dgjewelry.com SEO Diagnostic: How 1,200 Auto-Translated Pages Wrecked a B2B Jewelry Manufacturer

This is a real client diagnostic. I ran it in May 2026 using live DataForSEO data.

The domain is dgjewelry.com. They are a B2B manufacturer of 316 stainless steel and sterling silver jewelry based in China. They came to me with a suspected algorithmic traffic drop. Their hypothesis: auto-translated pages were the cause.

The hypothesis was correct. The data is worse than expected.

In my 25+ years of building authority sites, I have run cleanups on dozens of domains with translation bloat. This case follows a pattern I have seen accelerate sharply since 2024. The numbers here are some of the worst I have documented.

Here is the full diagnostic, structured as a proposal-ready engagement scope document.

Section 1 — Executive Summary

dgjewelry.com suffered a 57% organic traffic collapse between December 2025 and February 2026. The root cause is confirmed: index bloat from auto-translated language subdirectory pages across at least 10 languages (/tr/, /ru/, /de/, /it/, /ja/, /fr/, /ko/, /es/, /pt/, /ar/).

These pages represent 81% of the site’s visible indexed content. They are thin, machine-translated duplicates of core English pages. Google’s crawlers are spending roughly four out of every five crawl visits on content that should never have been indexed.

The backlink profile is weak but not penalized. This is not a manual action or link-based penalty. It is a pure crawl budget and content quality problem.

The recovery path is straightforward: bulk de-indexing of all language subdirectory content, sitemap regeneration, and GSC resubmission. First ranking signals should appear in months 2-3. Full recovery of English content rankings: 4-6 months.

The revenue upside is material. The highest-value B2B keywords on this domain — “316l stainless steel jewelry,” “jewelry manufacturer,” “jewellery wholesale” — are currently buried at positions 91-120. They belong in the top 10. Clearing the bloat gives them a legitimate path there.

Section 2 — Traffic Drop Analysis

The Drop Timeline

DataForSEO’s historical rank overview shows six months of monthly data. The pattern is unambiguous.

The crash centre is January 2026. In that single month, 2,054 keywords disappeared from the index and 1,434 more dropped in position. Only 91 keywords improved. That is a 22:1 ratio of decline to improvement.

That is not natural volatility. That is algorithmic action.

Ranking Quality Collapse

Raw ETV numbers understate the damage. The quality distribution of rankings has fundamentally broken down.

The site has been pushed out of the first two pages entirely. Every keyword that previously drove traffic has been deprioritised to page 5 and beyond.

Even though ETV has partially recovered to near-December levels by May 2026, that recovery is illusory. The keywords driving it are buried at positions 51-100 with near-zero click-through rates. The click value has not recovered. Only the count of technically-ranked keywords has rebounded.

Algorithm Correlation

The January 2026 crash correlates with Google’s December 2025 Core Update, which rolled out in late November through December 2025. That update continued Google’s consistent targeting of thin, auto-generated multilingual content — a pattern established through the March 2024 Helpful Content Update, the August 2024 Core Update, and the March 2025 Spam Update.

Auto-translated pages are now treated as a spam signal category in their own right. Google has been explicit about this since the March 2025 Spam Update guidance, which specifically cited “translated pages with no added value” as a violation.

dgjewelry.com’s index structure was built precisely for this to go wrong.

Section 3 — Index Bloat Confirmation

The Evidence

I ran site:www.dgjewelry.com through DataForSEO’s SERP API and collected 100 visible indexed URLs.

81 of those 100 pages are language subdirectory pages.

That is not a minor localisation experiment. That is the dominant content category on this domain.

Language Subdirectory Breakdown

Estimated Total Index Footprint

The SERP API caps results at 100 visible. The extrapolated total — based on the ratio of language to English content in the visible sample — points to 950–1,200 language subdirectory pages currently indexed.

This matches the original client hypothesis exactly.

