Product Configuration Explained: A Guide to Virtual Tabulation, Rules, and Constraints

Designing, selling, manufacturing and servicing complex products requires a robust product configuration solution.

And choosing the right product configuration technology can be the difference between achieving digital transformation goals or spending years trying to make the wrong technology work.

How do you find the best solution for your product portfolio and processes?

This blog post will provide definitions of important concepts, such as rules and constraints, compare product configuration solutions based on rules- and constraints-based algorithms, and provide advice on modeling by highlighting how modern modeling solutions like Configit Ace® with Virtual Tabulation® technology supports both rules and constraints while ensuring a fast and responsive performance.

The product configuration challenge

Today, both B2C and B2B customers expect the ability to personalize a product themselves.

Providing a self-service product configuration application that can guide customers to the right solution is an obvious solution. But for manufacturers of complex products with many options and interdependencies, this can be a challenge. As products become more complex, the user experience can suffer from long wait times after each selection and the potential for errors or “dead ends” where there is no valid solution.

Different technologies and approaches have been proposed, but which is the most appropriate for your needs?

Before debating the best methods, approaches and technologies for high performance product configurators, it is best to first clear up confusion around concepts such as ‘rules’ and ‘constraints,’ as well as modeling concepts and configuration engine technology approaches.

The ‘rules versus constraints’ debate

First, some definitions:

• Rules are explicit logical expressions or conditional statements used to guide configuration behavior.
• Constraints are declarative logical statements that define validity conditions for a configuration. Unlike rules, constraints do not prescribe actions or evaluation order.

Some vendors promote a constraints-focused approach, pointing to the potential explosion in rules that can ensue, while others advocate a hybrid approach, combining the two. Which is correct? The answer is both!

Both interpretations can be true because rule-based approaches have evolved over time.

The following table provides a definition and comparison of both traditional and modern declarative rule-based approaches with constraints-based approaches to product modeling and configuration.

While it is true that traditional rule-based approaches were more rigid, requiring great care when defining rules and understanding the order in which rules are executed, the combination of constraints- and rules-based approaches in modern product configurators provides a great deal of flexibility.

A third option: Attribute-based configuration

In addition to rules- and constraint-based approaches is ‘attribute-based.’

Attribute-based configuration (ABC) defines products using a set of characteristics (attributes) and values that describe the product space. The valid combinations of attributes are often specified via dependencies or rules.

This is a common approach in ERP-integrated systems like SAP LO-VC, and in many early CPQ systems.

• Products are described not by discrete parts or SKUs, but by features with attribute values.
• Dependencies between attributes are modeled using rules or constraints.
• For example:
  • Car.color = ‘Red’
  • Car.engine_power = ‘150HP’
  • Dependency: IF engine_power = ‘200HP’ THEN transmission = ‘Automatic’

Attribute-based models tend to use either rule-based logic (traditional or declarative) or constraint-based solvers to evaluate valid configurations based on attribute values.

Virtual Tabulation® (VT™) is a type of symbolic AI and is the only configuration technology that compiles every possible product configuration within your entire product portfolio. Because your product models can be attribute-, rule-, or constraint-based, VT provides the most flexible modeling experience in the industry.

VT™ then takes these millions of product combinations and compresses them into a compact, portable file, creating a complete solution space of your full product portfolio in one small, shareable file, guaranteeing that anything you sell is buildable and profitable. Sales, engineering, manufacturing and service are all working from the same information, eliminating errors and duplication of work.

The pre-compilation advantage

The pre-compilation feature of Virtual Tabulation provides several advantages that cannot be achieved using any other configuration solution.

For example, for a product model to compile, there needs to be consistency and completeness in the product model itself. Any inconsistencies, gaps or other errors in logic will be detected by the compiler and reported as errors to the modeler with advice on how to resolve the issue.

Not only does this detect the vast majority of issues before they reach the product configurator, but it also enables modelers to react quickly and fix the issues. This could save weeks of debugging and optimization.

Pre-compilation also enables individual modelers to work on their version or portion of a product model independently without affecting the overall product model. Modelers can pre-compile their work and ensure that it is correct before committing it to the overall product model for release to the product configurator. This enables many modelers to work on a product model at the same time.

Lastly, pre-compilation provides the ability to perform deep analysis of what is being offered. Because the VT file provides a full overview of what is currently being offered to customers, analysis can be performed to:

• Compare against what is being sold to optimize the product portfolio
• Compare against previous versions of the product model to see what’s changed
• Understand the impact of changes and ensure that there are no unintended consequences
• Understand where features and modules are being used across the product portfolio

These unique analytical capabilities are not possible using solver-based approaches.

Because Virtual Tabulation supports attributes, constraint and rule logic, there is no trade-off from a modeling and logic perspective in choosing a Virtual Tabulation approach. The real distinction is in the logic execution method; whether it is an interpreter-based approach using attribute-, constraint- or rule-based solver, or pre-compilation methods like Virtual Tabulation.

Solvers are perfectly fine for most products, and even moderately complex products. Several manufacturers have successfully created product configurators for complex products using solvers. However, the cost of validation, maintenance, continuous optimization and debugging, as well the time to introduce new capabilities in complex products, can be an issue.

While there is an extra cost in pre-compiling product models, it can be managed by making several changes to product models each day without issue.

The advantages of Virtual Tabulation are being able to:

• Detect inconsistencies, gaps and other errors during compilation
• Work independently and collaboratively on product models
• Enable lightning-fast runtime performance
• Handle multiple conflicts with rich guidance

This makes Virtual Tabulation ideal for highly complex products, but equally applicable to other products as it incorporates all the advantages of other approaches with very few downsides.

Daniel Joseph Barry is VP of Product Marketing at Configit. Dan Joe has more than 30 years of experience in engineering, sales, marketing, product management and strategy roles within IT and telecom companies.