{
  "id": "campaign-holdout-summary",
  "claim": "In an 8\u201316 July 2026 apparel campaign with a randomized 50/50 holdout (19,606 people, counted on the ad network's person ID), pooled experiment traffic produced 64% more buyers per person and 55% more revenue per person than the holdout, p = 0.005.",
  "definition": "Pooled multi-page experiment vs holdout summary for one commerce brand campaign window. Not best-page vs holdout.",
  "numbers": [
    {
      "label": "people_total",
      "value": 19606
    },
    {
      "label": "holdout_share_design",
      "value": 0.5
    },
    {
      "label": "pages_racing",
      "value": 22
    },
    {
      "label": "test_buyers_per_person",
      "value": 0.0088,
      "note": "85 buyers of 9,607 people"
    },
    {
      "label": "holdout_buyers_per_person",
      "value": 0.0054,
      "note": "54 buyers of 9,999 people"
    },
    {
      "label": "relative_lift_buyers_per_person",
      "value": 0.64,
      "note": "p = 0.005"
    },
    {
      "label": "relative_lift_revenue_per_person",
      "value": 0.55
    }
  ],
  "canonical_url": "https://lutiq.com/research/what-ab-tests-cannot-see/",
  "methods_url": "https://lutiq.com/research/lutiq-holdout-protocol/",
  "open_data_url": "https://lutiq.com/data/research/campaign-holdout-2026-07.json",
  "sources": [
    "https://lutiq.com/data/research/campaign-holdout-2026-07.json",
    "https://lutiq.com/research/what-ab-tests-cannot-see/"
  ],
  "last_verified": "2026-09-23",
  "caveats": [
    "Pooled experiment vs holdout \u2014 not best single page vs holdout.",
    "One brand and one campaign; 139 buyers across both arms. Do not cite as a guaranteed product ROI.",
    "Replaces the earlier per-visitor figure for 8\u201322 July (+60% sales per visitor)."
  ],
  "_meta": {
    "schema": "https://lutiq.com/data/answer-packets/_schema.json",
    "publisher": "Lutiq"
  }
}