Abstract
Search engine optimization (SEO) is the work of getting a business’s pages into Google’s unpaid results, where many local customers start their search. A new website has almost no links from other sites, and only 1.74% to 6.11% of new pages reach Google’s first page within a year. We coded the evidence behind 41 popular SEO posts on X (formerly Twitter) and checked each theme against Google’s published rules and measured studies. Of 30 posts that recommend a tactic, 17 rest on anecdote and 9 offer no evidence, and 13 of their authors sell something related. Tips for appearing in answers written by artificial intelligence (AI) drew the most likes, 17,464 across 9 posts, yet Google says its search does not use llms.txt, a text file meant to guide AI tools, and a matched study found no gain in AI citations from adding structured data, code that labels a page’s facts. Posting on large sites to borrow their rankings, mass-produced AI pages and keywords added to a business name break Google’s written policies. As a case we examined In Motion Automation, a small automation practice in Tampa Bay, Florida, whose site has one indexed page (as of 29 October 2026) and almost no links. Sites with weak link profiles held 2 to 5 of the first 10 results for trade answering searches such as “roofing answering service” and “hvac answering service”, and none of the US first 10 for broad terms such as “answering service for small business”. The evidence points a new local service site to one page per narrow search, a Google Business Profile (Google’s free business listing) with recent reviews, and links earned from real sources.
Introduction
When a homeowner needs a plumber, or a law firm needs someone to answer its phone at night, a common first step is a search. In a 2025 survey of 1,000 US consumers by BrightLocal, 45% said Google is where they go by default for local searches1. BrightLocal sells local marketing software, so its survey is a vendor study, research published by a firm that sells products tied to the finding; we label such sources throughout. Google answers a search with a results page. Its main parts are ads, which advertisers pay for by the click; organic results, the unpaid links Google ranks itself; for local searches, a local pack, a map with about three nearby businesses drawn from Google Business Profiles, the free listings where a business gives its name, category, hours, service area and reviews; and, more and more, an AI Overview, a summary written by Google’s artificial intelligence (AI) above the links. Search engine optimization (SEO) is work on a site, its pages and its reputation that aims to move it up the organic results and into the local pack.
A new website gets little of this traffic. Google ranks a page by how well it matches the search and by signals of trust, among them links from other sites, which Google’s PageRank system, one of its original ranking systems, still counts2. A link from another site is called a backlink, and the distinct sites that link to a page are its referring domains. A new site has few of either and no history. In an index of about 14 billion pages kept by Ahrefs, which sells SEO software, 96.55% of pages got no visits from Google3. Of a random sample of new pages, 1.74% reached Google’s top 10, the first page of results, within a year; among new English pages with real content, 6.11% did4. Demand is also spread thin. Almost 93% of the keywords, the phrases people type, in Ahrefs’s US database draw fewer than 10 searches a month5. Specific, low-volume phrases such as “roofing answering service” are called long-tail keywords, and they make up most of what people search for.
Advice on beating these odds spreads on X, the social network formerly called Twitter, where posts with thousands of likes promise traffic within days. Some promote programmatic SEO, thousands of pages built from one template and filled from a data set. Some promote parasite SEO, publishing on large sites such as Reddit or the blogging site Medium to borrow their rankings. Many promote GEO and AEO, generative engine optimization and answer engine optimization, names for efforts to appear inside answers written by AI tools such as ChatGPT and Google’s AI Overviews. A like records that a reader approved of a post; it carries no information on whether the tactic works. Google publishes the rules its search follows, and SEO software firms publish measured studies, so most popular claims can be checked against a written source.
