GITNUXSOFTWARE ADVICE
Data Science AnalyticsTop 9 Best Decision Optimization Software of 2026
Ranked roundup of Decision Optimization Software for operations teams, with OR-Tools, Gurobi Optimizer, and IBM Decision Optimization compared by features.
How we ranked these tools
Core product claims cross-referenced against official documentation, changelogs, and independent technical reviews.
Analyzed video reviews and hundreds of written evaluations to capture real-world user experiences with each tool.
AI persona simulations modeled how different user types would experience each tool across common use cases and workflows.
Final rankings reviewed and approved by our editorial team with authority to override AI-generated scores based on domain expertise.
Score: Features 40% · Ease 30% · Value 30%
Gitnux may earn a commission through links on this page — this does not influence rankings. Editorial policy
Editor’s top 3 picks
Three quick recommendations before you dive into the full comparison below — each one leads on a different dimension.
OR-Tools
Routing solver with time windows and capacity constraints via the RoutingModel
Built for teams needing high-performance optimization models in code for operations planning.
Gurobi Optimizer
Editor pickGurobi’s branch-and-cut engine for mixed-integer programming with advanced cut control parameters
Built for operations research teams building fast optimization-driven decision systems.
IBM Decision Optimization
Editor pickIBM Decision Optimization Studio for building and validating optimization models before deployment
Built for large operations teams needing enterprise-grade optimization for planning and routing.
Related reading
Comparison Table
This comparison table ranks decision optimization tools by integration depth, focusing on how each platform connects to data pipelines and exposes model building via API. It also compares the data model and schema choices, plus automation and extensibility surfaces such as configuration patterns, provisioning, and sandbox options. Admin and governance controls are reviewed across RBAC and audit log support to show how teams manage throughput and change control across models and solvers.
OR-Tools
open-source optimizationGoogle OR-Tools provides optimization models and solvers for routing, scheduling, assignment, and knapsack problems via a Python and C++ API.
Routing solver with time windows and capacity constraints via the RoutingModel
OR-Tools stands out for turning mathematical optimization and constraint programming into production-ready Python and C++ APIs. It supports mixed-integer linear programming, constraint satisfaction, vehicle routing, scheduling, and routing with time windows and capacities.
Model building, solving, and search control are exposed through consistent interfaces across problem types. Integration is strengthened by a design that fits into existing data pipelines and custom solver workflows.
- +Broad coverage across routing, scheduling, and constraint programming
- +Strong Python and C++ APIs for custom model building and control
- +Good solver orchestration with search strategies and callbacks
- +Efficient formulation options for routing problems and assignment models
- –Math and constraint modeling knowledge is usually required for best results
- –Debugging infeasibility or poor bounds can be time-consuming
- –Advanced tuning of search parameters needs solver and problem insight
Operations research engineers
Solve mixed-integer planning with constraints
Reduced cost and constraint violations
Logistics optimization teams
Route fleets with time windows
Faster delivery planning
Show 2 more scenarios
Manufacturing planning analysts
Plan production scheduling under limits
Higher throughput and feasibility
Analysts express sequencing and machine constraints to generate feasible production plans quickly.
Backend platform developers
Integrate solvers into data pipelines
Automation of decision optimization
Engineers embed OR-Tools solving flows into services that transform inputs and return assignments.
Best for: Teams needing high-performance optimization models in code for operations planning
More related reading
Gurobi Optimizer
commercial MIP solverGurobi Optimizer solves mixed-integer programming, linear programming, quadratic optimization, and related models through a high-performance solver interface.
Gurobi’s branch-and-cut engine for mixed-integer programming with advanced cut control parameters
Gurobi Optimizer provides decision optimization for linear, quadratic, conic, and mixed-integer models using a single solver core with consistent APIs across Python and other interfaces. Model building features support algebraic formulation, constraint and variable management, and solver parameter control for tuning runtime and solution quality. Solution retrieval covers primal and dual information, and it supports export paths for downstream analysis and automated decision workflows.
A key tradeoff is that performance gains often require model formulation discipline and parameter tuning, especially for large mixed-integer instances. It is a strong fit when repeated solves are needed, such as iterating on formulation changes in Python, running what-if scenarios, or embedding optimization inside batch pipelines for scheduling and resource allocation.
