Know when a number moves.
Prism checks your alert rules every time Search Console syncs, watches for anomalies alongside them, and turns whatever fires into a task with a name against it.
How something gets from a number to a job.
- A rule watches a metric
- Clicks, impressions, click-through rate or average position, with a condition and a number against it. Checked every time Search Console data lands, against the current period and the one before it.
- Anomaly detection runs beside it
- Today is scored against the last 28 days, measured from the middle value rather than the average, so one freak day cannot drag the baseline far enough to hide the next real drop.
- What fires lands in triage
- Everything raised arrives in one queue carrying a severity.
- Then it becomes somebody's job
- An owner, a status and a due date, and the work can land in whichever tracker your team already uses.
Prove the thing you shipped did something.
Log the migration on the day you ship it. The annotation sits on every performance timeline from then on, with a before-and-after click verdict over its own measurement window.
- Logged on ship day
- The migration, the content launch, the template change, marked on the timeline the day it happened.
- Its own window
- Each annotation compares the clicks before and after over its own measurement window and states the verdict.
- On every plan
- Alerts, tasks and annotations run on every plan, and none of them uses credits.
And the history to argue from.
Search Console forgets after 16 months. Budget conversations do not.
- Year-on-year that keeps working
Every daily sync is stored permanently, so the comparison you need in year three is still sitting there.
- GA4 on the same dates
The 4 traffic reports share the property picker, date range and comparison period with your search data, so a drop in clicks and a drop in revenue sit on one timeline.
- Fix First to spend the sprint well
Findings sorted by effect: the sprint goes on pages that are broken while earning, then on recoverable losses.
Daily, on purpose.
Alerts run on the daily Search Console sync, aligned to when Google finalises its numbers, so what fires is a real movement in a finalised figure.
Anomaly detection builds about a week of history before it commits to an opinion, and it skips very low-volume metrics.
- metrics a rule can watch, checked on every sync
- 4
- metrics a rule can watch, checked on every sync
- days behind the anomaly baseline, measured from the median
- 28
- days behind the anomaly baseline, measured from the median
- of history at connect, then kept for good
- 16mo
- of history at connect, then kept for good
- credits used by alerts, tasks or annotations
- 0
- credits used by alerts, tasks or annotations
The rest of the platform
- AI Visibility Monitor brand mentions across ChatGPT, Claude, Gemini and Perplexity: prompts on a schedule, every answer kept in full, every cited source listed. 18 reports.
- Search Performance 17 reports over your Search Console data: queries in striking distance, cannibalised pages, position changes, month on month, and the day Google swapped the page that ranks.
- Site Health Crawl and Lighthouse-score every page, then sort the findings by what each does to traffic: broken while earning, recoverable losses, upside, housekeeping. 14 reports.
Set one rule and find out first.
Connect a property, pick a metric and a threshold, and Prism checks it every time new data lands. Alerts, tasks and annotations run on every plan, and the Google access behind them is read-only. When the question is what the answer engines say about you, here is how far to trust a monitoring number.