Both the rules and the results page changed between 2024 and 2026. Google keeps spam policies, a written list of practices that can get a page demoted or removed from its index, the stored copy of the web that it searches. In March 2024 it added three: scaled content abuse, many low-value pages made to manipulate rankings; site reputation abuse, third-party pages placed on a strong site to borrow its rankings; and expired domain abuse, buying an old domain to reuse its reputation for low-value content6. It ran five core updates, broad changes to its ranking systems, and five spam updates between March 2025 and October 20267, and it stopped showing FAQ rich results, expandable question lists that a page could earn with structured data, code that labels a page’s facts for machines, on 7 May 20268. AI summaries now take clicks. In the Pew Research Center’s study of the browsing of 900 US adults in March 2025, people who saw an AI summary clicked a result on 8% of visits, against 15% when no summary appeared9. Advice written for the results page of 2023 may no longer hold, and advice written for AI answers is new and mostly untested.
This report asks which SEO advice can bring a new local service website its first search traffic. We test the answer on one case: In Motion Automation, a small practice in Tampa Bay, Florida, that sets up call answering, intake and follow-up automation for law firms, trade contractors, dental practices and other local businesses, with a website at agentsinmotion.ai. We collected 41 popular SEO posts from X and coded the evidence each one offered (Section 2). We checked each theme against Google’s documentation and published studies (Sections 3 to 7). We then measured the practice’s site and the Google results for 15 searches tied to its services, in the US and in Tampa (Section 8).
Most popular tips rest on anecdote or nothing, and several of the most liked either have no measured effect or break Google’s written rules (Figure 1). Tips that the evidence supports, such as refreshing pages that already rank and setting up a Business Profile, drew 2,722 likes; tips that fail or risk a penalty drew 20,192. The search results for the practice’s services show where a new site can enter: narrow trade searches where small sites already hold places in the top 10, while national vendors hold the broad terms. The report makes these contributions:
- A coded sample of 41 popular SEO posts on X, with likes, evidence type and whether the author sells something related (Section 2).
- A verdict on each popular tactic from Google’s rules and measured studies (Section 3, Table 1), with the evidence on time to rank (Section 4), AI answers (Section 5), local search and reviews (Section 6) and links and mentions (Section 7).
- A case: the baseline of one new local service site and the 15 searches open or closed to it (Section 8), with the pages and fixes the evidence supports (Section 9).
The most-liked SEO tips on X rest on anecdote, and likes do not follow the strength of the evidence
We gathered posts from 22 keyword searches on X for SEO tactics, ranked by engagement, plus the threads and articles the posts linked. We kept 41 posts from January 2024 to October 2026 that make an SEO claim: 30 that recommend a tactic and 11 that push back on one or bring data against it. Each post was coded by its main theme and by the strongest evidence it or its linked article offered. A data study states a method and a sample. Official documentation is Google’s own. A case study with numbers names a site and gives before-and-after figures. An anecdote is a story, or numbers with no named site or source. Videos and images attached to posts were not reviewed, and a post whose only proof was an image was coded as anecdote.
The 30 recommending posts drew 46,746 likes and the 11 pushback posts drew 2,035 (Figure 2). Tips about AI answers drew the most: 9 posts and 17,464 likes. Three of them, with 7,889 likes together, tell readers to publish an llms.txt file, a proposed text file meant to tell AI tools what a site contains101112. One says the file brought Vercel, a web hosting company, 10% of its traffic from AI tools, without a source11. Others report that AI search visitors convert to buyers 6 times as often as Google visitors13, or at 10% to 40% against 1% to 2%14, and that named companies grew through AI answers: Granola by 4,600% in new users from AI search in seven months15 and Railway from $20 million to about $100 million in annual revenue in a year16. None of these posts links a source for its number.
The single most-liked post, with 10,078 likes, describes services that drive cars full of phones around a town to fake requests for directions to a business, citing a Google patent on ranking places by direction requests; it shows no result17. The second, with 4,013 likes, demonstrates an AI tool rebuilding a site’s internal links, the links between a site’s own pages, for $0.2118; it reports a cost, not a change in traffic. Programmatic pages drew 7,365 likes over 6 posts. Two give self-reported numbers on named sites: one author’s sites drew 15,600, 9,000 and 4,500 visitors with no time window given19, and an AI writing company reported weekly clicks from Google rising from 971 to 5,500 in 60 days after it published more than 13,000 template pages, about half of them indexed at the time20. Two others offer only arithmetic: 10,000 pages, times 30 visits each, times a 2% purchase rate, times $10, gives $60,000 a month2122. Parasite posting drew the lowest median likes of the recommended themes, 220 a post, with claims such as one site going from 0 to 87,000 visitors “in days” through pages on Perplexity, an AI search engine, and no site named2324252627.
