plugin.json

{
  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
  "name": "fiera-incrementality",
  "version": "1.0.0",
  "description": "Design and read advertising incrementality tests: platform lift studies, geo experiments, switchbacks and directional pre/post checks, with a bundled, validated analysis script.",
  "author": {
    "name": "Fiera",
    "url": "https://fiera.digital"
  },
  "homepage": "https://fiera.digital",
  "license": "Apache-2.0",
  "keywords": [
    "incrementality",
    "incrementality testing",
    "geo lift",
    "geolift",
    "lift study",
    "conversion lift",
    "brand lift",
    "holdout test",
    "matched market test",
    "iroas",
    "marketing mix modeling",
    "meridian",
    "paid media",
    "google ads",
    "meta ads"
  ],
  "extensions": {
    "com.openai": {
      "interface": {
        "displayName": "Fiera Incrementality",
        "shortDescription": "Incrementality tests for ads",
        "longDescription": "Measure whether your ads drive results that would not have happened anyway.\n\nFiera Incrementality helps paid media teams choose the right test, design it, and read it:\n\n- Planner: turns a question (is brand search incremental? should we cut Meta prospecting?) into a test plan. Recommends the platform's own lift study when one is available, a geo experiment when channels must be compared on business data, and a switchback or a directional before/after read when nothing else is possible.\n- Geo test design: from a CSV of KPI by region and day, picks treated regions by stratified randomization, estimates the minimum detectable effect on held-out history, the test length and the cost, and writes a plan to approve before launch.\n- Geo test readout: estimates the effect with two methods (time-based regression and synthetic control), intervals from placebo tests, incremental CPA and ROAS, and robustness checks.\n- Quick incrementality: a directional read on a pause or budget change, compared with several baselines and their normal noise, plus switchback analysis for single markets.\n- Lift result review: an independent check of a lift study, vendor test or internal claim.\n\nResults feed attribution (incrementality factors) and marketing mix models (Meridian, Robyn, PyMC-Marketing) as calibration. Methods follow the published work behind Google's TBR and Meridian GeoX, Meta's GeoLift and the synthetic control literature, with sources cited in the skills.\n\nThe analysis runs in a bundled Python script on the data you upload. It is calibrated on simulated panels: false positive rate in A/A tests, coverage of the true effect, and power at the predicted detectable effect.\n\nIntended for performance marketing, growth and analytics teams. Limits: results are as good as the KPI data you provide; directional methods are labeled as such and do not prove causation; platform lift studies are run in the ad platform, not by this plugin.",
        "developerName": "Fiera",
        "category": "Data & Analytics",
        "capabilities": [
          "Recommend an incrementality method for a channel and decision",
          "Design geo experiments with power analysis",
          "Analyze geo experiments with placebo-based intervals",
          "Directional before/after and switchback analysis",
          "Review lift study results"
        ],
        "websiteURL": "https://fiera.digital",
        "defaultPrompt": [
          "What is incrementality, and how do I test it for my ads?",
          "Is our brand search incremental? Help me design a test.",
          "Here are our orders by region and day. Design a geo holdout."
        ],
        "brandColor": "#8A6D00",
        "logo": "./assets/logo.png",
        "composerIcon": "./assets/composer-icon.png",
        "supportURL": "https://fiera.digital/contact",
        "privacyPolicyURL": "https://fiera.digital/privacy",
        "termsOfServiceURL": "https://fiera.digital/terms"
      }
    }
  }
}