Logic-Based Configurator on the Web,

¶ … Logic-Based Configurator on the Web, authored by McGuiness, Isbel, Parker, Patel-Schneider, Resnick and Welty (1998). The authors are researchers at AT&T Bell Labs, MIT, and Vassar College, and together the authors posit that a more constraint-based approach to defining online configuration logic through an abbreviated set of constraint linkages that comprise collectively an engine make more sense than having a vast library of rules that guide the definition of a given product under consideration by a customer using web-enabled tools. The authors, while not specifically discussing the variations in rules-based vs. constraint-based configuration logic, point towards the growth of the latter mainly as a result of the need for simplification of complex configuration as is evidenced by the authors' research at AT&T Bell Labs experimenting with telecommunications products configuration logic. In product configuration arena the development and execution of telecommunications configurations are orders of magnitude more complex than many other products' configuration logic, trucks and buses included. The only complexity that approaches telecommunications is the definition of aircraft assemblies, and many companies rely not on configuration software engines but instead use Product Lifecycle Management (PLM) in conjunction with Manufacturing Execution Systems (MES) all layered into production and planning systems to accomplish the same result at a much higher level of abstraction necessary for aircraft production.

The authors deserve credit for their foresightful analysis of the migration from rules-based to constraint-based web-based configuration and the need for included large nested object structures corresponding to highly structured and complex, interconnected configurations. Today this is called nested configuration logic, and is critical for the development of constraint-based logic in configuration engines now increasingly being used across the Internet. The authors also accurately call out of the need for more extensive reasoning components, support for explanation generation, and the critical need for knowledge management tools.

Critique of the Article

What is missing from the article are several key points, the two most glaring are the impact product configurators have on the financial performance of companies over time. What has since been proven from the growth of constraint-based configurators is the ability to tie back the financial performance of a company directly to the influences of their configuration strategy on customer satisfaction. The following table, derived by Columbus (2003) during his research of the impact of configurators on the global performance of companies adopting constraint-based configuration engines, shows that a company's entire value chain can be positively influenced by the growth of configurators as part of their online selling efforts. The authors, working in MIT, AT&T Bell Labs, and Vassar, concentrated only on telecommunications configuration challenges, yet if they would have opened up the scope of their research they may have found the correlation of configurator performance on financial results for a company.

Areas of Measurement

Baseline: What to Measure

Example of Benefits

Company-specific

Project costs and expenses

Use as a baseline for defining ROI

Number of orders per year

Determine configuration's impact on inventory turns

Current inventory and costs

Inventory turn savings

Customer Data

Lifetime cost per customer; avg. deal size by customer

Sales

Order cycle time

Order cycle times reduction of 65% or more recorded with mftrs contacted

Cost of Sales

Days Sales Outstanding reduction from 60 to 29 days on average

Cross-sell and up-sell revenue

Increase of 33% on aggregate

Average sales price per order

Increase from 9% to 26%

Quote and Order

Average costs to complete an order

95% reduction in cost per order

Special Pricing Requests

Over 100% ROI on automating Special Pricing Requests

Bad or incomplete orders

Incomplete order reductions of 20%

Customer Service

Number of customer complaints

98% reduction in cost of simple requests

Revenue lost to churn

60% when cross-selling is used with quote-to-order

Number of calls on order status

Median level of 500 per week to 70

Warranty and Returns

Reduction in warranty cost on customized products

10% reduction at a minimum

Labor cost reductions

Decrease order re-work from 15% to 2%

Future Research

What is most lacking in the article however is a progression of where their research could be applied throughout the complexity levels of mass customization. The graphic below based on analysis from Columbus (2003) shows the varying levels of mass customization with color gradations applied to the definition of guided selling, quote-to-order or sales configuration, and product configuration. The message of this chart is that constraint-based engines can scale from the most simplistic (Assemble to Order) functions to the most complex (Engineer-to-Order) through the more exhaustive and extensive use of constraint logic in conjunction with an order state engine to track progress of nested configurations.