Index of every page on this site: llms.txt
All skills
google avatar

/meridianbudgetoptimization

@71777a2
by googlegoogle/meridian1.6k stars
314
Run it yourself

Loads a fitted Meridian model and runs budget optimization scenarios. Use when the user wants to load a saved model and find the optimal budget allocation or target ROI spending. Don't use for visualizing model performance (use meridian-result-visualization) or building/fitting the model (use meridian-model-building).

  • 3 files
  • 8.2 KB
  • Updated last month
  • GitHub
Use this Skill: https://skilld.dev/gh/google/meridian/meridianbudgetoptimization

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈87 tokens for metadata: the name and description. ≈1.2k when used: this file. ≈861 more on demand in 2 files.

Description uses 4.4% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈350 characters, or 4.4%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Meridian Budget Optimization

A skill for loading a fitted Meridian model and running budget optimization reports.

Core Workflow

Interactivity Checkpoint Rule

Throughout this workflow, you will encounter CRITICAL INTERACTIVE CHECKPOINTs. At each checkpoint, you MUST:

  1. Present the current proposed configurations, report paths, script path, or status to the user for approval.
  2. Ask the user if they are ready to proceed using the available user-interaction tool (e.g., ask_question), structured as a multiple-choice question. Do NOT use raw chat text.
  3. Wait for their response before proceeding.
    • MANDATORY: You MUST pause at every checkpoint regardless of the initial prompt instructions (even if the user request contains phrases like "run autonomously", "execute directly", "fix autonomously", etc.). The initial request does NOT bypass these interactive checkpoints.
    • Note: If the user replies to a checkpoint with a generic approval (e.g., "proceed", "do what you think is best"), proceed with the proposed defaults.

1. Initial Setup

  • Prompt the user for the path to the serialized model file (meridian_model.binpb by default).
  • Prompt the user for the output path for the Budget Optimization report (optimization.html by default).
  • Prompt the user for the desired path for the generated Python script. If unspecified, default to model_build/run_budget_optimization.py (or relative to the model directory).
  • CRITICAL INTERACTIVE CHECKPOINT: Present the gathered paths to the user and obtain confirmation before configuring optimization parameters.

2. Configure Optimization Parameters & Interactive Checkpoint

  • Prompt the user to configure the optimization scenario. See optimization.md for the full list of available parameters. Key configurations include:
    • Fixed vs Flexible Budget: Is the total budget fixed, or are we looking for a budget to hit a target ROI/mROI? (Defaults to Fixed).
    • Total Budget: If fixed, what is the total budget? (Defaults to historical spend).
    • Spend Constraints: What are the upper/lower bounds for spend shift per channel? (Defaults to 0.3 for fixed, 1.0 for flexible).
    • Target ROI / Target mROI: If flexible, what are the target ratios?
    • Reach & Frequency Parameters: Should optimal frequency be used? (Defaults to True for RF channels).
  • CRITICAL INTERACTIVE CHECKPOINT: Present the proposed optimization configurations to the user and obtain approval before continuing to generate the code.

3. Add Model Loading Code

  • Use meridian_serde.load_meridian() to load the model.
  • See load_model.md for code template.
  • CRITICAL INTERACTIVE CHECKPOINT: Present the model load path configuration and proposed Python code snippet to the user, and obtain approval before continuing to specify budget optimization.

4. Add Budget Optimization Code

  • Use optimizer.BudgetOptimizer to run optimization and generate the summary, applying the confirmed configuration.
  • See optimization.md for code template.

5. Pre-execution Checkpoint

  • CRITICAL INTERACTIVE CHECKPOINT: Ask the user for final confirmation to execute the budget optimization script now.

6. Execution & Script Setup

  • Write the accumulated Python script to the user-specified path. If your file-writing tool can ask the user to review the written file, don't request that review: the pre-execution checkpoint already asks for approval, and an extra prompt pauses the workflow.
  • Execute the script with the Python interpreter of the environment where google-meridian is installed: the one the user named; else python3 if a virtual environment is already active (VIRTUAL_ENV or CONDA_PREFIX is set); else the project directory's virtual environment (.venv/ or venv/, e.g., .venv/bin/python3) if there is one; else python3. If import meridian fails, ask the user which environment to use instead of searching the machine or installing packages unprompted.
  • CRITICAL: Do NOT delete the generated reports, the script, or the output directory at the end of the task. These are the deliverables requested by the user and must be preserved.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at 71777a2. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated last month

README badge

README badge for google/meridian/meridianbudgetoptimization