Skip to main content

CAtobuildCA Decision Guide: Self-Optimizing Roadmap

This guide explains how to leverage the CAtobuildCA feature to drive a self-optimizing development roadmap, using insights from your forecasting decisions and seamless integration with Linear.

1. Overview: What is CAtobuildCA?​

CAtobuildCA (pronounced "Cat-to-build-CA") is ChainAlign's intelligent system for automatically identifying and prioritizing high-impact feature enhancements based on the predictive power of your data. It closes the loop between forecasting outcomes and your development backlog by:

  • Analyzing Feature Importance: Automatically determines which external data factors (covariates) are most influential in your forecasts.
  • Generating Actionable Insights: Translates these insights into human-readable summaries.
  • Creating Linear Issues: Automatically creates high-priority tasks in your Linear backlog for further investigation or implementation of high-impact data sources.

This ensures your product roadmap is continuously aligned with value creation, focusing development efforts on features that provide the most predictive lift.

2. Web UI Interaction: The "Frontend Wow Card"​

The primary user interface for CAtobuildCA's insights is the Feature Importance Wow Card, displayed on the decision details page.

2.1. Accessing the Card​

Navigate to any decision's detail page within the ChainAlign Web UI. If the decision's outcome has been recorded and contains feature importance data, the "Feature Importance Wow Card" will be prominently displayed.

2.2. Understanding the Insights​

The card provides a concise summary, typically highlighting:

  • Key Drivers: A list of the top 1-3 most impactful data features (e.g., "Weather Temperature", "Holiday Flag").
  • Contribution Percentage: The percentage of predictive lift each feature contributed to the forecast.
  • Total Importance: An overall score indicating the combined influence of all analyzed features.

Example Display:

✨ CatobuildCA Insight: Key Drivers for Decision [Decision ID] ✨
Our analysis reveals the most impactful data features influencing this decision's outcome.
Focusing on these areas can yield significant improvements.

- **Weather Temperature**: Contributed 45.00% to the predictive lift.
- **Holiday Flag**: Contributed 30.00% to the predictive lift.
- **Value Target Lag-1**: Contributed 18.00% to the predictive lift.

Total analyzed importance: 1.00

This visual summary helps you quickly grasp which external factors are truly driving your business outcomes.

3. Linear Integration: Automated Task Creation​

Beyond just displaying insights, CAtobuildCA integrates directly with your Linear task management system to automate the creation of actionable development tasks.

3.1. Triggering Task Creation​

When the OutcomeAnalysisService processes a decision outcome and identifies impactful features (e.g., features contributing above a certain threshold), it automatically triggers the creation of a new issue in Linear.

3.2. Linear Issue Details​

The automatically generated Linear issue will typically include:

  • Title: [CatobuildCA] Prioritize features for Decision [Decision ID]
  • Description: A detailed explanation of the analysis, listing the impactful features and their contribution percentages. It will also include a recommendation for further investigation.
  • Priority: Issues created by CAtobuildCA are assigned a high priority (e.g., 2) to ensure they are reviewed promptly by the development team.
  • Project/Milestone: The issue will be assigned to the relevant project and, if configured, a specific milestone (e.g., M44 - Self-Optimizing Roadmap (CAtobuildCA)).

Example Linear Issue Description:

Outcome analysis for decision [Decision ID] identified the following impactful features:

- **Weather Temperature**: 45.00% contribution
- **Holiday Flag**: 30.00% contribution
- **Value Target Lag-1**: 18.00% contribution

Further investigation is recommended to leverage these high-impact features.

4. End-to-End Workflow​

Here's how the CAtobuildCA decision loop works:

  1. Decision Execution: A decision is made and executed within ChainAlign.
  2. Outcome Recording: The actual outcome of the decision is recorded, including the feature_importance data from the forecasting engine.
  3. Outcome Analysis (Automated): The OutcomeAnalysisWorker periodically runs, identifies decisions with new outcomes, and triggers the OutcomeAnalysisService.
  4. Feature Importance Analysis: The OutcomeAnalysisService analyzes the feature_importance data, identifies top impactful features, and generates insights.
  5. Linear Issue Creation (Automated): If impactful features are found, the OutcomeAnalysisService calls the LinearJudgmentEngineService to automatically create a high-priority Linear issue.
  6. Development Action: The development team reviews the Linear issue, investigates the identified features, and potentially implements new data integrations or model enhancements.
  7. Improved Forecasting: These enhancements lead to more accurate forecasts, closing the loop and driving continuous improvement.

5. Next Steps for Development​

  • Implement DecisionRepository.findDecisionsWithNewOutcomes(): This method is crucial for the OutcomeAnalysisWorker to identify decisions that need analysis. (Currently a TODO in OutcomeAnalysisService.js)
  • Test the full workflow: Once the test environment issues are resolved, comprehensive integration tests should be run to verify the end-to-end CAtobuildCA loop.

Key Message: CAtobuildCA transforms your S&OP process into a self-optimizing system, ensuring your development efforts are always aligned with what truly drives your business outcomes.