Most online casino “review” sites are paid a share of the losses generated by every player they refer, typically 15% to 35% of that player’s lifetime revenue. That single fact explains most of what looks wrong with the sector’s rankings: the operators willing to pay the highest commissions tend to cluster at the top, regardless of payout speed, complaint history, or licensing conduct. GamblScout’s rating engine removes commission as an input entirely. Every score is generated from scraped, machine-readable operator data — never from a human writer whose income depends on the placement of the “Visit Site” button.
Key takeaways
- Traditional casino affiliate revenue models pay reviewers a percentage of player losses, creating a direct financial incentive to rank paying operators favorably.
- The FTC’s 2023 Endorsement Guides revision confirms that affiliate-linked content is treated as a paid endorsement requiring clear disclosure, and its 2024 rule separately bans company-controlled review sites that falsely purport to be independent.
- UK regulators have separately sanctioned gambling affiliate sites for hiding commercial intent behind an “independent reviews” framing and for unsafe ad placement.
- GamblScout’s algorithm scores operators from scraped licensing, payout, complaint and terms-and-conditions data, with commission rate excluded from every weighting model.
- No score is editable by an operator; the only route to a higher score is a measurable change in the underlying data.
Table of contents
The commission problem in plain terms
A conventional casino review site does not charge readers. It is paid by the operators it writes about, usually through a revenue-share deal tied to how much a referred player deposits and loses over time.
affiliates in 2025 typically earn a 15% to 35% revenue share from referred players, with some networks offering commissions as high as 45%
. The online casino market itself is large enough to make this worth optimizing for:
global online casino market size was valued at $19.1 billion in 2024 and is projected to grow from $23.9 billion in 2026 to $38.0 billion by 2030
.
When a writer’s income scales with a specific operator’s player losses, the incentive is not to find the best product for the reader — it is to find the operator with the best commission terms and the highest conversion rate. We covered the structural version of this problem in The Affiliate Problem: Why You Can’t Trust Human-Written Casino Reviews. This article is about the alternative we built: a scoring system where commission rate is not a variable that exists anywhere in the model.
What regulators have already said
The conflict of interest in affiliate-funded review sites is not a theory GamblScout invented. It is now the explicit subject of consumer-protection rulemaking in the United States and enforcement action in the United Kingdom.
The FTC position
The US Federal Trade Commission revised its Endorsement Guides in 2023 and confirmed directly that
the same guidance applies anytime you endorse a product and get paid through affiliate links, and if it’s clear that what’s on your site is a paid advertisement, you don’t have to make additional disclosures
— implying that if it is not clear, disclosure is mandatory. The Commission went further in 2024 with a dedicated rule. Effective October 21, 2024, the
final rule relating to unfair or deceptive acts involving consumer reviews prohibits selling or purchasing fake consumer reviews, certain insiders creating reviews without disclosing their relationships, and creating a company-controlled review website that falsely purports to provide independent reviews
. That last clause is the closest a regulator has come to describing the exact business model this article is discussing: a site funded by the products it appears to independently rank. Violators face
civil penalties of up to $51,744 per violation
.
The UK ASA position
The UK’s Advertising Standards Authority (ASA) has taken parallel action against gambling comparison sites specifically. In one recent ruling,
the ASA upheld a complaint against a gambling affiliate’s casino comparison site after an ad appeared in search results for “help with gambling,” ruling that the safeguards it had in place, including negative keyword lists and compliance monitoring, were not robust enough
. In a related case involving a comparison site in a different vertical, the ASA found that
a paid-for search ad for a review website failed to make their commercial intent clear and falsely implied they were acting for purposes outside their trade by presenting websites used for marketing purposes as independent review sites
— the same “fake independence” problem the FTC’s rule now targets directly. Regulators on both sides of the Atlantic have converged on the same conclusion: a site that looks like independent journalism but is funded by the products it rates has to prove it, not just claim it.
