Per-ID eCTR + eCPM Overrides

Overview

Kevel’s ad server is designed to balance automation with control. While Kevel predicts ad performance using historical click-through and conversion data, we understand that retailers often have deeper, real-time context that isn’t visible to the Kevel decision engine.

The Per-ID eCTR and eCPM Overrides feature allows you to bring your own understanding of expected ad performance—whether from internal machine learning models, business logic, or domain expertise—into Kevel’s auction engine. By doing so, you gain the benefits of Kevel’s targeting, pacing, and reporting systems while fine-tuning performance based on data only you have.

How It Works

Override predicted eCTR (expected click-through rate) or eCPM (effective cost per mille) values per request by including an eCtrOverride or eCpmOverride parameter in the placements block of in your Ad Decision Request (ADR). These values can be scoped to specific IDs, such as advertiser, campaign, flight, ad, or creative template fields.

This mechanism supports:

Example Use Cases

Integration Notes

Example Ad Decision Request

Ad Decision Request

{
  "placements": [
    {
      "networkId": 123,
      "siteId": 456,
      "adTypes": [5],
      "eCtrOverride": {
        "forceOverride": false,
        "idAttribute": {
          "adId": {
            "9945867": 0.025,
            "8893776": 0.03
          },
          "ctFoo": {
            "bar": 0.015,
            "baz": 0.05
          }
        }
      }
    }
  ]
}

Parameters

Updated 10 months ago