Every change, graded
SEO change tracking that grades every change
SEO change tracking means recording each change to a site when it ships and checking later whether it moved the metric it targeted. Growth Radar freezes the baseline at merge, waits 28 days, then compares the changed pages with the most similar untouched pages of the same site and calls it win, loss, flat or inconclusive.
What it measures, and how
How a change is recorded, and how its verdict is reached.
- Recorded at ship time
- A change is recorded when its pull request merges, when you mark a fix shipped, or when an agent calls the MCP server. The metric it was meant to move (clicks, CTR, position, impressions, broken links or AI citations) is read for the pages it touched over the 28 days before.
- Compared with untouched pages
- After 28 days the same metric is read again on the changed pages and on the most similar untouched pages of the same site. It is a win when the changed page beats at least 95% of them and moves at least 10% more than they typically do, a loss in the mirror case, flat otherwise.
- Inconclusive is a result
- When the test is too weak to tell a real change from noise, which on low-traffic pages is usual, or there's nothing comparable to check against, the verdict is inconclusive: shown with its numbers and left out of every count.
- Still holding, or decayed
- A win is measured again after a second 28-day window. If at least half the lift remains, it's still holding; otherwise it's marked as decayed.
Why before-and-after isn't enough
Traffic moves for reasons that have nothing to do with your change: a Google update, a season, a competitor's launch. A before-and-after chart credits all of it to whatever you shipped that week.
Comparing the changed page with similar untouched pages of the same site takes the shared movement out. What remains is the part the change can claim, and when that part is too small to separate from noise, the honest verdict is inconclusive.
What it can't see
- Low-traffic pages usually come out inconclusive; the test needs enough clicks or impressions to detect a real change.
- A change that touched the whole site has no untouched pages to compare with, so it's inconclusive.
- Search Console trails real time by two to three days, so a verdict lands a few days after the window closes.
- A change whose starting number was typed in by hand rather than measured is kept out of the headline counts.
Growth Radar or another tool
AI-visibility trackers that report movement over time, per their own pages: Checked October 28, 2026.
| Tool | Price | AI engines | Best for |
|---|---|---|---|
| Growth Radar | Free $0 (1 site); Pro $49 a month; Agency $199 a month; 14-day free trial, no card | Narrow: Perplexity, ChatGPT, Gemini and Claude, checked daily | Teams that ship SEO and GEO fixes as code on several sites and want each one graded win, loss or flat 28 days after merge |
| Profound | Free 7-day trial on 50 fixed prompts, then Enterprise at a custom price. The $99 Starter and $399 Growth plans listed on 14 September 2026 are gone. | Up to 9 on Enterprise; the trial covers ChatGPT, Gemini and Google AI Overviews | Enterprise brand teams that want panel-based prompt volumes, log-based AI crawler analytics and content agents under one contract |
| Peec AI | Starter $95 a month (50 prompts, 3 models, unlimited seats); Pro $245; Advanced $495; Enterprise custom | Pick 3 models on self-serve plans, each extra one a paid add-on; up to 13 on Enterprise | Marketing teams and agencies that want prompt tracking plus ranked Actions at a self-serve price |
| Ahrefs Brand Radar | Index data from $199 a month, custom prompts from $50 a month, all AI models $699 a month, $0.02 per extra check | Google AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot | Teams that want market-wide share of voice from a pre-collected index of 514M+ prompts, with no prompt setup |
Sources: Profound pricing · Profound pricing, archived 14 September 2026 · Profound Prompt Volumes · Peec AI pricing · Peec AI Actions · Ahrefs Brand Radar
Common questions
What is SEO change tracking?
Keeping a record of each change to a site, the date it shipped and the metric it was meant to move, then checking that metric once search engines have had time to react. Growth Radar waits 28 days and compares against similar untouched pages, because a change that only beats its own past may be riding a site-wide trend.
Why compare against untouched pages?
Because they experienced the same Google updates, seasons and outages as the changed page. Whatever moved them too wasn't caused by your change. The difference between the two is the part your change can claim.
Why are so many SEO verdicts inconclusive?
Most pages on small sites get too few clicks for a test to tell a real change from noise. Calling those wins or losses would be guessing. Growth Radar shows them as inconclusive with their numbers and keeps them out of the totals.
Can an AI agent record the SEO changes it ships?
Yes. The MCP server has 41 tools, including one that records a shipped change with its baseline and one that reads its verdict back, so an agent in Claude Code or Cursor closes the loop on its own work.
Is there a free way to grade one change?
Yes. The free grader takes a Search Console export and grades one change as won, lost, flat or inconclusive with the same rules, in the browser, without an account.
Related
Try it on your own sites
Connect Search Console and the last 28 days import straight away. The trial needs no credit card.