Search Result Optimization. Stand out to surprise customers.Search

How to optimize search engine results

Contents 13

Search Result Optimization (SRO) is the practice of making a brand a source that search engines and AI answer engines cite. Where the results repeat the same consensus, SRO adds information those results do not contain, backed by evidence a machine can verify.

Kyle Roof ranked a Lorem Ipsum site first for "rhinoplasty Plano" in 2018

In a 2018 public SEO contest, Kyle Roof built a new site for the term "rhinoplasty Plano" whose on-page text was about 98 percent Lorem Ipsum. His team calculated how often the target term, its variations and contextual terms had to appear in specific places and placed them into the placeholder text. The site reached position one in organic search and in Maps.

Roof distinguishes ranking from conversion: in his account, ranking is a mathematical game driven by algorithms, while good content is what converts visitors.

SERP collapse: when results repeat only the consensus

When the documents on a results page repeat only the consensus, the consensus collapses into one answer restated many times: the same facts, with nothing one result adds to the others. On a results page, I call this consensus collapse a SERP collapse. The term does not refer to the consensus itself but to the absence of anything beyond it. A test: if a searcher who has read the first result learns nothing new from the next five, the page is in SERP collapse.

Three findings relate to this condition:

Shumailov et al. (Nature, 2024) showed that models trained on model-generated data lose the tails of the original data distribution. The term SERP collapse borrows this picture for results pages; it is not a measurement from that study.

Pew: Google users clicked a result in 8 percent of visits with an AI summary, 15 percent without

Pew Research Center analysed 68,879 Google searches by 900 U.S. adults in March 2025. With an AI summary on the page, users clicked a traditional result in 8 percent of visits; without one, in 15 percent. Links inside the summary were clicked in 1 percent of visits. 60 percent of searches beginning with who, what, when or why produced a summary. Pew reports an association, not a cause. With or without a summary, the largest share of searches ended without a click on a result link.

In the GEO study, cited sources, statistics and quotations raised visibility in generative answers by up to 40 percent

Aggarwal et al. (KDD 2024) tested content changes on GEO-bench, a benchmark of 10,000 queries. Adding cited sources, statistics and quotations from credible sources raised visibility in generative engine answers by up to 40 percent. Keyword stuffing performed below the unmodified baseline. The size of the effects differed by domain.

Pew found Wikipedia, YouTube and Reddit to be the most frequently cited sources in Google's AI summaries. Evidence is one factor among others; it is the factor a brand can add on its own page. By the definition above, a page leaves a SERP collapse only by adding information the other results lack.

Vercel and MERJ measured no JavaScript rendering by GPTBot, ClaudeBot or PerplexityBot

Vercel and MERJ analysed crawler traffic on nextjs.org, the Vercel network and two job board sites (published December 2024). None of the major AI crawlers they measured rendered JavaScript: OpenAI's OAI-SearchBot, ChatGPT-User and GPTBot, Anthropic's ClaudeBot, Meta-ExternalAgent, ByteDance's Bytespider and PerplexityBot. GPTBot and ClaudeBot fetched JavaScript files (11.50 and 23.84 percent of requests) without executing them. Googlebot, whose infrastructure Gemini uses, and AppleBot render JavaScript.

Schema markup written into the delivered HTML is part of what a non-rendering crawler receives. Schema markup that a plugin, tag manager or script inserts after the page loads exists only in the rendered page, so a crawler that does not execute JavaScript does not receive it. The Vercel study measured rendering, not schema specifically; the consequence for schema follows from the rendering data. Google documents JavaScript and Google Tag Manager as ways to generate structured data and recommends testing with the Rich Results Test, whose code explorer shows the rendered source. A page can pass that test while a non-rendering crawler receives no markup. The delivered source code shows what a non-rendering crawler receives.

Your page is fetched by retrieval, the generated answer is written from it, and it may name you or invent a false claim
What a crawler receives decides whether the answer carries your name or a false claim.

An ignorance graph publishes the documented gaps of a consensus

An ignorance graph is a documented map of what a field has not yet asked or answered. Every gap in it is tied to a verifiable source: peer-reviewed research, primary data, or documented practice.

Publishing an ignorance graph has two parts. The first states the consensus critically and without opinion. The second publishes the documented gaps as a new source of evidence.

Evidence for such gaps also sits outside the open web: in clinical practice, interviews, archives and internal data. Search Result Optimization begins with research before technical or semantic work: capture what people know and do, embed it in web content, and make it retrievable and citable for searchers and machines.

