Do Buying Extra Domains and Building Microsites Increase LLM Mentions?

Why do people think microsites work?

The logic runs like this. Assistants build answers from many sources. If you own several of those sources, you appear more often, and your product is described your way. There is a grain of truth to it. Research has shown that language models can be swayed by content crafted to promote a product. Nestaas and colleagues demonstrated "preference manipulation attacks" on Bing's and Perplexity's AI search using website content alone.

But the same research describes why it fails as a strategy. The authors call it a prisoner's dilemma. Everyone is tempted to do it, and the collective result degrades answers for all, which is exactly what search and AI companies then work to prevent.

What happens to microsite networks in practice?

• They are easy to connect. Shared owners, hosting, templates, analytics tags, writing style and outbound links reveal common control. Search engines have long identified networks this way.
• They breach search policies. Google's spam policies cover expired domain abuse, meaning domains bought and repurposed mainly to manipulate rankings with little-value content. They also cover doorway pages, scaled content abuse and link schemes. Google's policies explicitly include attempts to manipulate generative AI responses in its search. A site removed from a search index stops feeding the AI answers built on it, and a demoted one is retrieved far less often.
• They add little evidence. An assistant weighing "is this the best option?" finds five sites saying yes, all with thin content and no independent references. They add no independent confirmation, and they are easy to expose.
• They cost more than they seem to. Each site needs content, updates, hosting, security and monitoring. Neglected microsites go stale and start contradicting your main site.

Where is the legal line?

The FTC's Rule on the Use of Consumer Reviews and Testimonials, in force since October 2024, includes a section on company-controlled review websites. It makes it unfair or deceptive for a business to materially misrepresent that a website it controls, owns or operates provides independent reviews or opinions about a category of products that includes its own. So a site called "Best Payroll Software Reviews", secretly run by a payroll vendor and ranking that vendor first, is exactly what the rule targets. The rule carries civil penalties.

Disclosed comparison content on your own site is fine. The problem is the pretence of independence. The same logic applies to fake "community" sites and "expert" blogs that are really your marketing.

When does a separate site make sense?

In each case the site has its own reason to exist, discloses who runs it, and earns mentions on its merits. That is very different from a network of thin sites repeating the same claim.

What works better than microsites?

Put the same effort into sources assistants already trust:

• Earned coverage: journalists, industry newsletters and podcasts that name you in context. Muck Rack's analysis of AI citations found earned media made up about 84% of cited links.
• Review platforms your category's answers cite: ask every customer for an honest review in their own words, on platforms that allow requests (never Yelp).
• Original research published on your main site, which others cite and link to.
• Expert contributions under named authors, on publications your buyers read.
• Partner and marketplace listings that describe you accurately.
• A clear, complete main site, with proof pages, pricing, comparisons and documentation.

These take longer to build than a microsite. But each is a genuinely independent signal, and none carries the risk of a penalty or a regulator's letter. How LLMs find and use content in the first place is covered in whether LLMs build their own search index or rely on external search engines. How to tell whether they actually use your content is covered in how you can tell whether LLMs use your content or only index it.

How can you tell whether a microsite plan crosses the line?

Ask four questions before registering a domain:

• Would the site deserve to exist if AI did not? A tool, a research hub or a separate brand would. A site made only to be one more voice saying you are the best would not.
• Does it say who runs it? If readers would feel misled on learning that your company owns it, the plan relies on deception.
• Does it offer anything your main site cannot? If the content could live on your main domain with no loss, it probably should.
• Can you maintain it? A site that will be abandoned in a year becomes a source of stale, contradictory facts.

If any answer is no, put the budget into the alternatives above instead.

What if we already run microsites?

1. List every domain you own and what each site contains.
2. Keep the sites that pass the test: a distinct brand, tool, research project or market, with clear ownership.
3. Add disclosure to any comparison or review content: who runs the site and that your products are included.
4. Merge the rest into your main site where the content has value, with 301 redirects.
5. Retire thin sites that add nothing, and let their domains lapse or redirect them to the relevant page.
6. Check your backlink profile for links between your own sites that look like a scheme.

What might an illustrative case look like?

Consider an illustrative home security company whose previous agency had built six "independent" review microsites ranking its alarm systems first. For a while, some AI answers repeated their rankings. Then the sites lost their search visibility, and the answers moved on. A competitor also pointed a journalist to the network, and the resulting article was itself cited in answers about the company. The company closed the microsites and redirected them. It published a disclosed comparison on its own site and invested in a genuine customer survey. It also asked every customer for an honest review on two established platforms that allow requests. Recovering its reputation took most of a year.

What do people ask about microsites and AI visibility?

Do exact-match domains help AI mentions?

There is no evidence they do. Assistants care about what a site says and who else vouches for it, not the domain name.

Is a separate blog domain a microsite?

It can be. Unless there is a strong reason, keeping content on your main domain concentrates its value.

Can we buy an expired domain with a good history?

You can, but repurposing it mainly to manipulate rankings is expired domain abuse under Google's policies. Using it for a genuinely related, useful purpose is different.

Does owning many domains protect our brand?

Defensive registrations against typosquatting are sensible. Just redirect them to your main site rather than filling them with content.