- +High-performance solving across LP, QP, QCP, and MILP with strong scalability
- +Flexible modeling via Python API for dense and sparse optimization structures
- +Rich parameter controls for tuning numerics, search, presolve, and cut strategies
- +Good diagnostics including infeasibility handling and solution quality signals
- –Model formulation demands optimization expertise for best results
- –Feature depth can increase configuration complexity for new teams
- –Solver output interpretation can be nontrivial for complex mixed-integer models
Supply chain planning analysts
Capacitated lot sizing with mixed-integer variables
Schedules produced automatically
Energy systems researchers
Conic optimal power flow studies
Viable dispatch candidates found
Show 2 more scenarios
Operations optimization engineers
Vehicle routing with time windows
Feasible routes generated
They build MILP formulations and iterate constraints in code to reduce feasibility issues.
Finance model automation teams
Quadratic portfolio optimization runs
Allocations updated regularly
They run repeated QP solves and export results for risk and allocation reporting.
Best for: Operations research teams building fast optimization-driven decision systems
IBM Decision Optimization
enterprise optimizationIBM Decision Optimization models optimization problems for operations research and delivers optimization solution capabilities for planning and scheduling use cases.
IBM Decision Optimization Studio for building and validating optimization models before deployment
IBM Decision Optimization distinguishes itself with a mature optimization stack that targets routing, scheduling, assignment, and workforce planning use cases. It provides model-driven optimization through constraint programming and mathematical programming solvers, plus decision automation components for deploying optimized plans into business workflows.
The platform supports graph-based problem modeling, scenario evaluation, and repeatable optimization runs for operations teams that need consistent outputs. Integration options connect optimization results to enterprise applications and data pipelines so decisions can be refreshed as inputs change.
- +Strong built-in solvers for constraint programming and mathematical optimization
- +Good support for routing, scheduling, and assignment problem structures
- +Scenario evaluation supports repeatable runs for changing operational data
- –Model setup requires optimization expertise for complex constraints
- –Workflow deployment can be heavier for small teams without DevOps support
- –Debugging infeasibility and performance issues may take optimizer tuning
Logistics planners and dispatch teams
Vehicle routing with time windows
Lower travel time and cost
Manufacturing operations scheduling leads
Job and resource scheduling optimization
Fewer schedule violations
Show 2 more scenarios
Workforce planning analysts
Shift assignment and staffing levels
Improved staffing adherence
Assigns workers to shifts using qualification rules and demand forecasts for consistent coverage.
Operations model builders
Scenario evaluation for decision automation
Faster decision cycles
Runs repeatable what-if experiments and automates deployment of optimized decisions into workflows.
Best for: Large operations teams needing enterprise-grade optimization for planning and routing
AMPL
optimization modelingAMPL is a modeling language and optimization environment that coordinates with solver back ends to solve linear, integer, and nonlinear decision optimization models.
AMPL Modeling Language for algebraic formulation with mixed-integer and nonlinear support
AMPL stands out with a modeling-first workflow for optimization that supports linear, nonlinear, and mixed-integer formulations. It pairs an algebraic modeling language with high-performance solvers for tasks like scheduling, routing, planning, and network design.
AMPL also offers data handling and model development tooling that helps teams maintain large optimization codebases. Deployment can be integrated into applications by calling solve workflows from external environments.
- +Expressive modeling language for linear, nonlinear, and mixed-integer problems
- +Strong solver performance support across diverse optimization categories
- +Good data-model separation for maintainable large optimization projects
- +Supports model generation patterns for complex scheduling and routing
- –Modeling syntax has a learning curve versus GUI-first tools
- –Workflow is less intuitive for decision makers without optimization expertise
- –Requires solver and model tuning for best runtime on hard instances
Best for: Teams building and maintaining advanced optimization models for operations and planning
Pyomo
Python modelingPyomo is an open-source Python optimization modeling framework that generates optimization models and interfaces with multiple solver engines.
Rule-based constraint generation with indexed components using Pyomo sets and Parameters.
Pyomo is a Python-based optimization modeling framework that stands out for representing optimization problems as algebraic expressions built in code. It supports linear, nonlinear, mixed-integer, and stochastic-style model patterns through extensible sets, parameters, variables, and constraints.
Model components can be exported to multiple solver backends, and solutions can be analyzed directly in Python. Pyomo also offers advanced modeling constructs such as rule-based constraint generation and decomposition-ready structures for larger formulations.
