Short answer: when traffic falls, the first question isn't "what should we rewrite?" It's "what changed, and on which day?" STORY reads a Search Console Performance export, finds the sharpest one-day fall in clicks, lists the searches where you're seen but not chosen, puts your money pages at the top, and writes up to seven ranked tickets. Each ticket points at the companion check that proves or disproves it.
For a general tour of the Performance report, see how to use Google Search Console for SEO. This piece is the triage logic.
What you drop in
The Performance export as .xlsx, .xls or .csv. Google's multi-sheet export works as it is: Chart (daily rows), Queries, Pages, and Devices and Countries if you include them. You can also paste a table. More sheets give better tickets. Chart plus Queries is the minimum for a drop.
The detectors, under the hood
- The cliff. STORY walks the daily rows and finds the single biggest day-on-day fall in clicks. It only calls it a cliff if there are at least 4 days of data and the fall is 20% or more. It shows the average position before and after where the export has it. A cliff goes to the top of the board as P0, because a one-day change usually means something changed on the site, not that the content got worse.
- Week on week. With at least 14 days of data, STORY compares the last 7 days with the 7 before. A move of 10% or more becomes a P2 ticket, but only when there's no cliff. A cliff already explains it.
- Visible but not chosen. Any query or page with 100 or more impressions and a CTR under 1% is flagged. That's a snippet and intent problem, not a ranking one.
- Money pages. The top pages by clicks are listed, and the top one gets a "live-check before touching" ticket. Status before opinion.
- Navigational demand. Single-word queries, and queries mentioning login, app, portal, dashboard, contact or careers, are labelled navigational, so you don't over-read them.
Tickets are capped at seven. The last one is always "bank this run as a baseline", because next week's export is only useful if you kept this one.
Where each ticket sends you
Every ticket has buttons into the companion that proves it:
- A cliff goes to DRIFT for redirects and VEXA for robots.
- Weak queries go to THINN and GATE.
- Weak pages go to THINN and RAUM.
- The site map ticket goes to ATLAS and LOCRA, with a sitemap guess built from the Pages export's origin.
Edge cases
- A one-day dip that bounces back still registers as a cliff if it's 20% or more. Read the before and after, and check whether it held.
- Weekends. A Sunday dip on a B2B site can look like a cliff on a short export. Use 28 days or more where you can.
- Penalty context. Tick "a manual action or thin-content review is in play" and STORY adds an Appeal readiness chapter. It doesn't change the numbers, only the order of work: contain the thin section, prove the site's shape in ATLAS, decide per section in writing, then verify in LOCRA. STORY can't see your Search Console messages and doesn't submit anything.
Worked example
Date,Clicks,Impressions,Position
2026-10-04,402,15100,6.2
2026-10-05,410,15350,6.1
2026-10-06,251,14900,8.9
2026-10-07,246,14800,9.0
...
Top queries,Clicks,Impressions,Position
sample widget pricing,6,2300,4.8VERDICT: P0 - explain the 2026-10-06 cliff (410 -> 251, 39% down)
The cliff is a 39% fall in a day, with position sliding from 6.1 to 8.9. That pattern points to a release, not a slow content decline. "sample widget pricing" is a P1 visible-but-not-chosen query: 2,300 impressions at 0.26% CTR. STORY's order of work is DRIFT and VEXA on the URLs that changed around 6 October first, and the title rewrite second.
Run it: open STORY and press Load sample
The money angle
Most traffic-drop spend goes on the wrong fix, done confidently: rewriting content that was never the problem. STORY's job is to stop that in the first hour. Name the day, check the plumbing, protect the pages that earn, then rewrite.
That's the same first pass I run in a traffic drop audit, before any number goes in front of a client.