A single domain firing sitewide links with commercial anchors — “sponsored by carchex extended auto warranty… get a free quote” appears 22,494 times. Paid-placement anchors at this scale are exactly what Google’s spam policies target. First step: verify whether the links carry rel=sponsored — if CarChex controls the placement, fixing the rel attribute is cleaner than disavowing. If not, this is the most egregious candidate in the profile.
posgradount.edu.peDISAVOWAS 6Trust 6372 linkssince Dec 2025
Half of a simultaneous spam burst from two Peruvian .edu domains that began linking eight months ago. Classic negative-SEO or scraper pattern; no legitimate explanation.
snap.edu.peDISAVOWAS 2Trust 2372 linkssince Dec 2025
The other half of the same burst — identical link count, identical timing. File together.
Why so short a list? Google already ignores most low-grade spam on its own; disavowing marginal links buys nothing and risks clipping quiet value. Only patterns that are both manipulative and at scale make the file. The low-value directory tier (webguiding.net, 1directory.org and similar) is deliberately left alone.
Healthy links — top 20
Authority Score 77–100 · the equity worth protecting and growing▾
Neutral — monitored, no action
legitimate but low-lift patterns▾
The dealer networkKEEP~10 domainsAS 9–32100–3,100 links each
cincinnatiusedautosales.com, udriveautomobiles.co, topvalueauto.com, vipautosales.com and similar — plus automanager.com (dealer-website software, AS 32). The pattern suggests a CarChex widget embedded in AutoManager dealer-site templates: a legitimate partnership footprint. Low authority per link, but natural and topically relevant.
Low-value but not manipulative at scale. Google discounts these on its own; disavowing buys nothing and adds risk. Left alone, re-checked quarterly.
3 tracked competitors·17× CarShield’s traffic advantage·#1 vs #24–29 on shared head terms·Semrush, Aug 18 2026
Head-to-head · the market in one table
Est. traffic /mo
Keywords
Traffic per keyword
Authority Score
Referring domains
Paid keywords
CarChex
12,187
10,257
1.2
33
1,582
4
CarShield
212,802
18,662
11.4
47
1,825
0
Endurance
167,118
79,693
2.1
48
4,009
~203
chaiz.com watch
8,526
13,219
0.6
—
—
0
18-month trend · us vs the leaders, month by month
Estimated organic traffic / month
Log scale — the gap is real, but so are the leaders’ declines: Endurance −44% since 2/25, CarShield −19%.
CarChexCarShieldEndurance
Organic keywords ranked
CarShield’s footprint is shrinking (43k → 19k) while its traffic holds — position concentration. Endurance swings with content pruning.
CarChexCarShieldEndurance
Domain Authority history isn’t available from the data source — current Authority Scores: CarChex 33 · CarShield 47 · Endurance 48. AS gets re-pulled with each monthly cycle and charted from 9/26 forward.
Their top non-branded keywords — and where CarChex stands
CarShield
wins on position, not coverage — 11.4 visits per keyword vs our 1.2
The gap is position, not coverage. CarShield earns 17× the traffic on fewer than 2× the keywords — the same keywords, ranked higher. That’s why this program is weighted toward moving existing pages over producing new ones.
The leaders are drifting. Endurance’s traffic is down 44% and CarShield’s 19% since Feb 2025, while CarShield’s keyword footprint halved. The category is losing clicks to AI Overviews and the winners are consolidating — a well-executed challenger program gains share during exactly this kind of churn.
CarShield owns #1 on every head term the 9/26 money-page rebuild targets. The rebuild doesn’t need to beat them immediately — moving from #25–29 to page one captures value at every step, and theirs is a beatable sales page, not a publisher.
Endurance proves the content playbook. #7 on a 90,500/mo repair query shows a warranty brand can win informational SERPs. Our blog slate follows the pattern they validated.
The real link gap is diversity. Endurance: 4,009 referring domains. CarChex: 1,582, 41% junk-tier. The PR program targets new referring domains, not link counts.
Nobody owns the comparison SERPs. 11,000+/mo of CarShield-brand demand and 12,500/mo of Endurance-brand demand with no strong comparison content — the 11/26 and 2/27 conquest pages fill a vacuum.
Baseline: Aug 18, 2026 (PSI report run by TKX)·CWV field assessment: PASSED·Lab: 52 mobile · 47 desktop·sprints 11/26 + 12/26 · re-audit 2/27
Site performance · PageSpeed Insights baseline
The headline: real users are fine — the 28-day field data passes all three Core Web Vitals (LCP 1.9s, INP 129ms, CLS 0), so there is no ranking emergency. But the lab scores (52 mobile / 47 desktop) expose how little headroom the site has: 3.3 seconds of render-blocking requests, over half a megabyte of unused JavaScript, and ~1MB of assets with no cache lifetime. The 11/26–12/26 sprints spend down that debt so the field numbers stay green as traffic and content grow — and it fixes the two insecure (non-HTTPS) requests dragging Best Practices to 50.
PageSpeed Insights baseline · Aug 18, 2026. Refreshed with each monthly cycle; re-audit lands 2/27.
—
Core Web Vitals — real-user field data (runs with the test)
LCP · loading
—
INP · interactivity
—
CLS · stability
—
Lab metrics
LCP
—
FCP
—
TBT
—
CLS
—
Speed Index
—
Top opportunities
Run the test to populate. Until then, what live verification already tells us: the homepage ships ~190KB of raw HTML before assets (/auto-warranty/: 175KB) on Avada/Fusion templates; guide images are full-size 2022 JPEGs with no WebP variants; and the two WordPress installs behind the proxy carry two different caching behaviors. Sprint targets: mobile LCP < 2.5s, CLS < 0.1, INP < 200ms, Lighthouse 75+.
8,763 of CarChex’s tracked SERPs carry an AI Overview·~241 existing citations (re-verified at 9/26 setup)·benchmark: LLMrefs
Pulls your tracked keywords, their generated prompts, and positions straight from LLMrefs. Data may still be gathering — re-run any time.
How the benchmark is read
This is the baseline, not the target. Each keyword generates buyer prompts; LLMrefs asks the AI engines and records whether CarChex is cited and at what position. Movement on this scoreboard gets reported next to rankings each month — the program feeds it with llms.txt (shipped 9/26), FAQ/Article schema (10/26), answer blocks in every rebuild, and key-facts formatting (1/27).