- +Python-native algebraic modeling with readable constraint and objective definitions
- +Flexible support for linear, nonlinear, and mixed-integer formulations
- +Extensible modeling blocks and rule-based generation for large indexed models
- +Solver integration enables direct solve loops and programmatic result extraction
- –Modeling requires coding discipline and understanding of optimization formulation
- –Debugging can be difficult when constraints are generated at scale
- –Advanced algorithmic features depend on external solver or add-on tooling
Best for: Teams building custom optimization models in Python with solver backends.
COIN-OR CBC
open-source MIP solverCBC is an open-source mixed-integer programming solver used through COIN-OR toolchains for branch-and-cut optimization problems.
Branch-and-cut with configurable cutting plane generation
COIN-OR CBC is a branch-and-cut mixed-integer programming solver with strong performance on pure and mixed integer models. It reads standard MIP formulations through common interchange formats and is commonly embedded in custom optimization pipelines.
CBC supports advanced presolve, cutting planes, heuristics, and parallel search settings that influence solution quality and runtime. This makes it a solver-focused decision optimization option rather than a visual or workflow automation suite.
- +High-quality MIP solving with branch-and-cut and presolve.
- +Supports rich solver controls for cuts, heuristics, and search.
- +Works well embedded into optimization pipelines via standard model formats.
- –Requires MIP formulation skills to achieve strong results.
- –No built-in visual modeling interface for business-friendly workflows.
- –Large-scale models can be sensitive to parameter tuning and data quality.
Best for: Teams needing controllable MIP solving inside custom optimization applications
FICO Xpress
commercial optimization solverFICO Xpress delivers high-performance optimization solvers for mixed-integer programming and linear optimization with modeling interfaces.
Advanced presolve, cutting planes, and parallel solving for large mixed-integer models
FICO Xpress stands out for production-grade optimization modeling that targets complex decision problems with strong solver performance. It provides a modeling layer for mixed-integer programming, linear and quadratic optimization, and optimization over networks.
It also supports advanced presolve, cutting planes, and parallel solving to improve time-to-solution on large instances. Built for optimization teams, it emphasizes model formulation accuracy and solver-grade behavior rather than business-user workflows.
- +Strong mixed-integer and quadratic optimization with industrial solver performance
- +Flexible modeling support for linear, quadratic, and network-style decision problems
- +Parallel solving, presolve, and cutting strategies improve solution times
- –Modeling complexity can slow time-to-first-result for non-optimization specialists
- –Decision workflows require integration outside the solver for business execution
- –Performance depends heavily on formulation quality and solver parameter choices
Best for: Optimization teams building custom decision models for scheduling, routing, and planning
Decision Optimization in Azure AI
cloud optimizationMicrosoft Azure AI decision optimization tooling supports optimization modeling workflows and integrates with Azure services for operational planning decisions.
Scheduling and planning optimization with constraint programming and mixed-integer models
Decision Optimization in Azure AI stands out by combining optimization modeling with Azure integration for data pipelines and enterprise governance. It supports task planning and scheduling using integer programming and constraint programming, plus scenario evaluation for operational decisions.
The service fits into an Azure-first workflow by consuming structured data, generating decision variables, and returning optimized solutions for downstream systems. The platform is most effective when decision logic can be expressed in mathematical constraints rather than in natural language.
- +Constraint-based modeling for scheduling and planning with strong optimization primitives
- +Azure integration enables connecting data sources and production systems
- +Scenario comparison supports evaluating tradeoffs across operational assumptions
- +Decision outputs are structured for programmatic consumption
- –Modeling complex domains can require significant optimization expertise
- –Limited coverage for unstructured inputs beyond structured data preparation
- –Debugging infeasibility and performance tuning can be time consuming
- –Iterative experimentation often needs code changes to refine constraints
Best for: Teams optimizing schedules, routing, and resource allocation from structured data
Optuna
optimization automationOptuna performs automated hyperparameter optimization and can drive decision optimization workflows using objective functions and search strategies.
Pruners with intermediate value reporting for early stopping of trials
Optuna stands out for turning hyperparameter and decision tuning into an experiment-driven optimization loop with flexible search spaces. It provides Bayesian optimization style samplers, pruning for early stopping, and study-level abstractions for repeated trials. It also supports multi-objective optimization and integrates common ML stacks through Python APIs.
