A content quality rubric for casino and betting pages

A good gambling page answers a specific question for a specific market, shows where its facts come from and when they were checked, and would still be worth reading if search engines did not exist. Google does not publish a single quality score, but its rater guidelines and spam policies describe what it rewards and what it penalises clearly enough to turn into a working checklist. The rubric below does that.

Our editorial standards use the same ten questions as a pre-publication check for CasinoQueryLab's own pages: a page that scores "no" on the first three should be rewritten, not tweaked.

Where the rubric comes from

Google's Search Quality Rater Guidelines, last revised on 11 September 2025, ask raters to judge experience, expertise, authoritativeness and trust, with higher standards for "Your Money or Your Life" topics: health or safety, financial security, and government and civic matters. Gambling is not listed by name. It is widely treated as YMYL-adjacent because readers risk money and wellbeing, and analysts such as Glenn Gabe describe it as an "ultra-YMYL" category given how strongly it moves in core updates. We treat that as convention rather than Google's wording.

The spam policies are more direct. Scaled content abuse covers producing many pages mainly to rank, "whether through automation, human efforts, or some combination of human and automated processes". Site reputation abuse covers third-party content placed on a host site to exploit its ranking signals, a pattern long associated with casino reviews on news and university domains. Expired domain abuse covers buying old domains to borrow their history. Each maps to a question below.

The ten questions

The questions that catch the most problems

Question 3, the swap test, is the simplest and the most revealing. Take a paragraph from a "best online casinos in Michigan" page and replace Michigan with Pennsylvania. If nothing else needs to change, the paragraph is not about Michigan, and a site built from such paragraphs is the kind of scaled content the spam policies describe. Real market pages fail the swap test because they contain facts that only apply there: the regulator, the launch date, the products available, the age limit and the help line.

Question 10 matters more since 2024. Google's site reputation abuse policy, tightened in November 2024, treats third-party content placed to exploit a host's ranking signals as a violation regardless of the host's involvement. On 28 August 2026 Google announced that, from 30 August, manual actions under the policy would no longer apply inside the European Economic Area, where affected sections may instead be separated to rank on their own, while manual actions continue elsewhere. Google also listed presentation, content quality, stated or implied authorship and duplication as factors in its review.

Question 9 reflects a recent change. Google's guidance on generative AI content, updated on 1 October 2026, states that generative models do not retrieve facts, that their output can contain errors, and that publishers should manually fact-check all AI-generated content, including titles, descriptions and structured data. Our own approach is described in our policy on AI use and corrections.

What the rubric deliberately leaves out

Word count is absent. A legality question can be answered well in 300 words and badly in 3,000. Keyword density is absent too. So are review stars: Google has not shown review rich results for an organisation's reviews of itself since 2019, so an operator's self-rating adds nothing.

The rubric also does not judge whether gambling content should exist in a given market. That is a legal question, answered by regulators; our guide to UK advertising rules for search content shows how one market's rules reach organic pages.

Using the rubric

Question 1 is easier to answer once the page's search intent is clear; our casino and sportsbook intent taxonomy helps there. Score a sample of a site's most important pages, not every page. Look for patterns: a site whose market pages all fail question 3, or whose guides all fail question 2, has a template problem rather than a page problem, and the fix is editorial policy rather than rewriting one page at a time. For a site that lost visibility in a core update, which our update timeline lists, scoring templates this way is one structured place to start.