Skip to main content

One Model — Portfolio Strategy & Profitability Optimization

Data classification: GM_Confidential
Domain: Finance Data Engineering — Portfolio Strategy, Profitability & Optimization
Primary business capability: Vehicle portfolio scenario planning, profitability analysis, and constraint-aware optimization
Also known as: PST/PBT Integration, One Optimizer, Portfolio Strategy Tool, Profitability Scenario Tool
ASMS: a230337
Catalog schema: finance_prod.gold_integrated_portfolio_optimization_gbl


Overview

One Model is GM's integrated platform for deciding how the vehicle portfolio should evolve over the next ten years. It brings the Profitability Scenario Tool (PST) and the Portfolio Balancing Tool (PBT) into one application so Finance, Product Portfolio, and Portfolio Planning teams can plan, compare, and optimize the portfolio in a single place. One Model unifies portfolio planning attributes with financial forecasts — variable profit, volume, capital, and engineering spend — into one authoritative view of GM's programs.

The platform is organized around three working surfaces: an Analyzer for exploring the current portfolio, a Scenario Builder for creating and editing portfolio scenarios, and an Optimizer that recommends the best portfolio under defined business constraints. It helps leaders answer questions such as:

  • Which programs should we launch, sustain, retime, or exit to maximize sustainable profitable growth?
  • What is the financial impact (variable profit, revenue, ROI) of a portfolio change across the next decade?
  • How do capital, engineering, and plant-capacity limits constrain what the portfolio can actually deliver?
  • How do scenarios compare, and which one is the most defensible decision?

Key benefits

One Model replaces fragmented, spreadsheet-driven portfolio analysis with a standardized data foundation and a repeatable scenario-and-optimization workflow. That matters because portfolio decisions depend on reconciling many moving parts — program timing, volumes, prices, profitability, capital and engineering budgets, plant capacity, and regulatory effects — that previously lived in separate models.

For business users, it provides one consistent place to build scenarios, run optimizations, and produce executive-ready financial comparisons. For data engineers, it provides a certified gold-layer foundation that integrates portfolio planning and finance forecast data into a consistent schema.

ThemeTypical questions
Portfolio strategyWhat is the best mix of programs to launch, sustain, retime, substitute, or exit?
Profitability tradeoffsHow does a change affect variable profit, net sales and revenue, EBIT, and contribution margin?
Constraint-aware optimizationWhat is the optimal portfolio given capital, engineering, plant-capacity, and cadence limits?
Scenario decisioningHow do baseline and alternative scenarios compare, and what changed and why?
Resource governanceHow are capital (ME/VT/PC) and engineering (PD/S&S/ME) dollars allocated by program and year?

Who uses it

Team / personaHow they use One Model
Product PortfolioEvaluate launch/sustain/exit decisions and portfolio mix across segments and propulsion types
Portfolio PlanningBuild and maintain program scenarios, timing, and volume assumptions
Finance / Finance 360Assess profitability, capital, and engineering impacts; tie results to financial forecasts
Finance ITApplication ownership, integration, and operational support
Finance Data EngineeringBuild and operate the data pipelines and gold serving layer that feed the tool

Core business scope

One Model integrates portfolio planning and finance forecast data, then supports analysis and optimization across the program lifecycle. Work is organized around three surfaces:

SurfaceWhat it does
AnalyzerExplore the integrated baseline portfolio — financials, volumes, capital, engineering, and timing across programs
Scenario BuilderCreate and edit scenarios: add master programs, change volumes, retime, substitute, and edit program details
OptimizerRecommend the best portfolio under constraints using objective-driven optimization

The optimizer manages five portfolio decisions — program retiming, deletion, addition, substitution, and production/wholesale volume — while respecting business constraints:

  • Plant production capacity
  • Yearly capital and engineering resource cost
  • Program cadence

Optimization targets one or more objectives (variable profit, ROI, volume, revenue), with multi-objective optimization and program-level capital/engineering allocation expanded in the R2 release.

Source systems by business subject

One Model homogenizes portfolio, volume, finance, capacity, and policy inputs into a consistent schema. Major source areas include:

Subject areaKey sources
Financial forecastOneStream (current finance source); eGPPS / GPPS and Hyperion legacy (migrating via the HYP2OS effort)
Long-term plan & volumeLong-Term Plan / GLTP, Moonshot volume forecasts
Portfolio & program structureProgram and master-program mappings, lead/donor-predecessor lineage, BCT
Plant capacityMoonshot plant capacity (annual and monthly, installed and assumed straight-time/overtime)
Diversions & elasticityBP25 product diversions, GM product diversion matrix, market and product-library price elasticities
Policy & regulatoryIRA tax credits and GHG regulatory cost estimates
User-maintained inputsBusiness-uploaded files for capital/engineering curves, price forecasts, master-program mappings, and saved scenarios

As part of the OneStream transition, One Model's ETL is being repointed from the legacy Hyperion eGPPS source to OneStream-based finance sources to preserve impacted Finance metrics.

