# Experiments and lift

Run disciplined ad tests with Adwarp's change journal and lift analysis. Log a hypothesis before a change, measure before vs after with compute_lift, and read results with the confounds stated.

Ask it: "did that change help", "measure the lift", "log this test", "what tests are running", "was the new headline better".

Ad accounts accumulate changes nobody can evaluate later because nobody wrote down
what changed and why. Adwarp journals every change it makes automatically; this skill
adds the discipline around it: hypotheses going in, honest verdicts coming out.

## How to reach Adwarp

Prefer the Adwarp MCP tools if loaded: `log_ad_change`, `list_ad_changes`,
`compute_lift`. CLI: `adwarp changelog list`, `adwarp changelog add`, `adwarp lift <level>
<entityId> [-w days]`. The journal lives at `~/.adwarp/changelog.jsonl`. Changes made
directly in Campaign Manager are invisible until logged by hand; that is the most
common reason a lift read comes back empty.

## Starting a test

1. **Hypothesis first**, 1 sentence: "outcome-led headline will beat feature-led on
   cost per conversion."
2. **Make the change.** Copy changes on LinkedIn are recreate, not edit
   (`delete_ad` + `create_image_ad`, drafts, user confirms deletes); Adwarp journals
   these automatically. Changes made in Campaign Manager get `log_ad_change` with the
   entity, field, and after-value.
3. **Label it** (`-l "outcome-led headline test"`) and tag related changes so the
   test reads as 1 unit.
4. **1 change per entity at a time** where possible. 2 simultaneous changes on
   the same entity cannot be separated afterwards; if it happened anyway, say so in
   the read-out.

## Reading a test

`compute_lift` compares the window before each recorded change against the same
window after (default 14 days) and reports per-metric deltas: CTR, CPC, conversions,
conversion rate, cost per conversion.

Rules for an honest read:

- **Wait for volume, not just days.** No verdict without ~1,000 impressions or a few
  conversions on each side; otherwise report "inconclusive, recheck on <date>".
- A `partial` after-window means the change is recent; present it as interim.
- **State the confounds every time:** seasonality, LinkedIn's learning phase
  restarting after an edit (early post-change days usually dip), concurrent budget or
  audience changes. This is a directional pre/post comparison, never proof.
- Verdicts are exactly 1 of: better, worse, inconclusive.

## The test registry

When asked "what tests are running", `list_ad_changes` is the answer: recent changes
still inside their after-window (open tests), completed windows nobody has read yet,
and unlabeled changes worth labeling before they are forgotten.

## Report format

Per change: label, date, entity, before vs after on the 4 or 5 KPIs that
matter, verdict, confounds. Close with 1 portfolio line: what the wins suggest
testing next.

Canonical page: https://adwarp.dev/docs/skills/experiments
Last updated: 2026-10-18