How xark.io scores every publisher across five dimensions before recruiting, commissioning, or scaling — and why raw traffic metrics alone miss 60% of program performance drivers.
Quick Answer
How does Xark evaluate publisher quality for affiliate programs?
Xark's Publisher Quality Score (PQS) rates publishers on five dimensions: traffic quality (DA, source mix, engagement), content relevance (vertical match, content depth), audience alignment (ICP match, geography), historical CVR (actual conversion data or EPC benchmark), and compliance record (FTC disclosure, attribution hygiene). Publishers scoring 85+ receive elevated commissions and dedicated support. Those below 40 are declined regardless of audience size.
The Publisher Quality Score: How We Evaluate Affiliate Partners
Most affiliate programs recruit publishers the wrong way: they filter by audience size, check for category relevance, and send an outreach email. The result is a publisher roster full of activated partners who never produce meaningful GMV — and no systematic way to know why.
Xark's Publisher Quality Score (PQS) is our answer to this problem. It's a five-dimension evaluation framework that scores every publisher before we recruit them, before we increase their commission, and before we scale budget toward them. It's the difference between intuition-based program management and a repeatable system.
Why Raw Traffic Metrics Fail as Filters
A publisher with 2 million monthly readers who drives a 0.2% CVR is worth less to your program than a publisher with 80,000 readers driving a 3.1% CVR. Traffic volume is a vanity metric in affiliate marketing. What matters is whether the publisher's audience converts, whether their content is relevant enough to drive intent, and whether they follow the program rules.
Programs that filter only on audience size consistently over-recruit T1 publishers who don't convert and under-recruit T2 publishers who do. PQS corrects for this by weighting five factors that actually predict program contribution.
The Five PQS Dimensions
Dimension 1: Traffic Quality (0–20 points)
Traffic quality evaluates the source and composition of a publisher's audience — not just how many people visit, but who they are and how they arrive.
What we assess:
- ◆Domain Authority (DA) as a proxy for organic search credibility
- ◆Traffic source split: SEO-driven content publishers convert 2.3x better than social-driven ones for considered purchases
- ◆Bounce rate and session duration: publishers with 3+ minute average sessions indicate engaged, high-intent audiences
- ◆Mobile vs desktop split: desktop-heavy publishers (common in comparison and review content) often drive higher AOV
Scoring rubric:
- ◆18–20: DA 50+, 70%+ organic search, <40% bounce, 3+ min sessions
- ◆12–17: DA 30–50, 50–70% organic, balanced traffic sources
- ◆6–11: DA 20–30, mixed sources, moderate engagement
- ◆0–5: Low DA, primarily social or paid traffic, high bounce
Why it matters: A fashion influencer with 1M followers and 100% social traffic will drive a 0.4% CVR on a consumer electronics product. A review site with 80K monthly readers and 75% organic search will drive a 2.8% CVR on the same product. Traffic quality predicts conversion before the publisher has ever promoted your brand.
Dimension 2: Content Relevance (0–20 points)
Content relevance measures how closely the publisher's primary content category matches your product vertical — and whether their editorial stance creates genuine purchase intent.
What we assess:
- ◆Vertical match: does the publisher cover your product category?
- ◆Content depth: do they publish long-form buyer's guides and comparison articles (high intent) or short-form lifestyle content (lower intent)?
- ◆Content recency: are they publishing in your category actively, or was the last relevant article from 18 months ago?
- ◆Brand safety: would your brand be comfortable appearing alongside their content?
Scoring rubric:
- ◆18–20: Primary vertical match, long-form comparison/review content, published in category last 30 days
- ◆12–17: Partial vertical match, mixed content formats, recent category content
- ◆6–11: Adjacent vertical, generic lifestyle or deal content
- ◆0–5: Off-vertical, thin content, content older than 6 months in category
Why it matters: A publisher writing detailed air purifier comparison articles converts on air purifier affiliate links at 4× the rate of a home improvement blogger who covers air purifiers occasionally. Relevance predicts intent.
Dimension 3: Audience Overlap / ICP Alignment (0–20 points)
Audience overlap evaluates whether the publisher's audience matches your ideal customer profile (ICP) — demographics, purchase behavior, income range, and geographic market.
What we assess:
- ◆Audience demographics: age, gender, income — pulled from SimilarWeb or reported publisher media kits
- ◆Geographic concentration: US-focused audiences for US programs, EU-weighted for European expansion
- ◆Purchase intent signals: does the publisher's audience actively research before buying (high intent) or browse passively (low intent)?
- ◆Competing product overlap: does the publisher heavily promote direct competitors? This indicates audience fit but creates loyalty conflicts.
