Skip to main content

Guide: segment an audience with filters

Goal: measure a brand among a specific audience — e.g. a demographic or a Profiles segment — and compare it to the total population.

Steps​

  1. Look up the filter values you need (filters catalogue) — for example the Profiles definition_name and its categories.
  2. Build the filter(s), remembering the combination rules: AND across filters, OR within a filter's values (Filters).
  3. Run the same query with and without the filter to compare.

Request (filtered)​

{
  "meta": { "version": "v1" },
  "data": {
    "queries": [
      {
        "id": "young-adults",
        "entity": { "region": "us", "sector_id": 1, "brand_id": 1007 },
        "period": { "type": "month", "offset": 1, "amount": 1 },
        "metrics_score_types": { "likelybuy": "positives" },
        "scoring": "aware",
        "filters": [
          { "definition_name": "profiles_us__age", "op": "in", "values": ["18-24", "25-34"] }
        ]
      }
    ]
  }
}

To get the "everyone" baseline, send the same query with "filters": [].

Notes​

  • Filters and the scoring population are independent: filters decide who enters the query; scoring decides who forms the base.
  • Combine filter types freely (a demographic filter AND a metric filter, etc.).