Of the 30 recommending posts, 17 rest on anecdote and 9 offer no evidence; 3 are case studies with numbers, all self-reported, and 1 is a data study, a summary of 14 studies by Ahrefs posted by an Ahrefs executive28. Thirteen of the 30 authors sell a product, course, agency service or sign-up offer in or behind the post, against 2 of the 11 pushback authors (Figure 3). Likes do not rise with evidence: the median recommending post with no evidence drew 1,289 likes, more than the median anecdote (1,059). The pushback posts cite Google’s documentation29, a list of manual penalties30, a reported penalty against a large publisher31 and an expert survey32, and drew 99 to 284 likes each.
Google’s rules and measured studies support pages matched to a buyer’s search and business profiles, and rule out several of the most-liked tips
Table 1 gives a verdict on each tactic. Holds means Google’s rules or measured studies support it. Partial means it works in some forms and not others. Fails or risky means Google says it does not use the thing recommended, the tactic breaks a written Google policy, or no outcome was shown. Local search and links have their own sections (6 and 7); this section covers the rest.
| Tactic as posted | Verdict | What the rules and studies show |
|---|---|---|
| One page per service, written for the buyer’s exact search | Holds | Google asks for content made for people, not to manipulate rankings33; experts surveyed by Whitespark rank a dedicated page per service first among local organic factors (opinion survey, vendor)34; small sites with such pages rank in the case searches (Section 8). |
| Long-tail searches where small sites already rank | Holds | Almost 93% of US keywords draw fewer than 10 searches a month (vendor)5; 2 to 5 of the top 10 for trade terms in the case are small sites (snapshot). |
| Google Business Profile with category, service area and reviews | Holds | Google ranks local results on relevance, distance and prominence, and reviews and links feed prominence35; details in Section 6. |
| Refresh pages that already rank | Holds, once a site has data | Rests on anecdote3637; it acts on the site’s own Google data, and no source read argues against it. A new site has few pages that already appear in searches. |
| Links and mentions earned from real sources | Holds | PageRank is still a core system2; links correlate with rank, more for local searches (vendor, correlation)38; Section 7. |
| Programmatic pages | Partial | Works where each page does a real job, such as a template or tool on a strong brand (vendor estimates)39; many pages without added value are scaled content abuse40. |
| Free tools and calculators | Partial | One vendor’s free tool ranked first “pretty much overnight” on a very strong domain (anecdote)3; small sites in the case use calculators. |
| AI-built internal links | Partial | Experts rarely name internal links a top factor (11.1% of 131, opinion survey)41; the post shows a cost, not a traffic change18. |
| Other AI-answer tips (mentions, answer pages) | Partial | Brand mentions correlate with AI visibility (vendor, correlation)42; Google says inauthentic mentions do not help43; conversion multiples on X have no source. |
| llms.txt and other files written for AI | Fails | Google: “Google Search itself doesn’t use them”43. |
| Structured data to win AI citations | Fails | Pages that added it gained no AI citations against matched controls (vendor)44; Google: no special markup is needed43. |
| FAQ markup for rich results | Fails | FAQ rich results stopped appearing on 7 May 20268. |
| Parasite posting on large sites | Risky | Breaks the site reputation abuse policy45; Reddit treats link-heavy self-promotion as spam46. |
| Mass AI pages | Risky | Scaled content abuse40; 1,446 of 79,000 checked sites drew penalties in March 2024 (vendor)47. |
| Keywords added to the business name | Risky | Breaks Google’s rule that the name match the real-world name48. |
| One cloned page per city | Risky | Google’s spam policies name city pages that funnel visitors to one page as doorway abuse40. |
| Lists of directory backlinks | Risky | Google lists low-quality directory links as link spam, links it treats as manipulation40. |
| “Rank in days” | Fails | “No one can guarantee a #1 ranking on Google”49; new pages rank slowly (Section 4). |
| Car full of phones faking direction requests | Fails | The post shows no result17; no source read tests it. |
Files and markup written for AI tools have no measured effect on Google
Google’s guide to its generative AI features, published in May 2026, answers the llms.txt tip directly: “You don’t need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search (including its generative AI capabilities), as Google Search itself doesn’t use them”43. One post reports that Lighthouse, Google’s free tool for auditing web pages, now checks for the file12. A check in an auditing tool is a different thing from an input to ranking, and the guide speaks to the second. The guide covers Google only; it says nothing about ChatGPT or other chatbots, and no study read measures the file’s effect on them.
