SearchEngineInfo All articles
Search Analytics

Ten-Day Teardown: A Ranking Postmortem for US Publishers

SearchEngineInfo
Ten-Day Teardown: A Ranking Postmortem for US Publishers

Ranking swings after a search update invite two opposite mistakes. One is panic editing: titles rewritten, pages consolidated, and internal links changed before anyone knows what moved. The other is waiting: teams assume the update will sort itself out and lose weeks of recoverable traffic. A ten-day postmortem is designed to avoid both.

The term postmortem can sound dramatic. In practice, it is an after-action review with a deadline. It gives US publishers a fixed window to confirm scope, separate signal from seasonality, document wins and losses, and assign a focused backlog. It is not a hunt for a single culprit. It is a structured reading of what changed and why.

Start before the swing: the baseline file

A postmortem is only as good as its baseline. Before the next confirmed or suspected update, save a snapshot of performance by query, page, template, and device. Include impressions, clicks, average position, conversions where available, and a small set of priority queries. Store it outside the live dashboard so later changes do not overwrite the comparison.

Also record editorial and technical changes from the preceding 30 days: launches, migrations, redirects, title changes, and content pruning. Without that context, the postmortem will confuse correlation with causation.

Days 1-2: Confirm scope and freeze changes

On the first day, confirm whether the swing is real and how broad it is. Compare week-over-week and year-over-year windows, and check whether the pattern appears across the site or in specific sections. If an update is confirmed by the platform, note the dates. If it is suspected, say so plainly in the record.

Freeze non-urgent changes for the ten-day window. This does not mean stopping all work; it means preventing a wave of reactive edits that will make the postmortem unreadable. Urgent fixes, such as broken checkout flows or accidental noindex tags, still proceed.

Days 3-4: Segment by query, template, and intent

Aggregate traffic numbers hide the story. Segment the change by query class, page template, content type, and intent. A publisher may see losses in news recaps but gains in evergreen explainers, or the reverse. Template-level analysis can reveal whether the issue is technical, editorial, or both.

Use a consistent classification: branded versus non-branded, informational versus commercial, fresh versus evergreen. Compare the segments against the baseline. The goal is to find the smallest useful pattern, not to explain every page.

Day 5: Run seasonality and demand checks

Before attributing a swing to an algorithm, check demand. Search interest may have fallen because a news cycle ended, a holiday passed, or a competitor released a major report. Compare query volume trends, not just rank positions.

Look for calendar effects, weather events, sports seasons, and US-specific moments such as election cycles or tax deadlines. If demand fell while rankings held, the problem may be market-driven. If rankings fell while demand held, the update is a stronger suspect.

Days 6-7: Inspect winners and losers with equal rigor

Most postmortems overstudy losses. Winners are equally informative. Sample pages that gained, pages that lost, and pages that did not move. Compare content depth, freshness, author signals, internal links, page experience, and SERP feature presence.

Ask whether the winners share a trait the losers lack: clearer sourcing, stronger entity signals, better structured data, or more complete answers. Avoid single-page anecdotes. Look for repeated patterns across the sample.

Day 8: Assign owners and write the record

Every finding needs an owner and a date. Technical findings go to the SEO or engineering lead. Editorial findings go to the content lead. Analytics gaps go to the measurement lead. The record should include what changed, what was checked, what was ruled out, and what remains uncertain.

Document both wins and losses. A win may be a template that held steady or a content type that gained. Those are evidence for future decisions, not just morale.

Day 9: Build the prioritized backlog

Convert findings into a backlog ranked by expected impact and effort. Separate quick corrections from structural work. A quick correction might be restoring an internal link or updating a stale date. Structural work might be revising a template or consolidating overlapping pages.

Do not let the backlog become a list of generic SEO tasks. Each item should tie to a specific finding from the postmortem.

Day 10: Publish the decision memo

Close the ten days with a short memo: scope, method, findings, actions, and the next review date. Share it with editorial, product, and analytics. The memo is the institutional memory that prevents the next swing from becoming another round of guesswork.

What not to do

Do not rewrite titles across the site on day two. Do not assume every loss is a penalty. Do not wait a quarter to document what happened. The ten-day postmortem is not a guarantee of recovery. It is a discipline for responding to volatility without overreacting to noise.

All Articles

Related Articles

Twelve Queries, One Answer: Mapping AI Search Fan-Out for US Publishers

Twelve Queries, One Answer: Mapping AI Search Fan-Out for US Publishers

Pricing the Answer: A Zero-Click Worksheet for US Publishers in 2026

Pricing the Answer: A Zero-Click Worksheet for US Publishers in 2026

Beyond the Click: Building a Search Visibility Scorecard for US Publishers in 2026

Beyond the Click: Building a Search Visibility Scorecard for US Publishers in 2026