Business measures and rules

One Model carries a full program-level P&L alongside resource and policy measures so scenarios and optimizations are financially complete:

Measure groupExamples
Profit & revenueVariable Profit (VP), Net Sales and Revenue, Contribution Margin, Income Before Interest and Tax (EBIT)
CostMaterial, Total Variable Manufacturing, Logistics, Structural Cost, Contribution Cost
ResourcesCapital spend by feature (ME — Machines & Equipment, VT — Vendor Tooling, PC — Program Contingency); Engineering spend (PD, S&S, ME), including labor hours
Volume & priceWholesale volume, average transaction price forecasts, price elasticities
PolicyIRA credits, GHG regulatory costs

Common conventions a new user should know:

  • Programs are analyzed individually and rolled up to master programs for portfolio-level views.
  • Timing is anchored to lifecycle milestones such as SORP (Start of Regular Production) and EOP (End of Production).
  • Capital and engineering totals are spread to monthly periods using allocation curves before constraints are applied.
  • Diversions model how volume shifts between programs when one is added, delayed, or removed.

Reporting, interfaces, and consumers

The primary interface is the One Model application itself (Analyzer, Scenario Builder, Optimizer), where users build scenarios, run optimizations, and compare results. Published scenarios and optimized outputs are persisted in the gold serving layer so they can be analyzed, compared, and consumed downstream by Finance, Finance 360, and portfolio dashboards. Data-quality and tie-out visibility for end users is delivered through supporting dashboards.

EnvironmentURL
Devhttps://a230337-t2-musea2-app-ui-oneoptimizer-dev.azurewebsites.net/
Testhttps://a230337-t2-musea2-app-ui-oneoptimizer-test.azurewebsites.net/
Prodhttps://a230337-p1-musea2-app-ui-oneoptimizer.azurewebsites.net/ (onemodel.gm.com)

Security and access model

  • Data is classified GM_Confidential and is intended for Finance, Product Portfolio, and Portfolio Planning audiences.
  • Access to the One Model application (GUI) is managed via myAccess under ASMS a230337. Three roles are available:
myAccess roleAAD groupEnvironments
230337-OneModel-Usersa230337-onemodel-usersProduction
230337-OneModel-Testersa230337-onemodel-testersTest and Production
230337-OneModel-Developersa230337-onemodel-developersDev, Test, and Production

Where the data lives

LayerLocationPurpose
Bronzefinance_prod.bronze_integrated_portfolio_optimization_user_files_gblRaw business-uploaded inputs: capital/engineering curves, price forecasts, master-program mappings, and saved scenarios
Silverfinance_prod.silver_integrated_portfolio_optimization_gblCleansed and integrated source data from OneStream, GLTP, Moonshot, BCT, and other upstream systems; conformed schema before gold promotion
Goldfinance_prod.gold_integrated_portfolio_optimization_gblCertified, business-ready portfolio, financial, capacity, and optimizer-output tables; primary serving layer for analytics and the One Model application

Start with the integrated combined-source tables for portfolio analysis, then move to optimizer-output and resource tables for deeper work.

Table / viewUse it for
primary_combined_source_yearlyAnnual integrated volume + financial forecast by product/plant/scenario — the primary yearly portfolio view
primary_combined_source_monthlyMonthly time-series volume and P&L for detailed analysis and optimization input
combined_monthly_portfolioIntegrated monthly volume, financial, and cost metrics across all published scenarios
optimized_portfolio_outputOptimizer's recommended production plan (volumes, VP, timing) by program/plant/propulsion/month
optimized_portfolio_metadataRun audit trail: session, user, baseline scenario, and constraint warnings
published_scenario_detailMaster list of published scenarios available for analysis and comparison
moonshot_monthly_plant_capacityPlant capacity constraints used to validate production volumes
finance_capital_spend_monthly / finance_engineering_spend_monthlyBaseline monthly capital and engineering resource profiles used as optimizer inputs

Practical guidance

  • Use primary_combined_source_yearly (or primary_combined_source_yearly_vw) for most portfolio-level analysis; drop to primary_combined_source_monthly only when monthly granularity matters.
  • Reference scenarios by their scenario_id from published_scenario_detail when loading baselines or comparing alternatives.
  • Use the optimized_* output tables and optimized_portfolio_metadata to interpret optimizer recommendations and trace them back to a run.
  • Use moonshot_* capacity and finance_*_spend_monthly tables when investigating why the optimizer respected or flagged a constraint.

Getting started path

  1. Read this overview and the One Model (PST/PBT Integration) Confluence page for charter, workstreams, and release context.
  2. Request access to the gold serving schema via the Atlan catalog entry for finance_prod.gold_integrated_portfolio_optimization_gbl (enforced by Immuta).
  3. Explore the portfolio in the Analyzer, then build a simple scenario in the Scenario Builder before running the Optimizer.
  4. For data work, start with primary_combined_source_yearly in finance_prod.gold_integrated_portfolio_optimization_gbl.
  5. Review optimizer outputs (optimized_portfolio_output, optimized_portfolio_metadata) to understand recommended changes and constraints.
  6. Move to resource, capacity, and diversion tables only when investigating constraint behavior, source logic, or reconciliation.

Support and ownership

RoleContact
Engineering DRIJie Du
Program DRIBrad Cromwell
Optimizer DRIPeling Wu-Smith
Analyzer DRIMustafa Mezaal
Data team lead / DRIWayne Smiles / Dave Olds
Finance ITJohn Brozanski, Rachel Curry

Collaboration: Slack #one-model-platform-development.

ResourceURL
One Model (PST/PBT Integration) Overview (Confluence)One Model Overview
PST-PBT Integration SharePointPST-PBT Integration
Jira project (SDEPPD)SDEPPD
Gold serving schema (Atlan)finance_prod.gold_integrated_portfolio_optimization_gbl
PST application repo (GitHub)GeneralMotors-IT/360Fin_230337_pst