How to Improve Search Rankings: Diagnose the Cause Before You Rewrite
June 15, 2026
Author: Shusaku Yosa
Improving search rankings starts with diagnosis. Establish which page and query changed, then separate indexing problems, changing demand, positions, and click-through behavior. A useful rewrite addresses a documented information gap; it is not simply a longer version of the existing page.
Separate a ranking problem from a traffic problem
Fewer search visits do not automatically mean worse rankings. Demand, result presentation, country mix, technical availability, and indexing can all change. Write down the observation first: fewer clicks, fewer impressions, or a position change for the same query.
Google’s traffic-drop diagnostic guide describes several possible causes. A single spot-check of one result is insufficient evidence for a site-wide diagnosis.
Compare the same pages, queries, dates, countries, and devices
Match the comparison window, search type, exact URLs, language, countries, and devices. Prefer equal durations with comparable weekday patterns, and consider a corresponding prior-year period when seasonality matters. Then filter the page and inspect its visible queries.
Hypothetical page, equal 28-day windows | Before | After |
|---|---|---|
Topic query A | 900 impressions; 9 clicks; position 10 | 500 impressions; 5 clicks; position 10 |
Query B | 100 impressions; 1 click; position 30 | 500 impressions; 5 clicks; position 20 |
Page total | 1,000 impressions; 10 clicks | 1,000 impressions; 10 clicks |
Query B improved, but the query mix also changed. The page average alone cannot describe the outcome. Inspect like-for-like queries and counts. Record visible site: searches separately and do not reconstruct anonymized queries.
Check indexing and technical issues before editing content
If a page has no Performance row, check its current indexing state instead of assuming it is excluded. Inspect redirects, selected canonicals, directives, and last crawl. Review manual actions or security issues when relevant.
An unintended noindex deserves attention before wording changes. Conversely, an indexed page with no exposure for a chosen query may have a relevance problem rather than a technical one. Google uses multiple ranking systems; there is no official fixed “200-factor” scorecard that guarantees improvement.
Choose different actions for demand, position, and CTR changes
Observation | Evidence to investigate | First action |
|---|---|---|
Impressions and positions fall | Matched queries, technical state, result changes | Inspect the URL and compare queries |
Impressions fall; matched positions hold | Demand, seasonality, audience mix | Check comparable periods and demand |
Impressions hold; clicks fall | Query mix, title, search presentation | Review actual results and expectations |
No page row | Filters, period, indexing state | Inspect the URL and reporting scope |
These are hypotheses, not causal verdicts. A CTR decline may reflect new queries or result features rather than a bad title. For sparse data, report the click counts before emphasizing percentage changes.
Improve the page around the information users still need
Describe a specific gap: missing eligibility conditions, unclear exclusions, an unexplained calculation, or a procedure without a way to confirm success. Add the evidence, comparison, or worked example that resolves it. Increasing word count is not an independent objective.
Use Google’s helpful-content guidance to review authorship and the basis for factual claims. In a regulated field such as legal services, verify qualifications, case claims, and advertising language before optimizing the copy. Rankings should not reward misleading promises.
Handle core updates and small ranking movements carefully
When a core update may be involved, check the official rollout status and completion date. Google’s core-update guidance recommends waiting at least a full week after completion before analyzing and avoiding drastic changes for small movements.
A modest change in one query is not a reason to delete an article or change its URL. Investigate substantial persistent shifts across relevant pages. Neither a rewrite nor an update finishing guarantees recovery to a particular position.
Record changes and decide how to evaluate them
Improvement ticket | Hypothetical example |
|---|---|
Target and gap | Pricing guide: extra-charge conditions are missing |
Evidence | Relevant visible queries and documented reader questions |
Change | Add an inclusion/exclusion table; owner verifies facts |
Review scope | Published version, equal windows, Web search, same language |
Decision | Assess impressions, clicks, and verified inquiries separately |
If titles and body content change together, their individual contributions cannot be isolated. Record advertising, seasonality, and other releases too. Begin with one URL and a clear chain from observation to gap, change, and review.


