Google’s August 2026 spam update hit while a lot of small sites were cruising on templated pages and syndicated content. Naveed Ganatra’s video is a good reminder to run a focused check instead of flailing: https://www.youtube.com/watch?v=6oOiOmOsOEo. One week in, here’s how I’d triage a small site without burning the house down 🔧.
If you’re seeing volatility, don’t guess. Measure, isolate, and fix patterns that actually look like spam footprints. I’ll show you where to look and how to decide what to prune vs. improve.
Short version first—then details.
- Map the damageSegment by page type and intent before you touch content.
- Kill obvious spam patternsDoorways, spun hubs, junk UGC. Don’t half-fix—remove or quarantine.
- Rebuild trust signalsBrand SERP, GBP, reviews, authorship, and citations.
- Watch lagging metricsUse GSC, logs, and branded CTR to judge recovery, not just sessions.
What this spam update likely hit (and why you should care)
Google’s spam updates historically clamp down on patterns, not one-off posts. That means:
- Scaled, near-duplicate or templated pages with thin differentiation 💥
- Doorway structures (city/service pages that funnel to the same lead form)
- Low-value UGC or forum spam bleeding into indexable pages
- Link schemes and junk affiliate/partner footprints
- AI-spun or stitched content that’s not adding unique evidence
If any of that describes parts of your site, assume scrutiny and audit by pattern, not URL. If you want my step-by-step rapid response, start with my 72-hour playbook: spam update triage: fix doorways, brand SERP, and AI surfaces and the fuller brief: August 2026 spam update + Gemini 3.7: my 72‑hour triage and what to fix now.
The five checks I’d run this week (in order)
1) Segment losses by page type and intent
Why: You need to see which pattern got hit. Pull GSC by URL folder and template. Group by:
- City/service pages vs. core service vs. blog vs. location hub
- Informational vs. transactional queries
- Pages created by the same generator/template
Outcome: A shortlist of patterns to fix. No pattern, no fix.
2) Nuke doorway and spun patterns first
Why: These are bright-red spam signals. If multiple city pages share 95% text and the same CTA, they’re liabilities.
- 410/redirect clusters that add zero unique value
- Consolidate to a single robust service page per region with real proof (work samples, photos, quotes, permits)
- If you must keep coverage, add on-page evidence per locale (PPC receipts, invoices redacted, geo EXIF on images, local testimonials) 🧭
3) Clean UGC and thin hubs
Why: Indexable tag pages, empty category hubs, and comment spam drag down trust.
- Noindex thin hubs; paginate or consolidate
- Moderate or auto-nofollow UGC links
- Rate-limit and archive dead forum threads
4) Fix brand surfaces you actually control
Why: When organic dips, strong brand signals cushion click-through and trust.
- Google Business Profile: match categories, add services, Q&A, and fresh photos; report obvious competitor spam
- Brand SERP: title/description, sitelinks hygiene, and favicon clarity—see: fix favicons and protect brand CTR 🔎
- Author and entity markup: tie content to real people and org facts; schema won’t save junk, but it helps machines disambiguate—proof here: does schema markup help ai search? practical proof and metrics
5) Evidence-first refresh on what remains
Why: Recovery requires unique, verifiable value.
- Replace generic claims with artifacts: screenshots, data tables, invoices (redacted), EXIF-stamped images, mini-case studies
- Add comparative context and decision criteria, not fluff
- Tighten internal links by task/outcome, not just keywords 🔗
How to measure impact without panicking
- Use GSC as your primary signal. Sessions can be noisy for weeks.
- Compare branded vs. non-branded separately. If branded holds while non-branded drops, quality/intent mismatch is the issue, not your reputation.
- Track by pattern: build simple dashboards per folder/template. If one template’s recovering, replicate its fixes.
- Check server logs for crawl behavior. Post-fix, you want deeper recrawls of cleaned sections.
- Annotate everything. If you need help structuring this, my SEO Audit process is built around repeatable, pattern-level measurement.
Stop trying to fix “the algorithm.” Fix the footprint that made your pages look interchangeable.
When AI content is the question, make evidence the answer
I don’t care if text started with a human, a model, or a legal pad. Google cares whether the page is useful, verifiable, and distinct. If a section reads like filler, replace it with proof: workflows, images you own, localized constraints, pricing realities, and before/after outcomes 🧪. And if a whole cluster is beyond saving, remove it cleanly and focus on pages that can earn links and mentions for the right reasons.
What I’d skip doing this week
- Disavow theater. Unless you’re in a paid link mess you created, this rarely moves the needle in a spam update.
- Sitewide rewrites. Fix one pattern well, then scale the approach.
- Publishing more of the same. Pause the factory. Rebuild on evidence.
Should I delete all my AI-written posts?
No. Keep what you can harden with evidence and unique angles; remove only what you can’t justify.
How long does recovery take after a spam update?
Weeks to months. You’ll usually see signals within recrawl cycles if fixes are real.
Do redirects save doorway pages?
Only if the destination is genuinely stronger. Redirecting junk to junk keeps the footprint.
What’s the fastest win for a local SMB?
Consolidate thin city pages into a strong regional service page and strengthen your GBP.
Bottom line: updates expose weak patterns. If you run this five-check triage and rebuild with evidence, you’ll come out leaner and safer for both Google and AI surfaces. If you want a battle-tested workflow for AI surfaces specifically, start here: How to show up in AI search (my field playbook after Marvin Chow’s chat). That’s how I connect the clean-up work you do today to durable SEO/GEO/AI visibility tomorrow.
