What is black hat SEO: the short answer
Black hat SEO means breaking Google's spam policies in order to rank higher. White hat SEO follows them. Gray hat is not Google's term; people use it for the methods in between, which are not clearly banned but exist mainly to move rankings.
Google does not use the colors. It publishes spam policies, 16 sections with examples, and says it detects violations "through automated systems and, as needed, human review that can result in a manual action". Sites that violate them "may rank lower in results or not appear in results at all". So black hat SEO is not a mood or a style. It is a list, and a list can be read.
Below: the nine methods I group from that list, why people use each, what Google says, what I could see of the on-page ones on 132 top sites, where my test goes blind, my own position on gray methods, and a five-minute check of your site. This is a description of the methods, not a guide to using them.
White hat, gray hat and black hat SEO: what the difference is
The difference is intent, not the tool. A hidden menu is normal design; hidden keywords are spam. A paid link is normal advertising when it carries rel="sponsored" or rel="nofollow", and spam when it is meant to pass ranking credit. Google says buying and selling links "is a normal part of the economy of the web" and not a violation as long as they are qualified with those attributes.
| Label | What it means | Where it comes from |
|---|---|---|
| White hat | Follows Google's spam policies and guidelines | Google's wording |
| Black hat | Breaks the spam policies to rank higher | Common usage, defined by Google's list |
| Gray hat | Methods that are not clearly banned but exist mainly for rankings | Common usage, not a Google term |
My opinion, not a measured fact: gray hat is a description of how much risk you accept, not a safe zone. Google acts on the policies, not on the colors.
The 9 black hat SEO techniques from Google's spam policies
The wording below follows Google's spam policies, which I read on 18 November 2026. The "why people use it" parts are my reading, not Google's.
1. Hidden text and links
Words placed on a page so that a crawler reads them and a visitor does not. Google's examples: white text on a white background, text hidden behind an image, CSS that positions text off the screen, font size or opacity set to 0, and a link hidden in a single small character. People used it because a crawler reads code and a person reads a screen, and for a long time the two disagreed. Google adds that many design elements hide content for good reasons, such as menus and tabs, so the test is whether the content is hidden solely to manipulate search engines.
2. Keyword stuffing
Filling a page with keywords or numbers: lists of phone numbers without added value, blocks of text that list cities and regions the page wants to rank for, or the same words repeated until the text sounds unnatural. It worked in the era when ranking leaned on how often a phrase appeared. Today the risk is the opposite: the text becomes useless to the reader, who leaves.
3. Cloaking
Presenting different content to users and search engines with the intent to manipulate rankings and mislead users. Google's examples: a page about travel destinations for the search engine and a page about discount drugs for people, or inserting text only when the visitor is a search engine. Intent is in the definition. A paywall is not cloaking if Google sees what a subscriber sees, and a hacked site often shows cloaked content to hide the hack from its owner.
4. Sneaky redirects
Sending a visitor to a different URL than the one they requested, maliciously, to show users and search engines different content or to show users something unexpected. Google's examples: showing search engines one kind of content while redirecting users to something significantly different, and showing desktop users a normal page while sending mobile users to an unrelated one. A normal redirect after a site move is not this.
5. Doorway pages
Sites or pages created to rank for specific, similar queries that lead users to an intermediate page less useful than the final destination. Google's examples: several websites with slight variations of the URL and home page, and several domains or pages aimed at regions or cities that funnel users to one page. People used them to cover every phrase and place with its own page.
6. Scaled content and scraping
Many pages generated mainly to manipulate rankings and not to help users. Google's examples: using generative AI tools to produce many pages without adding value, scraping feeds or search results to create pages, stitching content from different pages, and creating many pages that make little sense but contain search keywords. Google says this applies "no matter how it's created", so the tool is not the line, value to the reader is. Thin affiliate pages with product descriptions copied from the merchant are in the same family. I covered where the line sits in programmatic SEO done right.
7. Link schemes and link networks
Creating links to or from a site primarily to manipulate rankings. Google's list includes buying or selling links for ranking, excessive link exchanges, automated programs or services that create links to your site, advertorials and guest posts with optimized anchor text, widely distributed links in footers or templates, and forum comments with optimized links. Google's page does not use the term PBN, but a network of sites built to link to one money site and chains of links pointing at those links fall under "creating links primarily to manipulate rankings". People use links because they are still a way to estimate trust, so someone will always try to fake them.
8. Expired domain abuse
Buying an expired domain and repurposing it primarily to manipulate rankings with content that gives users little or no value. Google's examples: affiliate content on a site previously used by a government agency, commercial medical products on a former non-profit's site, and casino content on a former elementary school site. The old domain brings a history; the new content borrows it.
9. Site reputation abuse
Publishing third-party content on a host site mainly because of that host's established ranking signals, so that the content ranks better than it would on its own. Google's examples: an educational site hosting a page of sponsored payday loan reviews written by a third party and distributed across other sites, and a medical site hosting a low-quality third-party "best casinos" page that is not integrated with the site. Google lists what is not inconsistent with the policy: syndicated news, user-generated forums, editorial columns, and affiliate links used throughout a page with links treated appropriately. For pages in this situation outside the EEA, Google says a manual action may apply; inside the EEA it describes categorizing such pages as separate from the main domain instead.
If you want to see what the cross-niche version of this looks like in practice, my casino affiliate site blueprint shows the white hat build of the same niche.