Scoring rubric:
- ◆18–20: Demographic match >80%, primary US/target market, clear buyer intent, limited competitor promotion
- ◆12–17: Partial demographic match, some geographic spread, moderate intent
- ◆6–11: Adjacent demographic, broad geographic, low intent signals
- ◆0–5: Mismatched demographics, wrong geography, passive audience
Dimension 4: Historical CVR (0–20 points)
Historical conversion rate is the most predictive single metric in the PQS framework — but it's only available for publishers who have promoted in your category before.
What we assess:
- ◆CVR from affiliate platform data: if the publisher is active in your category on any network, we pull their category EPC and reverse-engineer CVR
- ◆Content publisher CVR benchmarks by format: review articles (2.5–4%), comparison articles (3–6%), gift guides (1.5–3%), social posts (0.5–1.5%)
- ◆Attribution window utilization: publishers whose CVR improves significantly with longer windows indicate high-consideration traffic
Scoring rubric:
- ◆18–20: CVR 3%+ (or EPC 40% above category benchmark), clear track record
- ◆12–17: CVR 1.5–3%, consistent performance history
- ◆6–11: CVR 0.8–1.5%, limited history or mixed category data
- ◆0–5: CVR <0.8%, no category history, or primarily coupon-type traffic
For new publishers without history: we default to Dimension 2 (content relevance) as a proxy and assign a conservative 10-point starting score, adjusted after the first 60-day performance cycle.
Dimension 5: Compliance Record (0–20 points)
Compliance evaluates whether the publisher follows program rules — FTC disclosure requirements, attribution best practices, and content accuracy.
What we assess:
- ◆FTC disclosure: is affiliate relationship clearly disclosed on all content?
- ◆Attribution hygiene: do they use unique tracking links, or do they resell traffic through sub-networks?
- ◆Coupon policy compliance: do they publish only authorized promo codes, or source codes from unapproved channels?
- ◆Past program history: have they been terminated from other programs for compliance violations?
Scoring rubric:
- ◆18–20: Full FTC compliance on all content, clean attribution, no compliance history
- ◆12–17: Minor disclosure gaps, corrected quickly, no material violations
- ◆6–11: Occasional compliance issues, responsive to corrections
- ◆0–5: Repeated violations, coupon code misuse, sub-affiliate opacity, or prior program termination
The PQS Scoring Matrix
| Score Range | Tier | Action |
|-------------|------|--------|
| 85–100 | Elite | Elevated commission + dedicated support + early launch access |
| 70–84 | Premium | Above-market rate + quarterly check-ins |
| 55–69 | Standard | Market-rate commission + standard onboarding |
| 40–54 | Conditional | Market rate, 60-day performance review |
| <40 | Do Not Recruit | Decline or deprioritize outreach |
How PQS Changes Program Outcomes
Programs that screen publishers through PQS before recruitment see three measurable improvements:
1. Higher publisher activation rates: Publishers scoring 70+ activate at a 73% rate within 30 days of approval, versus a 38% activation rate for the general publisher population.
2. Better EPC from day one: Publishers recruited based on PQS score generate 1.8x the EPC of publishers recruited by audience size alone, because they're better matched to the program's conversion profile.
3. Lower publisher churn: PQS-recruited publishers have 55% lower 90-day churn than unscreened publishers. Churn is expensive — it resets the onboarding investment and reduces program GMV continuity.
Using PQS for Commission Tier Decisions
PQS is not just a recruitment filter — it is the foundation for commission tier decisions. When a publisher applies for an elevated rate, we re-run their PQS with their actual program performance data and use the score to make the commission recommendation:
- ◆PQS 85+: Justify 1–3% above baseline with guaranteed 12-month rate lock
- ◆PQS 70–84: Standard elevated tier, quarterly performance review
- ◆PQS 55–69: Market rate, performance triggers for future increases
- ◆PQS <55: No rate increase until score improves
This removes commission negotiation from the realm of gut feel and puts it in a data-driven framework that both the brand and the publisher can understand.
Xark's PQS Process
For every new publisher we evaluate, we run PQS in two stages:
Pre-outreach (automated): We use SimilarWeb, Ahrefs, and affiliate platform data to score Dimensions 1, 2, and 4 before sending a single email. This filters our target list from 500 prospects to the top 150.
Post-application (manual): When a publisher applies to the program, we score Dimensions 3 and 5 manually using their media kit and compliance audit. This creates the final PQS score that drives commission tier assignment.
The full PQS cycle takes 20–30 minutes per publisher. On a program recruiting 50 publishers per quarter, that is 17–25 hours of structured evaluation — versus an unstructured process that takes the same time but produces worse outcomes because it lacks a repeatable decision framework.