Domain Quality
- Authority and traffic relative to a niche-calibrated floor
- Stability through Core Updates and AIO rollouts
- Resistance to traffic-manipulation patterns
- Health of the keyword profile that drives the traffic
Methodology · Proprietary to Earned Media
The Provenance Score. The rubric behind every link.
Most agencies build links on judgment that drifts — one team member's "high quality" is another's "spammy". We replaced judgment-by-feel with a 50-point rubric run on every prospect, every time. The result: the same standard on week 1 and week 52, on every account, regardless of who's prospecting.
Why it exists
Quality is a system, not a vibe.
Eight years of link building taught us the same lesson Google learned a decade ago: human judgment about "good links" looks consistent on day one and drifts hard by month six. People get tired. Quotas get tight. The DR floor quietly shifts. The Provenance Score exists to remove the drift — the rubric scores prospects, the scorer just observes.
The rubric is proprietary. The criteria are not. We publish what we score; we keep the anchors, weights, and edge-case logic internal — the same way a credit-rating agency publishes its categories but not its model. Below is the public face: enough to show you the work, not enough to copy.
The five criteria
What the rubric scores.
Editorial Standards
- Whether the page reads like a real publication
- Discipline of outbound link patterns across the site
- Integrity of content production — human vs slop
- Consistency of editorial conventions and disclosure
Contextual Fit
- Topical match between the linking page and your target page
- Category alignment between the publisher and your industry
- Keyword overlap with your domain
- Audience overlap with your buyer profile
Author E-E-A-T
- Real, traceable authors behind the byline
- Recognised expertise in the relevant category
- Editorial provenance — bylines elsewhere, real history
- Entity resolution signals readable by Google and LLMs
LLM & AI Presence
- Whether ChatGPT, Perplexity and AI Overviews already cite the publisher
- Domain visibility in the corpora LLMs train on
- Strength of structured-data and entity signals
- Presence in category-specific buying-intent queries
14-point Triage runs first
- Hard-signal screen for casino, gambling and adult footprints
- Paid-placement detection on contributor and contact pages
- PBN and IP-cluster cross-references
- Subdomain-blog and orphan-folder traps
Score bands
Three outcomes. One standard.
The Provenance Score — outcomes
Out of 50Below threshold
Rejected
Doesn't get pitched. We log the score, the failed criteria, and the screenshots — and we don't re-prospect the domain for 90 days. Most prospects land here, by design.
Pass band
Approved
Solid editorial placement worth pitching. Reported to you with the QA log attached — including the score and which criteria contributed to it. No mystery placements.
Premium band
Tier 1
The placements we screenshot. High-authority publisher, on-topic, real authors, already cited by LLMs. The links that compound across both engines instead of just one.
Where it shows up in your account
Built into the work, not bolted on.
Every prospect, scored before outreach.
- Triage check log on every prospect we touch
- Provenance Score stored against every approved domain
- Rejected prospects logged with reason and re-prospect lockout
Metrics captured the day the link goes live.
- Ahrefs DR, OT, traffic chart at time of placement
- Screenshot of the live page with the link in context
- Final Provenance Score archived against the placement
Every live link re-scored every 90 days.
- Domain re-checked against the current rubric
- Score drift surfaced before Google notices
- Disavow recommendations if a publisher degrades materially
You can defend every link in your profile.
- Each placement comes with its scorecard and evidence
- Aggregate score distribution reported monthly
- Premium-band ratio tracked as a portfolio quality metric
What we won't show
Why the anchors stay private.
Scoring criteria are gameable when they're public.
The moment a low-quality publisher knows the exact thresholds, the exact patterns we screen for, and the exact LLM-citation tests we run, they engineer their site to pass them. The rubric works because the bar moves faster than the workaround.
It's eight years of pattern recognition.
The anchor descriptions are calibrated against thousands of placements and dozens of Core Updates. It's the part of our work that compounds — and the part competitors would copy first if we handed it over.
The receipts on your account.
Every score, every criterion contribution, every screenshot. You see the rubric applied to your prospects in full detail — you just don't see the rubric applied to other agencies' prospects.
Free Provenance audit
Send us your last 10 links. We'll score them on Provenance.
CSV or list of URLs. We'll run the 14-point Triage and the Provenance Score on each, and send back a written audit telling you which would pass our rubric, which we'd kill, and why — with the screenshots. You'll know in 48 hours whether your current provider is doing real work or quietly failing.
Request the audit →
No sales call required for the audit itself.
Related