What I could see of black hat SEO on 132 top sites
I checked the on-page methods, because only those can be read from outside. On 18 November 2026 a Python script requested the homepages of 100 big US sites and 32 casino and loan sites twice: once as a Chrome browser, once with a Googlebot user agent. It compared the words, looked for hidden blocks, counted the most repeated word and read the link tags.
The first finding is about access: 77 of 132 homepages (56 of the 100 big sites and 21 of the 32 casino and loan sites) could be read by a plain request; the rest answered scripts with a block page or an error. Everything below is about those 77.
- Cloaking by user agent. 62 sites gave a readable page to both. On 50 the words matched on 99% or more. 9 differed by more than 5%: three were empty JavaScript shells, nothing to compare, and on six the vocabulary differed. I looked at three of those six: on a beauty store and a recipe site the fake Googlebot received more category and product names, and on a casino review site different promo and legal blocks. That can be a dynamic page or something else, and a raw request cannot tell the intent, so I do not call any of them cloaking.
- A closed door for the fake bot. 15 of 132 sites (12 big, 3 casino) answered my fake Googlebot with an error while giving Chrome a normal page, among them Wikipedia, Reddit and CNN. My guess, not a measurement: they check who is really knocking by its address.
- Hidden text. I looked for blocks of 25 words or more hidden by the page's own inline styles. 5 of 77 sites had one. The first block on each was a slide-out menu, a browser notice, a menu, a news ticker and a brand paragraph. Not one was a keyword list.
- Keyword stuffing. Among 62 pages of 300 words or more, one word took 8% or more of the text on 7. Six were interface words and brand names ("loading", "save", "watchlist"), one a casino guide where "casinos" is the subject.
- The dead tag. 15 of 77 sites still have a
meta keywordstag. Google's page on supported tags says it is not used by Google Search and has "no effect on indexing and ranking at all". - Links. On the homepages of 69 sites I counted 1,799 outside links; 291 (16%) had
nofollow,sponsoredorugc. For big sites it was 225 of 1,535 (15%), for the casino and loan sites 66 of 264 (25%). Links with URLs like/go/,/out/or?aff=appeared on only two casino sites: 95 links, all withnofolloworsponsored. A first, wider pattern also caught ordinary/play/and/bonus/sections, so I dropped it.
What this check cannot see
It cannot say that top sites play clean. It says the classic on-page tricks are rare among the sites a script can read. Four limits:
- Intent is not in the code. A hidden block or a different page for a bot proves nothing by itself.
- A fake Googlebot is not Google. The real one comes from Google's addresses, and a site can treat it differently.
- Raw HTML only. I did not run JavaScript, did not read external CSS, and looked at homepages only.
- Links, old domains and borrowed reputation do not show on a page. That is most of methods 7, 8 and 9.
Black hat SEO today: where it went
My reading of this data, with the caveat that it is an opinion: the on-page tricks are cheap to catch and have been caught for years, so what remains lives off the page. Links come from other sites. An expired domain looks like any site until you know its past. Borrowed reputation is a page on a trusted site that has nothing to do with it. None of it shows in the source of one homepage, which is why a test like mine cannot find it, and why I did not try to name sites that "use black hat". I cannot prove it about anyone.
In the casino and loan niches, which everyone suspects, I found the same on-page picture: no hidden keyword text, no stuffing. The difference was in links, where a quarter of outside links carried a nofollow or sponsored tag against 15% on big sites. For the build of such a site that stays inside the rules, see my teardown of casino affiliate sites.
My position on gray hat methods
I do not practice black hat. I do use gray methods, such as a network of my own sites and tiered links. I treat that as a risk I accept knowingly: it works until Google decides it does not, and the decision is Google's. If a site is the only income of a business, the arithmetic is different. I describe the methods here and do not teach them; if you build links, build them for readers first.
How to check your own site in five minutes
- Manual actions. In Search Console open Security and manual actions, then Manual actions. If a human reviewer has taken action, you are notified there and in the message center. An empty report means no manual action is recorded; it does not cover automated demotion.
- A live URL test. In URL Inspection test the live page and look at what Google rendered. Compare it with your browser: the same words and the same links is what you want.
- Hidden blocks in the source. Search your page source for
display:none,visibility:hiddenand negative offsets, and read what sits inside. A menu or a banner is fine; a paragraph of keywords is not. - Paid and affiliate links. Check that they carry
rel="sponsored"orrel="nofollow", as Google asks. - Who built your links. If someone built links for you and cannot say where they came from, ask. That is the risk I would look at first.
If your pages disappear from results and you do not know why, start with why Google is not indexing your pages; a spam action is one cause among several.
Mistakes I see in this topic
- Treating every hidden element as spam. Menus, tabs and banners are hidden by design. Judge the content and the intent.
- Calling every difference between Chrome and a bot "cloaking". Dynamic pages differ for ordinary reasons.
- Believing in the meta keywords tag. It has no effect on ranking.
- Marking paid links as normal links. That is the one link practice Google asks you to fix with an attribute.
- Hiring links and not asking where they come from. The responsibility for the result stays with the site.
What I did not measure
- Whether any of the methods works. I did not test rankings.
- Anything off the homepage. Category pages, reviews and templates may differ.
- One day. 18 November 2026; sites change.
- Casino and loan sites are a sample of 32, of which 21 were readable. It is a hint, not a rule, and I do not name them.
If you want your site checked against these policies, including links and templates, I do it as part of an SEO audit.
Files for this article
Below the article: the spam-signals-check script that repeats my check on your pages, a cheat sheet with the policies, how to check each and what to do, and the dataset of the 100 big sites under CC BY 4.0 with the measuring script.