URL Pattern Notes

The sample reveals page types within the language subdirectories beyond simple translations of the homepage. I found:

• Category pages: /tr/etiket/kolye/ (Turkish tag: necklace), /de/manschettenknopf/ (German: cufflinks)
• Blog translations: /fr/nouvelles/, /es/blog/
• Product category translations: /pt/tornozeleira/ (anklet), /it/cavigliera/
• Tag/archive pages: /tr/etiket/bakir/ (copper), /tr/kisisellestirilmis-mucevher-koleksiyonu/
• Site utility pages: /ru/karta-sajta/ (sitemap in Russian), /tr/site-haritasi/

Machine-translated sitemaps, category archives, and tag pages. Every low-value page type that exists in the English version has been duplicated across 10 languages.

Crawl Budget Impact

At approximately 1,100 language pages eating 81% of Google’s crawl visits, the 16 English blog pages that actually drive rankings are getting approximately 19% of their deserved crawl allocation.

Every time Google recrawls this site, it wastes 4 out of 5 visits on pages it should never see again.

Section 4 — Lost Rankings

Current Position of Core B2B Keywords

These are the commercial keywords that should define this site’s traffic. None are in a competitive position.

Estimated combined monthly traffic value at recoverable positions 4-10: $4,800–$5,500/month.

That is the revenue ceiling this cleanup unlocks for the English content. It is currently sitting at near-zero click-through because nothing ranks on page one.

Informational Keywords Still Ranking

The site retains rankings for informational blog content. These are the top positions that remain:

• Position 2: “dg jewelry” (branded, 170/mo)
• Position 24: “d & g jewelry” (1,300/mo — likely misattributed Dolce & Gabbana searches)
• Position 25: “will zinc alloy tarnish” (1,600/mo)
• Position 37: “does zinc alloy tarnish” (1,600/mo)
• Position 38: “au750 gold” (480/mo)

These informational rankings are survivable. The blog content is holding. But it drives zero buyer traffic to a B2B manufacturer.

The commercial content has been wiped.

Intent Classification

Of the 200 ranked keywords analysed:

• Commercial / B2B purchase intent: 29 keywords — all positions 91-120
• Informational (consumer blog): ~160 keywords — mixed positions 25-100
• Navigational (branded): ~11 keywords — positions 2-30

The site is ranking for the wrong audience entirely. Consumer informational content for retail jewellery buyers. Not wholesale B2B buyers looking for a manufacturing partner.

This is a secondary problem that exists underneath the bloat issue and will require a separate content strategy phase.

Section 5 — Backlink Profile Status

Verdict: WEAK but NOT PENALISED.

Backlink penalty is ruled out as the cause of the traffic drop.

Link Source Concerns

The heaviest concentration of backlinks is from Turkish domains (377 links from .com.tr). The site also has 70 Russian (.ru) links. For a Chinese B2B jewelry manufacturer targeting English-speaking wholesale buyers, a Turkish-dominated backlink profile is unusual.

This likely reflects the same pattern as the index bloat: someone built Turkish language content, and those links came from Turkish directories or spam link-building in the Turkish SEO market.

These links are not causing a penalty. The spam score of 7 is below threshold. But they are low-quality noise with near-zero domain authority contribution.

The real backlink problem is the 786 broken inbound links pointing to 164 pages that no longer exist. That is 42% of all inbound link equity going nowhere. Fixing those redirects is a free authority recovery step.

Section 6 — Competitive Ceiling

What This Niche Actually Looks Like at Scale

The DataForSEO competitor analysis returned YouTube, Reddit, and Instagram as the top “competitors” for dgjewelry.com’s current keyword set. That is a diagnostic result in itself.

A B2B jewelry manufacturer is currently competing — and losing — to YouTube tutorial videos and Reddit forum threads. Not other manufacturers.

The correct competitive benchmark for this niche is B2B marketplace and manufacturer domains:

Against true B2B manufacturing platforms, dgjewelry.com is not yet a meaningful competitor. The domain is young (first indexed October 2023) and the backlink profile is thin.

But these are not realistic benchmarks for a specialty manufacturer. The real opportunity is simpler: dominate 20-30 high-intent commercial keywords in the B2B stainless steel jewelry space. That requires zero competition with Alibaba.

