- Informational
Why Keyword Rankings Change: How to Investigate a Drop
Editorial responsibility: SearchCub
Keyword rankings change as competing pages, your website, and Google's systems change. Before fixing a reported drop, confirm that you are comparing the same query, market, page, and kind of position. A different measurement can look like lost visibility even when nothing comparable has declined.
1. Check whether the two observations belong together
Record these details beside the old and new numbers:
Detail | What must match |
|---|---|
Query | The actual search phrase, not a similar variant |
Target | The same page, or the same domain if comparing domain visibility |
Audience | Country and device where available |
Measurement | The same source and position definition |
Timing | Dated observations or comparable, complete reporting periods |
Coverage | Enough comparable results to support the conclusion |
For an illustrative example, a desktop observation at position 4 in the U.S. and a mobile observation at position 12 in the U.K. do not establish an eight-position drop. Both the audience and device changed.
Google Search Console's average position summarizes positions across impressions. It is different from the position returned by one search. Adding more impressions for harder queries can change an overall average without proving that a particular query declined.
In SearchCub, retained Domain or Page research can provide a comparison when compatible observations exist. Check their dates and coverage. A first observation is a baseline; a missing row in partial results does not prove a lost ranking. Retained evidence cannot reconstruct a period that was never observed.
2. Find out what actually fell
For a site you manage, open Google Search Console separately and compare complete periods in its Search results performance report. Filter to the affected query or page, then check country and device. Look at clicks, impressions, and position together.
Pattern | Useful next check |
|---|---|
One retained row is missing | Check collection coverage before declaring a loss |
Clicks fell while impressions and position stayed similar | Inspect the displayed result and competing search features |
Impressions fell while position stayed similar | Check seasonal demand and changes in the query mix |
Several comparable queries lost position persistently | Examine affected pages and the results now above them |
A whole section disappeared after a site change | Check access, redirects and indexing first |
These are investigation paths, not diagnoses. Google's traffic-drop guidance distinguishes technical problems, shifting demand, search-system changes, and other causes. The graph's shape helps narrow your search; it cannot identify the cause by itself.
3. Match the timing to something you can verify
Start a short incident note with the affected URLs, comparison dates, and first confirmed change. Then investigate in this order:
Your own changes. Did a release alter URLs, internal links, canonical tags, access rules, or the main answer? Check the actual affected pages against the change record. A release date is a lead, not proof.
Google's view of the URL. Use URL Inspection to examine indexed status and the selected canonical. If Google is treating a different URL as the main version, investigate that before rewriting the article.
The current competition. Open SearchCub's SERP overview for the query and check the displayed market. Read the pages now competing with yours. Note a concrete difference, such as a working calculator replacing static explanations, rather than assuming every higher result has better SEO.
A documented search update. Check the dates in Google's core-update guidance. Google recommends waiting at least a full week after an update finishes before analyzing the effect. An overlapping update does not prove that it caused your particular drop.
4. Make a change you can evaluate
If a broken redirect is confirmed, repair it and verify the destination. If the page works but no longer answers the query well, use the competing results to identify the missing task; our SERP analysis walkthrough helps turn that finding into a brief.
Write down the hypothesis, action, and observation that would count as improvement. For example: “The guide stopped showing the setup instructions after a template change; restore them and check the rendered page, then review comparable search data.” This is an illustrative investigation, not a reported result.
Avoid changing the title, URL, structure, and content at once just to make the chart move. Keep a dated record, allow new observations to accumulate, and judge whether useful visits recover alongside visibility. One better position on one check is a result to watch, not a completed recovery.