Six checks you can run on your own search results today

Run the checks in this order. Each step ends with a checkpoint; do not continue until it is met.

  1. Test your main query for SERP collapse. Search your most important query and open the first six results. Read the first one completely. For results two to six, write down every fact that the first result does not contain. Checkpoint: if your list is empty, the results page is in SERP collapse, and any page that adds a documented fact has room to stand out.
  2. Ask the AI answer engines. Put the same need as a full question, beginning with who, what, when, why or how, to ChatGPT, Perplexity, Gemini and Google. Pew found that question searches produce an AI summary 60 percent of the time. Checkpoint: for each engine, note whether your brand is named and which sources are cited.
  3. Check what non-rendering crawlers receive. Open the source code your server delivers for your page: type view-source: in front of the URL in the browser address bar. Search for application/ld+json and for your three most important statements. Checkpoint: schema markup and statements are in the delivered source. If they appear only in the browser inspector, they are inserted by script, and the crawlers measured by Vercel and MERJ as non-rendering do not receive them. Google's Rich Results Test will not show this, because it evaluates the rendered page.
  4. Write down what only you can document. List facts from your own data, practice, customer records or studies that none of the six results contain. Checkpoint: every fact has a source you can name, such as a dataset, a publication with DOI, or a documented case.
  5. Turn each fact into a citable passage. Write a heading that states the fact, followed by two or three sentences with the number and the source. In the GEO study, adding statistics and cited sources raised visibility in generative answers by up to 40 percent. Checkpoint: the passage is understandable without the rest of the page. You can test how a passage is retrieved with the What a retriever reads off each of your passages – Free online tool.
  6. Repeat check 2 after publication. Ask the same questions again once the page has been recrawled. Checkpoint: compare the brand mentions and cited sources with your notes from check 2.

If check 4 leaves your list empty, the documented gaps of your field have not been mapped yet. That is the work of an Information Gain Audit for a domain and of the thematic gap stack for a market. Questions: info@johannesfaupel.com.

What standard SEO, GEO, AEO and digital PR metrics do not measure

Each paragraph lists the working vocabulary of a discipline and its standard metrics, then names the quantity these metrics are not defined to capture. The comparison concerns the metrics, not the work of every practitioner.

SEO works with search intent, SERP features, structured data, meta descriptions, on-page optimization, internal linking, crawlability, Core Web Vitals, keyword research, information architecture and PageRank. Its standard metrics, such as clicks, impressions, organic traffic and domain rating, describe visibility, demand and link strength. None of these metrics is defined to capture whether a page adds information the other results lack. Google's information gain patent describes a score for this quantity.

GEO works with AI Overviews, citation rate, source authority, entity linking, knowledge graphs, retrieval-augmented generation, brand mentions, semantic search, vector embeddings and the context window. Its metrics, such as citation rate, source attribution and AI brand share, count whether and how often a brand appears in AI answers. A count is not defined to capture what the cited passage says about the brand. Pew found that links inside Google's AI summaries were clicked in 1 percent of visits.

AEO works with featured snippets, direct answers, zero-click search, schema markup, conversational search, knowledge panels, voice search optimization, entity recognition, natural language processing and intent disambiguation. Its metrics, such as featured snippet share and position zero capture, measure a brand's share of answers that already exist. A question for which no answer exists yet is outside that share. Pew found that 60 percent of Google searches beginning with who, what, when or why produced an AI summary.

Digital PR works with unlinked brand mentions, co-citation, backlink profiles, domain authority, linkable assets, digital outreach, citation building, thought leadership content, content syndication and media coverage. Its metrics, such as coverage count, referral traffic and earned media value, measure reach. None of them is defined to capture whether the coverage carries original data. In the GEO study, adding statistics and adding cited sources raised visibility in generative answers by up to 40 percent.

Definitions: Search Result Optimization, SERP collapse, ignorance graph

How Search Result Optimization differs from SEO

SEO, as measured by its standard metrics, optimizes a page's position among the documents already in the index. Search Result Optimization adds information those documents lack and makes it citable.

SERP collapse

SERP collapse is consensus collapse on a results page: the documents repeat only the consensus, so that no result adds information the others lack.

Ignorance graph

An ignorance graph is a map of the questions and facts a field has not yet covered, with each gap tied to a verifiable source.

Sources

First published 2022-08-05. Last revised 2026-09-21.

Search Result Optimization for your domain, enterprise, product or service: info@johannesfaupel.com. Or order the thematic gap stack for your market at https://overlookedrevenue.com/.

Search engine optimization