing said sales data; and generating said preferred set of prices for said at least one product utilizing said demand model, wherein said demand model is a price elasticity model.

2. The method of claim 1 further comprising: providing deal history data for at least one deal; and generating a win probability model utilizing said deal history data, wherein said win probability model is used in conjunction with said price elasticity model in generating said preferred set of prices.

3. The method of claim 2 further comprising: providing competitive behavior data; and generating a competitive behavior model utilizing a Nash equilibrium computation, wherein said competitive behavior model is used in generating said preferred set of prices.

4. The method of claim 3 further comprising: providing price guidance data for said at least one product; generating a preferred set of guidance prices utilizing said price guidance data; and reconciling said preferred set of prices with said preferred set of guidance prices to generate a reconciled set of prices.

5. The method of claim 4 further wherein said selected market segment is selected from a preferred set of market segments.

6. The method of claim 5 further comprising: providing a set of goals and constraints wherein said preferred set of prices is generated to meet said goals and constraints.

7. The method of claim 6 further comprising: providing said set of preferred prices to a price control and management system; and generating a quotation utilizing said price control and management system such that said set of preferred prices is incorporated into said quotation.

8. The method of claim 7 further wherein said set of preferred prices may be overridden by user input.

9. The method of claim 8 further comprising: generating a deal utilizing said price control and management system wherein said deal includes a set of final prices.

10. The method of claim 9 wherein said set of final prices are used to calibrate said demand model.

1 1. The method of claim 9 wherein said set of final prices are used to calibrate said win probability model.

12. The method of claim 9 wherein said set of final prices are used to calibrate said competitive behavior model.

13. A computer program product for use in conjunction with a computer system for generating a preferred set of prices in a business to business market environment, said computer program product comprising a computer readable storage medium and a computer program mechanism embedded therein, said computer program product comprising: instructions for selecting at least one product in a selected market segment; instructions for providing sales data corresponding to said at least one product; instructions for generating a demand model for said at least one product utilizing said sales data; and instructions for generating said preferred set of prices for said at least one product utilizing said demand model, wherein said demand model is a price elasticity model.

14. The computer program product of claim 13 further comprising: instructions for providing deal history data for at least one deal; and instructions for generating a win probability model utilizing said deal history data, wherein said win probability model is used in conjunction with said price elasticity model in generating said preferred set of prices.

15. The computer program product of claim 14 further comprising: instructions for providing competitive behavior data; and instructions for generating a competitive behavior model utilizing a Nash equilibrium computation, wherein said competitive behavior model is used in generating said preferred set of prices.

16. The computer program product of claim 15 further comprising: instructions for providing price guidance data for said at least one product; instructions for generating a preferred set of guidance prices utilizing said price guidance data; and instructions for reconciling said preferred set of prices with said preferred set of guidance prices to generate a reconciled set of prices.

17. The computer program product of claim 16 further wherein said selected market segment is selected from a preferred set of market segments.

18. The computer program product of claim 17 further comprising: instructions for providing a set of goals and constraints wherein said preferred set of prices is generated to meet said goals and constraints.

19. The computer program product of claim 18 further comprising: instructions for providing said set of preferred prices to a price control and management system; and instructions for generating a quotation utilizing said price control and management system such that said set of preferred prices is incorporated into said quotation.

20. The computer program product of claim 19 further wherein said set of preferred prices may be overridden by user input.

21. The computer program product of claim 20 further comprising: instructions for generating a deal utilizing said price control and management system wherein said deal includes a set of final prices.

22. The computer program product of claim 21 wherein said set of final prices are used to calibrate said demand model.

23. The computer program product of claim 21 wherein said set of final prices are used to calibrate said win probability model.

24. The computer program product of claim 21 wherein said set of final prices are used to calibrate said competitive behavior model.