How-to: Compare audit runs
Two raw audit results are noisy (“dimension X went 6→7”). Audit Compare turns them into a narrative — what changed, why, what to try next. This walks through the two flows you’ll use most: temporal compare (same URL over time) and variant compare (draft vs. live).
Audit Compare is on every paid tier. Each compare run costs 1 audit-unit (the LLM narrative call) — the underlying audit runs are reused if they already exist.
Temporal compare — same URL, two points in time
Use this after shipping a copy change to verify whether it actually moved positioning.
1. Open the page workspace
Sidebar → Landing Pages → click the row for the URL.
2. Go to the History tab
You’ll see every audit run for this URL, newest first. Free plans only see the latest run; paid plans keep all history indefinitely.
3. Compare two runs
Each row past the first has a Compare to previous button on the right. Click it on the newer run and PlumPMM compares it against the run immediately before in the list.
Wait ~10–15 seconds for the LLM narrative call. The result drops in below the row:
- Overall delta (regression / improvement / mixed / noise) and before/after scores
- Per-dimension deltas — biggest movers called out
- Plain-prose narrative referencing the actual quoted page text
If the narrative call fails (rate limit, outage), you’ll see a score-only result — the numeric deltas without the prose. Click Re-run narrative to retry without re-paying for the underlying audits.
Variant compare — draft vs. live
Use this before shipping a draft, to verify the rewrite would actually move the score.
Requires the draft to be linked to a Landing Page. Open the draft workspace → Landing Pages tab → link a page if you haven’t already.
1. Open the Preview tab
Inside the draft workspace, click Preview. You’ll see the draft rendered as a grayscale landing page mock.
2. Generate a preview audit
Click Generate preview audit. This audits the draft as if it were live — same 13-dimension pipeline, same scoring. ~30 seconds.
3. Click “Compare to live page”
PlumPMM also runs a fresh audit on the linked live page (so the comparison uses two recent baselines, not a stale one), then generates the narrative. Total wall time ~60 seconds.
You’ll see the same shape as a temporal compare — overall delta, per-dimension swings, narrative — but framed as draft-vs-live. Use it to pick the right draft variation before you ship.
vs.-Competitor compare
Briefly: from the Research surface you can compare any audited competitor against your own URL or draft preview. Same engine, tagged as the cross-page kind. → How-to: Track a competitor
What’s next
| Goal | Read |
|---|---|
| Understand the three compare kinds and the score-only fallback | Concept: Audit Compare |
| Get compares fired automatically when a page drifts | How-to: Set up a monitor |
| Skip the manual compare loop on a fresh URL | Concept: Fasttrak |