- +Flexible search spaces with samplers for Bayesian and tree-based exploration
- +Pruning reduces compute by stopping unpromising trials early
- +First-class multi-objective optimization with Pareto front handling
- +Study persistence supports reproducible runs and resumable experiments
- –Python-centric API can slow adoption for non-Python decision teams
- –Complex callback and pruner logic increases implementation difficulty
- –High performance depends on correct objective design and training loops
Best for: ML and operations teams optimizing models and decision variables via Python experiments
Conclusion
After evaluating 9 data science analytics, OR-Tools stands out as our overall top pick — it scored highest across our combined criteria of features, ease of use, and value, which is why it sits at #1 in the rankings above.
Use the comparison table and detailed reviews above to validate the fit against your own requirements before committing to a tool.
How to Choose the Right Decision Optimization Software
This buyer’s guide covers OR-Tools, Gurobi Optimizer, IBM Decision Optimization, AMPL, Pyomo, COIN-OR CBC, FICO Xpress, Decision Optimization in Azure AI, and Optuna.
The focus is integration depth, data model design, automation and API surface, and admin and governance controls. Each section maps concrete evaluation checkpoints to named tools used for routing, scheduling, assignment, and MIP solving.
Decision Optimization software for routing, scheduling, and MIP-style decision plans
Decision Optimization software turns constraints, objective functions, and decision variables into solver-ready optimization models. It targets planning tasks like routing with time windows and capacities, scheduling, assignment, and workforce planning using constraint programming and mathematical programming solvers.
Tools like OR-Tools provide production-oriented Python and C++ APIs for routing and assignment models, while Gurobi Optimizer exposes a consistent solver interface for linear, quadratic, conic, and mixed-integer models in code.
Evaluation criteria that map to integration, schema control, and automated execution
Integration depth decides whether optimization runs become part of existing pipelines or remain a detached modeling workflow. Data model clarity decides whether constraints stay maintainable across schema changes in operational inputs.
Automation and API surface decide whether decisions refresh automatically inside batch jobs or service calls. Admin and governance controls decide who can run, modify, and audit optimization configurations across teams and environments.
Solver API coverage across routing, scheduling, and mixed-integer formulations
OR-Tools concentrates on routing solver capability through RoutingModel with time windows and capacity constraints, which directly matches operations planning needs. Gurobi Optimizer spans LP, QP, QCP, and MILP through one solver core, which matters when optimization models must support more than one problem family.
Data model separation and maintainable model structure
AMPL emphasizes data-model separation so large optimization codebases remain maintainable when datasets or formulations evolve. Pyomo uses Python-native algebraic expressions with indexed sets and parameters, which keeps constraint and objective definitions tied to a structured model representation.
Automation surface and programmatic solve loops
Gurobi Optimizer supports repeated solves in Python for what-if scenarios and embedding optimization inside batch pipelines, which fits iterative planning workflows. OR-Tools exposes consistent model building, solving, and search control interfaces across problem types, which helps when routing and scheduling logic must be generated and executed by application code.
Extensibility for solver behavior through tuning and callbacks
Gurobi Optimizer provides rich parameter controls for presolve, search, and cut strategies, which supports repeatable runtime tuning on large mixed-integer instances. OR-Tools exposes solver orchestration with search strategies and callbacks, which supports custom search control for difficult routing and assignment problems.
Governance-ready model building and validation workflow
IBM Decision Optimization includes IBM Decision Optimization Studio for building and validating optimization models before deployment, which creates a controlled path from model authoring to execution. Decision Optimization in Azure AI fits Azure-first operational governance by consuming structured data and returning structured decision outputs for downstream systems.
Branch-and-cut control and presolve for throughput on hard instances
COIN-OR CBC provides configurable branch-and-cut settings through advanced presolve, cutting planes, and parallel search parameters, which fits custom pipelines that need solver control. FICO Xpress emphasizes advanced presolve, cutting planes, and parallel solving for time-to-solution on large mixed-integer models.
Choose by mapping your integration and model lifecycle to the tool’s execution surface
The selection starts with how optimization work must be produced and executed. Some tools focus on in-code modeling and solver orchestration, while others add a studio-like authoring and deployment workflow.
