Forensic SEO Audit

I anchor every forensic SEO audit on a single date: the day impressions or clicks first declined in Google Search Console. That point gives me a reference to correlate against events. Without it, you are chasing symptoms.

Step one: open the Search Console performance report and set the date range to the last six months. Look for a sharp change. If the drop was gradual, compare week over week. I set a custom date range to find the exact day. For example, a client's organic traffic fell 43% on 12 March. The Search Console graph showed a clear cliff on 11 March. That was my anchor.

Step two: match that date against three categories: your own changes (deploy logs, CMS updates, server moves), search engine updates (check the list from Moz or Semrush for that date), and third-party signals (competitor launches, backlink loss). In the client case, their deploy log showed a CMS update on 10 March that introduced a noindex tag on all category pages. Rolled back the change, traffic recovered in six days.

Step three: verify indexing status using Search Console's coverage report and URL inspection tool. I look for spikes in "Excluded" pages or "Submitted URL not indexed". Also check for manual actions. This step often reveals the real cause before you run a crawl. In a different case, I found 30% of pages were unintentionally blocked by a new Disallow directive in [robots.txt](/robots-txt/).

Step four: run technical checks on the affected page set. I focus on crawlability, robots.txt, redirects, mobile usability, and [Core Web Vitals](/core-web-vitals/). But I only do this after confirming the indexing status, because the crawl data might mislead if indexing was already removed. For a site using React, I checked the rendered HTML and found Googlebot was seeing empty shells due to a missing pre-rendering service. The fix was a server-side rendering update.

Step five: when no internal change explains the drop, look outward. Check backlink profile for sudden losses or toxic links. Use Ahrefs or Majestic to compare link growth vs competitors. A friend's site lost 60% of traffic; I found no technical issue. Competitor analysis showed a new comparison page that outranked them for key queries. The forensic audit flagged it, and they built better content to reclaim positions.

That path has saved me from wasting hours on irrelevant fixes. The key is to follow the timeline, not guess. I also keep a log of all changes for future audits.

The biggest mistake I see is assuming a traffic drop is always an algorithm update. Many times it is self-inflicted: a bad deploy, a changed robots.txt, a missing sitemap. One client lost 40% of traffic after a hosting migration that changed the server response from 200 to 503 for key URLs. The team blamed Google until I checked the server logs.

Do not skip the timeline. Without knowing when the drop started, you are guessing. I once spent a week [analysing on-page issues](/seo-audit/) when the decline began on a date that matched a competitor's new ad campaign. The timeline would have shown the correlation.

Do not ignore external factors. A competitor might have launched a better product or a new feature that steals impressions. Or the market shifted seasonally. Check Google Trends for your niche. For a travel site, the drop aligned with a change in search behaviour post-pandemic. No technical fix could solve that.

Be careful with automated crawl tools. They flag many issues — broken links, missing meta descriptions, slow pages — but they do not show timing. I once ran a crawl that found 200 broken links. The drop, however, happened months before those links appeared. The real cause was a page template change that removed structured data from product pages.

Edge case: what if Search Console data is delayed? Use server logs or CDN logs to find the traffic drop date. Cloudflare logs show request counts per URL. I cross-check with Google Analytics if available.

Another watch-out: don't ignore developer changelogs. Many teams do not document changes. Push for a commit history review. A missing noindex tag removal can take days to find.

Finally, confirm manual actions even if you think the drop is technical. A manual action for unnatural links can hide behind a clean coverage report. Check the Manual Actions panel directly.

Early in my career, I assumed every traffic drop was a Google penalty. I spent weeks rewriting content and building links when the cause was a 404 error on the sitemap file after a migration. The sitemap was not submitted, so Google stopped indexing new pages. I learned to check the [sitemap](/sitemap/) status first.

I also overlooked the importance of change history. A client's site lost 50% traffic after a redesign. I ran full crawls, found duplicate content issues, and fixed them. No recovery. Then I asked the developer: "What changed on 1st June?" They admitted they moved the site to a new hosting provider and forgot to update DNS. The site was unreachable for three days. That caused a massive drop in indexing. Now I demand a full change log before I start.

I used to rely too much on automated tools. A tool flagged 500 pages as duplicates due to URL parameters. I spent hours implementing [canonical tags](/canonical-tag/). Turns out the real problem was a rel="canonical" pointing to the wrong URL on the homepage. The tool did not catch that. I now manually inspect a sample of pages for each issue.

Another mistake: ignoring JavaScript rendering. I thought all modern crawlers could handle it, but a client's React site appeared fine in my browser. Googlebot, however, saw an empty page. The [JavaScript SEO](/javascript-seo/) guide taught me to use the URL Inspection Tool's "Live Test" to see what Googlebot actually sees. That caught the problem immediately.

I also changed my mind about competitor analysis. I used to think it was optional for a forensic audit. Now I consider it mandatory. If no internal change correlates, the external shift is almost always the answer. I add a competitor gap analysis to every forensic workflow.

Finally, I used to rush into fixes. Now I validate the root cause by reverting the suspected change first, if possible. One client wanted to update 200 product pages. I said: "Revert the last deploy." Traffic recovered within 48 hours. Patience saves time.