the copilot

Analyze an MMM output

An agentic copilot that interprets Marketing Mix Model outputs and answers your questions with grounded reasoning and explicit confidence.

Load a sample MMM output

…or analyze your own MMM output

expected format

{
  "metadata": {
    "model_type": "bayesian",
    "data_span_weeks": 52
  },
  "channels": [
    {
      "name": "Search",
      "spend_weekly": [
        12000,
        12500,
        11800,
        13000
      ],
      "roi_point": 3.2,
      "roi_ci": [
        2.8,
        3.6
      ],
      "adstock_decay": 0.4,
      "saturation_params": {
        "half_saturation": 40000,
        "shape": 1.1
      },
      "contribution_pct": 0.45
    },
    {
      "name": "Social",
      "spend_weekly": [
        8000,
        8200,
        7900,
        8100
      ],
      "roi_point": 1.9,
      "roi_ci": [
        1.2,
        2.6
      ],
      "adstock_decay": 0.6,
      "saturation_params": {
        "half_saturation": 25000,
        "shape": 1
      },
      "contribution_pct": 0.3
    }
  ],
  "model_diagnostics": {
    "r_squared": 0.86,
    "mape": 0.12
  }
}

Pick a sample or upload your own MMM output to see a grounded, streamed analysis and start chatting about it.