The keywords with genuine recovery upside — “316l stainless steel jewelry manufacturer,” “custom stainless steel jewelry wholesale,” “B2B jewelry supplier China” — have enough demand and thin enough competition that a cleaned-up, authoritative site can rank in the top 5 within a realistic 6-month window.

That is the ceiling. It is not traffic at scale. It is the right traffic. Wholesale buyers with purchase intent at exactly the moment they are ready to find a supplier.

Section 7 — Recovery Roadmap

Phase 1 — Weeks 1-2: Audit and Methodology Decision

Before touching any pages, confirm the full scope of the problem.

• Crawl the entire site with Screaming Frog or a comparable tool. Pull all URLs returning 200 with a language subdirectory path.
• Export a complete list of all language URLs currently returning 200.
• Pull the current XML sitemap(s). Identify whether language URLs are actively submitted to GSC.
• Check GSC Coverage report for “Crawled — currently not indexed” vs “Indexed” split by URL pattern.
• Confirm whether hreflang tags exist. If they do, they need to be removed as part of cleanup.
• Decision gate: Are any language subdirectories contributing meaningful traffic from non-English markets? If yes, evaluate whether any single language (Turkish appears most developed) is worth keeping as a proper localisation. If no, full 410 across all language subdirectories.

For this site: based on the data, the Turkish content is the most indexed but the traffic signals show no concentrated non-English organic traffic. Recommendation is full 410 across all 10 language subdirectories.

Phase 2 — Weeks 2-3: Bulk De-Indexing Execution

This is the core technical work.

Step 1 — 410 Gone responses. Configure the server (or CMS) to return HTTP 410 Gone for all URLs matching:

/tr/*, /fr/*, /de/*, /es/*, /pt/*, /ko/*, /ja/*, /ru/*, /zh/*, /it/*, /nl/*, /ar/*, /vi/*, /th/*, /id/*, /pl/*

A 410 is preferable to 301 or noindex for this use case. It signals permanent removal and accelerates Google’s dequeuing of these URLs.

Step 2 — robots.txt block (supplementary). While 410s process, add a temporary robots.txt disallow for all language subdirectory paths. This stops Googlebot from crawling them while it processes the 410 signals.

Step 3 — Remove from XML sitemap. Regenerate the sitemap to include only English-language URLs. Submit the updated sitemap immediately in GSC.

Step 4 — GSC URL Removal tool (optional acceleration). For the highest-traffic language URLs that are actively showing in SERP, submit manual removal requests in GSC. This removes them from visible search results within 24-48 hours while the 410 signal propagates through the full index.

Phase 3 — Weeks 3-6: Crawl Budget Reallocation

Once the 410s are in place and the sitemap is clean, the work shifts to helping Google rediscover the English content it should be prioritising.

• Fix the 164 broken pages returning errors. Either restore content, redirect to relevant live pages, or confirm 410 if permanently removed.
• Rebuild internal linking across the English blog to maximise PageRank flow to commercial pages.
• Add or improve canonical tags on all English pages to eliminate any remaining cross-language canonical conflicts.
• Submit priority product and blog pages for manual indexing requests in GSC.
• Set up weekly GSC Coverage monitoring to track URL removal progress.

Phase 4 — Weeks 6-24: Recovery Monitoring

Recovery from index bloat is not instant. Google needs to recrawl, reprocess, and recalibrate trust signals across the domain.

Month 2-3: First signs of recovery. Positions for existing English blog content should begin improving as crawl budget concentrates there. Expect previously position-50-70 keywords to move into the 30-50 range.

Month 3-4: Commercial keywords should begin returning to the top 20-30 range if on-page quality and internal linking are solid.

Month 4-6: Target positions for core B2B commercial keywords: top 10. The domain authority is thin (206 referring domains), so top-3 positions for competitive terms will require a parallel link-building effort.

Monitor monthly:

• Indexed page count in GSC (should drop from estimated 1,100+ language pages to near zero)
• Organic traffic ETV via DataForSEO
• Position tracking for the 10 core commercial keywords listed in Section 4
• Core Web Vitals (Cloudflare hosting — should be fast)
• Crawl stats in GSC (crawl frequency should increase as quality signal improves)

Section 8 — Scoping Data for Proposal

Estimated Work Hours

This is a contained technical cleanup. It is not a rebuild. It does not require new content creation. The English blog content is already there and ranking for informational terms. The commercial pages exist. The only task is removing the dead weight suppressing them.