The next step checks how decisions must refresh and how much control the platform exposes for solver behavior tuning. The guide then matches those needs to concrete capabilities like OR-Tools RoutingModel time windows or Gurobi branch-and-cut cut control parameters.
Define the decision types and constraint shapes that must be solved
Routing with time windows and capacity constraints maps directly to OR-Tools RoutingModel and its routing-focused solver API. Mixed-integer models that require advanced cut control and diagnostics map to Gurobi Optimizer and its branch-and-cut engine.
Pick the data model approach that matches the way operational data changes
If model development needs strict separation between algebraic formulation and changing datasets, AMPL’s data-model separation helps maintain large codebases. If the organization needs a Python-first representation where constraints and objectives are defined as algebraic expressions, Pyomo’s sets, parameters, rule-based constraint generation, and solver back ends fit that workflow.
Evaluate the automation and API surface needed for scheduled refresh and batch or service execution
For repeated solves inside batch pipelines and iterative what-if planning in code, Gurobi Optimizer’s Python API is designed for programmatic solve loops and solution retrieval. For routing and scheduling logic executed inside existing application workflows, OR-Tools provides consistent interfaces for model building, solving, and search control.
Check governance controls for model authoring, validation, and controlled deployment
When model validation must happen before deployment, IBM Decision Optimization Studio provides a studio workflow for building and validating optimization models. When the decision system must integrate into an Azure-first environment with structured input consumption and structured output return, Decision Optimization in Azure AI aligns with that operational pattern.
Confirm extensibility for solver performance tuning and explainability signals
If tuning requires runtime control over presolve, cuts, and search strategies, Gurobi Optimizer offers detailed parameter controls and diagnostics for infeasibility and solution quality signals. If the pipeline must control branch-and-cut mechanics directly through a solver-focused approach, COIN-OR CBC’s cutting plane generation controls and presolve heuristics provide that control
Add hyperparameter search only when the goal is decision-tuning, not direct constraint solving
Optuna fits decision tuning when optimization problems depend on hyperparameters or when multiple trials must be coordinated through samplers and pruning. For direct operational constraint solving like routing and scheduling, prioritize OR-Tools, Gurobi Optimizer, IBM Decision Optimization, or AMPL over Optuna’s experiment-driven search loop.
Who benefits from decision optimization tooling in real operations workflows
Decision optimization tools concentrate value where constraints, routing, and scheduling must be recalculated from structured inputs under time and capacity limits. The right tool depends on whether the organization needs an in-code API-first workflow or a governed authoring and deployment workflow.
The tool fit also depends on whether the optimization work is primarily constraint programming, mathematical programming, or experiment-driven tuning.
Operations research teams embedding optimization into systems with repeated solves
Gurobi Optimizer fits teams that build optimization-driven decision systems because it provides high-performance solving across LP, QP, QCP, and MILP with extensive solver parameter controls for repeated what-if iterations.
Planning teams that need production routing with time windows and capacities in code
OR-Tools fits teams that need high-performance optimization models in code because RoutingModel directly supports time windows and capacity constraints with a Python and C++ API built for application workflows.
Enterprise operations groups that need a validated model workflow and repeatable runs
IBM Decision Optimization fits large operations teams because IBM Decision Optimization Studio supports building and validating models before deployment and scenario evaluation supports repeatable optimization runs for changing operational data.
Teams maintaining large optimization formulations with clear model-data separation
AMPL fits teams building and maintaining advanced optimization models because the modeling language supports mixed-integer and nonlinear formulations with maintainable data-model separation for large codebases.
Azure-first teams that optimize schedules and planning from structured data feeds
Decision Optimization in Azure AI fits teams that must integrate optimization into Azure pipelines because it consumes structured data, generates decision variables, returns optimized solutions for downstream systems, and supports scenario comparison.
Failure modes that show up when integration, model design, and governance are mismatched
Many decision optimization failures come from model and integration choices, not solver runtime alone. When the tool and the operational lifecycle do not match, teams spend time on avoidable formulation rework and debugging.
The recurring pitfalls below map to concrete constraints described across OR-Tools, Gurobi Optimizer, IBM Decision Optimization, and the modeling frameworks.