How the algorithm actually works
GamblScout’s scores are produced by a pipeline, not a writer. The process runs in three stages, each documented on its own hub page for readers who want the technical detail:
- Collection. Structured and unstructured data is pulled directly from operator sites, license registers, payment processor documentation and public complaint databases. This is covered in depth on our Data Scraping & The Technical Engine hub.
- Interpretation. Free-text sources — terms and conditions, help-center wording, player complaint threads — are parsed with natural language processing to flag ambiguous wagering language, sentiment shifts in complaint volume, and inconsistent bonus terms. See Natural Language Processing (NLP) & Sentiment Analysis for the method.
- Scoring. Every data point is converted into a weighted variable inside a published scoring model, detailed on the Scoring System & Algorithmic Weights hub. Commission rate, affiliate network membership, and sponsorship status are not variables in that model. They are excluded by design, not by policy statement.
Fraud and fair-play checks — RNG certification status, licensing sanctions, and dispute-resolution track record — run as a separate module described on our Security, Fraud Detection & Fair Play hub, and feed into the same score rather than a separate “trust badge” that can be purchased.
What changes in practice
The table below sets out the structural difference between the affiliate-review model described above and the pipeline GamblScout runs. It is not a claim that every human-written review is dishonest — only that the two models respond to different incentives, which is the point regulators have also made.
| Factor | Typical affiliate review model | GamblScout algorithmic model |
|---|---|---|
| Primary revenue source | Revenue share or CPA commission from rated operators | No commission-linked funding used as a scoring input |
| Who assigns the score | A staff writer or freelancer, often anonymous | A documented, published weighting model |
| Update trigger | Manual re-review, often infrequent | Automated re-scrape on a fixed schedule |
| Disclosure of financial relationship | Varies; subject to FTC and ASA enforcement when absent | Not applicable — no per-operator commission exists |
| Ability for an operator to influence its own score | Possible through commission negotiation or content requests | Only by changing the underlying operational data itself |
What the algorithm cannot fix
Removing commission from the model removes one source of bias, not all of them. An algorithm inherits the quality of its inputs: if a license register is outdated or an operator’s terms page is deliberately vague, the score reflects that ambiguity rather than resolving it. We treat unclear source data as a negative signal in itself — opacity is scored, not smoothed over — but we do not claim the pipeline is infallible. Our approach to weighting different regions is covered separately on the Regional Deep Dives & The Global Split hub, since licensing standards and enforcement intensity vary sharply between the jurisdictions covered on our Licensing & Jurisdictions hub. A model built for one regulatory environment will misjudge another if it is not explicitly adjusted, which is why regional weighting is treated as its own documented layer rather than a single global formula.
Frequently asked questions
Does GamblScout accept payment from the operators it rates?
GamblScout does not accept revenue-share or CPA commissions tied to individual operator scores. The site’s commercial model is separate from the scoring pipeline described above, and no per-operator payment enters the weighting model used to calculate a score.
Can an operator pay to improve its GamblScout score?
No. A score changes only when the underlying scraped data changes — a license status update, a shift in complaint volume, a rewritten terms page, a certified RNG audit. There is no manual override or “sponsored placement” tier inside the scoring model itself.
How is this different from a standard “editorial independence” disclaimer?
A disclaimer is a statement; the pipeline is a structural constraint. The FTC’s 2024 rule specifically addresses sites that
falsely purport to provide independent reviews
despite disclaimers, which is why GamblScout removes the commission variable from the model rather than relying on a policy statement alone.
Do humans play any role in the process at all?
Humans design and audit the model, define which data sources count as reliable, and review flagged anomalies. No human assigns or edits an individual operator’s final numeric score.
Methodology note
This article’s scoring-model description draws on the same data signals GamblScout’s algorithm applies elsewhere on the site: scraped licensing records, payout and complaint data, NLP-parsed terms and conditions, and fraud/fair-play indicators, all combined through a published weighting model that excludes commission or sponsorship status as an input variable.
Gambling involves risk. Only play with money you can afford to lose and use the deposit limits and self-exclusion tools available in your jurisdiction.