Structured data is also called schema, after the schema.org vocabulary it uses; it labels facts such as a business’s name, address and services so that machines can read them. Several posts recommend it for AI citations, the links an AI answer shows as its sources. Ahrefs (a vendor) tracked 1,885 pages that added structured data between August 2025 and March 2026 against 4,000 matched pages that did not. Citations in AI Overviews fell 4.6%, a small but significant change, and citations in Google’s AI Mode, its chat-style search, and in ChatGPT rose 2.4% and 2.2%, both indistinguishable from zero44. Google says “Structured data isn’t required for generative AI search, and there’s no special schema.org markup you need to add”43. Structured data still has a job in ordinary search: Google documents markup that describes a local business on the business’s own pages50. FAQ markup no longer earns anything visible. Google limited FAQ rich results to well-known government and health sites in August 202351 and removed the feature from search on 7 May 20268.
The same guide rules out other AI-era tips. It says content need not be broken into small chunks for AI, that making a page for every variation of a question in order to manipulate AI answers violates the scaled content abuse policy, and that “seeking inauthentic ‘mentions’ across the web isn’t as helpful as it might seem”43. What it recommends is ordinary: “Creating content that people find unique, compelling, and useful will likely influence your website’s presence in generative AI search in the long run more than any of the other suggestions in this guide”43.
Programmatic pages work where each page does a job, and many thin pages count as spam
Google defines scaled content abuse as “when many pages are generated for the primary purpose of manipulating search rankings and not helping users”, “no matter how it’s created”, and lists “Using generative AI tools or other similar tools to generate many pages without adding value for users” as an example40. The programmatic cases that grew, in Ahrefs’s estimates, are strong brands where each page is a working tool or template: Canva’s feature pages grew from 23 ranking pages in early 2022 to about 310, and from about 20,000 to roughly 13 million monthly visits; Notion has about 600 template pages with about 204,000 visits and QuillBot about 130 tool pages with about 464,00039. The same article quotes John Mueller of Google: “Programmatic SEO is often a fancy banner for spam”39. The 60-day case on X is self-reported, covers a short window and had half its pages indexed20. Practitioners on X report AI content case studies that “succeeded at first” and then fell months later52, and one history traces the same pattern from content spinning in 2008 to 2011 to programmatic pages and AI content53.
AI-assisted pages are allowed, and mass-produced AI pages draw penalties
Google allows AI in writing. Its 2023 guidance on AI content says “Appropriate use of AI or automation is not against our guidelines”54. It asks that AI-written content, including titles and descriptions, be checked by a person for accuracy before publishing55. Mass production is the risk. After its March 2024 update, Google reported 45% less low-quality, unoriginal content in its results56. Originality.ai, a firm that sells an AI-text detector, found that 1,446 of 79,000 sites it checked on three advertising networks received manual actions in March 2024, penalties applied by a Google reviewer that here removed the sites from the index; all had some AI-written posts, and half had 90% to 100%47. The AI share was measured with the vendor’s own detector. In a study of 331,000 pages, Ahrefs found that 5.3% of pages ranked first to third were entirely AI-written, while pages with very high AI content were indexed less often (40.35% against 49.28% for low AI content), with AI share again estimated by a detector57. On X, one author reports that 127 of 200 AI articles were dropped from the index (anecdote)58.