Treating solver APIs as plug-and-play when formulation discipline is required
Gurobi Optimizer performs best when model formulation is disciplined and parameter tuning is applied for large mixed-integer instances. OR-Tools also needs math and constraint modeling knowledge to achieve best results and to debug infeasibility or poor bounds efficiently.
Picking an overly solver-focused tool when a governed authoring to deployment workflow is required
COIN-OR CBC and FICO Xpress concentrate on solver behavior like branch-and-cut controls, presolve, and parallel solving, which often requires extra integration work for business execution. IBM Decision Optimization’s Studio workflow and scenario evaluation reduce friction when model validation and repeatable runs must be managed across an operations organization.
Overcomplicating model maintenance by ignoring model-data separation or indexed constraint generation patterns
AMPL prevents tangled formulations by separating data and model development, which helps when datasets evolve across planning cycles. Pyomo’s rule-based constraint generation with indexed components using sets and Parameters is the maintainable pattern when constraints must expand across large indexed structures.
Using experiment search tools for direct operational constraint solving
Optuna targets hyperparameter and decision tuning via samplers and pruning and it runs trials through an experiment loop. Direct routing with time windows and capacities and direct scheduling should be implemented with OR-Tools RoutingModel or IBM Decision Optimization constraints rather than wrapping constraint solving inside Optuna trials.
Assuming infeasibility debugging and performance tuning are automatic
IBM Decision Optimization and Decision Optimization in Azure AI still require optimization expertise for complex domain modeling and tuning, especially when debugging infeasibility and performance issues. Even with high-performance solvers like Gurobi Optimizer, interpreting solver output and diagnosing complex mixed-integer behavior can take time without formulation discipline.
How We Evaluated and Ranked Decision Optimization tools
We evaluated OR-Tools, Gurobi Optimizer, IBM Decision Optimization, AMPL, Pyomo, COIN-OR CBC, FICO Xpress, Decision Optimization in Azure AI, and Optuna using criteria centered on features, ease of use, and value. Features carry the most weight because integration depth, data model control, and automation and API surface determine whether optimization runs become dependable components in real workflows. Ease of use and value each matter enough to influence the rank when multiple tools support similar modeling patterns.
OR-Tools earns the top placement because it pairs high-performance routing through RoutingModel with time windows and capacity constraints with consistent Python and C++ APIs for model building, solving, and search control. That combination lifted the features score and ease-of-use fit for teams that need routing and scheduling optimization expressed directly in application code.
Frequently Asked Questions About Decision Optimization Software
Which tools are best when optimization must run inside a Python or code-first workflow?
How do OR-Tools and Gurobi differ for mixed-integer and runtime tuning in iterative what-if runs?
Which option is most suitable for enterprise routing and scheduling with model validation before deployment?
What is the practical difference between using AMPL versus building models directly in Pyomo for large model codebases?
How do COIN-OR CBC and FICO Xpress compare for embedding a solver into a custom optimization pipeline?
Which tools provide clear pathways for exporting solutions to downstream automation and analysis?
How do integration and API needs differ between IBM Decision Optimization and Decision Optimization in Azure AI?
What should teams consider for SSO, RBAC, and audit logging when selecting a platform?
How should data migration be handled when moving existing optimization formulations between tools?
When does extensibility matter more than raw solver speed, and which tools support that best?
Tools reviewed
Primary sources checked during evaluation.
Referenced in the comparison table and product reviews above.
Keep exploring
Comparing two specific tools?
Software Alternatives
See head-to-head software comparisons with feature breakdowns, pricing, and our recommendation for each use case.
Explore software alternatives→In this category
Data Science Analytics alternatives
See side-by-side comparisons of data science analytics tools and pick the right one for your stack.
Compare data science analytics tools→FOR SOFTWARE VENDORS
Not on this list? Let’s fix that.
Our best-of pages are how many teams discover and compare tools in this space. If you think your product belongs in this lineup, we’d like to hear from you—we’ll walk you through fit and what an editorial entry looks like.
Apply for a ListingWHAT THIS INCLUDES
Where buyers compare
Readers come to these pages to shortlist software—your product shows up in that moment, not in a random sidebar.
Editorial write-up
We describe your product in our own words and check the facts before anything goes live.
On-page brand presence
You appear in the roundup the same way as other tools we cover: name, positioning, and a clear next step for readers who want to learn more.
Kept up to date
We refresh lists on a regular rhythm so the category page stays useful as products and pricing change.