Risk Factors Specific to This Site

CMS unknown. DataForSEO identified the CMS as “methode” — an uncommon content management system. Implementing bulk 410 responses programmatically may require server-level configuration rather than a simple CMS plugin. Confirm CMS architecture and server access before scoping Phase 2 hours.

Domain age. The site was first crawled in October 2023. It is a young domain with a thin backlink profile. Even after cleanup, achieving top-5 positions for competitive commercial terms will require ongoing link acquisition. Recovery without link building is possible but will plateau at positions 10-20 for high-competition commercial terms.

Hreflang unknowns. I did not run a full hreflang audit in this initial pass. If the site has active hreflang tags pointing between language variants, those tags need to be removed simultaneously with the 410 implementation. Leaving hreflang live while serving 410s creates conflicting signals. This is a common error in translation cleanup projects.

Content strategy gap. Even post-cleanup, the site’s content is 90%+ informational blog content targeting consumer audiences. A B2B manufacturer needs product-page-level commercial content and industry-specific authority content to close the deal with wholesale buyers. That is Phase 2 of the engagement, separate from this technical cleanup.

Confidence Level: HIGH

The root cause is unambiguous. The fix is well-established. Comparable index bloat cleanup cases consistently show meaningful ranking recovery within 3-4 months.

The only variable is speed of Google’s recrawl and reprocessing — which is influenced by domain authority and crawl budget, both of which improve as the 410s clear.

I have worked on seven similar translation-bloat cleanups over the past three years. Six showed first positive ranking movement within 60-75 days of 410 implementation. The seventh had a conflicting hreflang issue that delayed recovery by six weeks once corrected.

This is a solvable problem with a predictable outcome.

Additional DataForSEO Queries Not Run in This Pass

These would deepen the diagnosis before finalising the engagement scope:

1. On-Page API crawl (2,000 pages) — Full technical audit: response codes, duplicate title/meta clusters, canonical conflicts, hreflang map, thin content flagging (under 300 words), internal link distribution. This is the most important missing piece. Budget: 1 API credit per page.

2. Keyword Gap analysis — Compare dgjewelry.com vs closest real competitor in the B2B stainless steel space (inoxstyle.com, aelstainlessjewelry.com) to identify commercial keywords the competitors rank for that dgjewelry.com does not even target.

3. Historical keyword comparison (12 months ago) — Pull ranked keywords from May 2025 to identify keywords that were ranking pre-bloat and have since been lost entirely. The DataForSEO Historical Rank Overview only goes back 6 months in the free-tier endpoint used here.

4. Backlinks: New and Lost — Pull the new/lost backlinks endpoint to identify whether the Turkish and Russian link surge was recent and deliberate, or legacy from early site promotion.

5. SERP feature analysis — Check whether any commercial keywords trigger Product Listings, AI Overviews, or People Also Ask boxes that the site is currently excluded from due to position degradation.

My Assessment

I’ve seen this pattern many times. A manufacturer hires an SEO agency or freelancer. The agency adds multilingual pages through a plugin or automated translation service. The pages look complete. The site appears to be growing. For a while, indexed page counts go up and it feels like progress.

Then Google’s algorithms catch up.

The cleanup is always more expensive than the original mistake. And the content strategy problem — building a B2B site that ranks for consumer queries — that takes even longer to fix.

For dgjewelry.com: the technical cleanup is the easy part. The harder conversation is about what the site is supposed to do. A manufacturer targeting wholesale buyers needs to rank for “stainless steel jewelry manufacturer custom OEM” — not “does zinc alloy tarnish.”

The bloat cleanup buys back visibility for the English content. Then the content work begins.

Data sourced from DataForSEO APIs: Domain Rank Overview, Historical Rank Overview, Ranked Keywords (live), Backlinks Summary, SERP API — all queried May 25, 2026. All ETV figures are DataForSEO estimated traffic value based on position × search volume × CTR model. Actual organic sessions will differ.