Report card

rtings.com

Payout–Rank not computable — in every article, every pick we can price carries the same rate (23 monetized articles, 0 with two different rates to compare)

Generated 2026-09-17 under SPEC: Scoring — v0.6.12. State: Not scored — no payout variation to correlate.

Monetization exposure — what was observed

A pick whose status was not established is counted apart from a pick observed to be unmonetized. Each reason is published on its own, because they are different situations and one of them is about us rather than about the publisher:

Marked sponsored by the publisher — what was observed on those links

123 direct merchant link(s) across 123 ranked pick(s) carry this publisher’s own rel="sponsored" and point at a registrable domain that is not the publisher’s. NONE OF THEM IS BADGED, whatever follows: rel="sponsored" is the attribute the HTML specification defines for paid placement of any kind, including placement that pays no commission, so it establishes that the publisher marked the placement as paid and establishes nothing about a commission. What was observed ON those links, counted apart rather than summarised as an absence: 0 carry a real attribution subtag that this publisher’s OWN client code wrote onto the anchor while the page ran in our browser. SPEC §0.1.0 counts those as monetized (m_i = 1) and still does not badge them, because a signal derived from executed JavaScript is weaker provenance than served markup: it is admitted to a corpus statistic we publish with its inputs, and refused to an assertion made to a reader about one link. 0 carry that same attribute with the literal value "undefined" — the publisher’s own code reporting that it recognised NO affiliate program for that destination. Those are m_i = 0 on the publisher’s own signal rather than on our failure to find one. 123 carry no such attribute at all. Absence is not a statement: those are m_i = 0 because nothing was observed either way, and that is the only one of these three counts about which we are saying we did not detect a mechanism.

Declared, not observed

No pick on this publisher is priced from a merchant declared in the link.

Payout–Rank correlation

The card records Payout–Rank not computable — in every article, every pick we can price carries the same rate (23 monetized articles, 0 with two different rates to compare) as the headline it would print, and prints no median, no interval and no Kickback Score. The inputs below are published anyway, so the same computation can be run against them.

Disclosure

§1.5 is purely positional. The A/B boundary here is measured against the first OUTBOUND link, which is at or before the first AFFILIATE link, so this test is stricter than §1.5 and can only understate a grade, never overstate one.

Diagnostics published beside the figures

Count of qualifying articles with ρ > 0.99 at n ≤ 5.

Articles excluded from the pool, by reason

Reasons may co-occur on the same article, so these counts do not sum to a total.

Limitations that apply to this publisher

1. The Payout–Rank correlation is not computable for this publisher, and that is not a gap in our data. The correlation is computed within each article, and in every one of this publisher’s articles all the picks we can price carry the same commission rate — so there is no variation in payout for the ranking to be compared against. A rank correlation needs at least two different rates inside one article; this publisher’s articles supply one each. Collecting more articles of the same kind would not change that. The publisher may still pay different rates in different articles, which a within-article rank correlation does not measure.

Right of reply

State: silent. We wrote to this publisher and the seven-day window closed without a reply. Not replying is not agreement with anything on this card, and we do not present it as one.

Written to feedback@rtings.com on 2026-09-10T00:11:02Z. Basis for that address: published site contact Seven-day window closes 2026-09-17T00:11:02.000Z.

Nothing has been received from this publisher.

Confounds, stated by us, first

1. Reverse causality. A publisher may negotiate a higher rate because it already ranks a product well. The §5 event study partially addresses this by measuring ranking changes after rate changes. It does not eliminate it.
2. Category confounding. High-commission categories may differ in quality.
3. Brand-size confounding. Products with affiliate programs skew toward larger brands, which may be better on average.
4. Rate coverage. We observe roughly half the market. The rate coverage figure on this card is the measured share for this publisher.
5. Editorial firewalls are real. Where the median ρ is at or below zero, that is stated on this card in the same size type as every other number.
6. Private-network rate invisibility. Merchants on Impact, CJ, Rakuten and Partnerize keep rates private, so their picks carry r = null and are dropped from ρ. Those are systematically different merchants from Amazon’s — large retailers rather than marketplace sellers. The sign of the resulting bias is indeterminate.
7. Per-transaction earnings caps and negotiated tiers mean a published nominal rate is an upper-bound proxy for effective compensation. eBay’s published rate card carries a per-transaction cap on every category (550 USD on Electronics, 100 USD on eBay Motors and on Real Estate) and states that exceptions to the rate card exist for certain partners. Roborock publishes 4%–7% and Narwal 4%–10%, and where a given publisher’s tier falls within either band is negotiated privately. On a high-priced item a per-transaction cap can put the amount actually earned, expressed as a percentage of price, below the published rate. This is a disclosure: it moves no estimator and no r_i.

Recompute this

All publishers · Methodology