Posting on large sites to borrow their rankings breaks Google’s site reputation abuse policy
Site reputation abuse is publishing third-party pages on a strong site to exploit its ranking signals. Google’s policy holds that “no amount of first-party involvement alters the fundamental third-party nature of the content or the unfair, exploitative nature of attempting to take advantage of the host’s sites ranking signals”45. Forbes Advisor’s pages fell soon after a critic published an exposé of the site; one analyst counted 1.7 million queries that dropped or were lost, though Google did not confirm a penalty5931. A practitioner on X reports that the policy stayed manual and that some penalized sections recovered60.
Reddit is highly visible. Google has had access to Reddit’s data through Reddit’s programming interface since a 2024 partnership61; Reddit gained more search visibility in Google’s US results in 2024 than any other site, reaching third place in SISTRIX’s measure (vendor)62; and Wikipedia, YouTube and Reddit made up 15% of the sources in the AI summaries Pew examined9. In Semrush’s study of 248,000 Reddit posts cited by AI tools (vendor), 80% had fewer than 20 upvotes63. Reddit’s own rules say that if a user’s posts are mainly links to a business the user runs, the user should limit them or buy ads46. In the second half of 2025, moderators made 40.7% of all removals, 68.6% of them automated, and spam was the most common reason for removals by Reddit’s staff, at 54%64. Google lists forum comments with optimized links as link spam40. No source read measured traffic from Reddit posts to a small business’s site.
Refreshing pages and adding free tools help a site once it has pages that rank
Search Console is Google’s free report of the searches a site appears for, how often it is shown (impressions) and how often it is clicked. Two posts tell owners to export pages ranking about 8th to 20th, or pages with many impressions and few clicks, and rewrite their titles and content3637. Both offer anecdote only. The tip uses the site’s own measurements and conflicts with no source read, but it needs pages that already have impressions, which a new site lacks. Internal links rank low in expert opinion: in a 2026 survey of 131 SEO practitioners, 11.1% named internal links among the most important ranking factors, against 57.1% for relevance and 54.8% for backlinks41.
Free tools have one vendor’s anecdote behind them. Ahrefs replaced a landing page with a free tool in 2018, and it “ranked #1 pretty much overnight”3; Ahrefs says its free tools are among its highest-traffic pages and have held steady as AI answers take clicks elsewhere39. Ahrefs’s domain is far stronger than a new site’s. In the case searches, two small sites that rank put a lost-revenue calculator in front of the reader (Section 8).
New pages take months to reach the first page, and most never do
The evidence on timing contradicts the fast claims on X (Figure 4). Ahrefs (a vendor) tracked new pages for a year: 1.74% of 1 million random new pages reached the top 10 within a year, against 5.7% in a 2017 study, and 6.11% of 2 million new English pages with real content did. Of the pages that did reach the top 10, 40.82% got there within a month, so a page that ranks quickly is the exception within a small minority. Older pages hold the first page: 72.9% of top-10 pages were more than three years old, 13.7% were less than one year old, and the average page in first place was 5 years old, up from 2 years in 20174. Links go with traffic: of about 20 million pages in Ahrefs’s index with no referring domains, 2,997 drew more than 1,000 search visits a month3.
Google gives no promise of speed. Its starter guide says “Some changes might take effect in a few hours, others could take several months. In general, you likely want to wait a few weeks to assess whether your work had beneficial effects in Google Search results”65. It also warns that “No one can guarantee a #1 ranking on Google”49 and that outside SEO tools “don’t have access to our internal ranking data”66. The author of the best-documented programmatic case on X replied to a reader that such pages take “3 months+ to take off”67. No post in the sample showed a ranking in days with a named site.
AI summaries cut clicks, appear less on buying searches, and cite pages that already rank
AI summaries reduce clicks on the results below them (Figure 5). In Pew’s panel, an independent study of the browsing records of 900 US adults, users who saw an AI summary clicked a regular result on 8% of visits and a link inside the summary on 1%, against 15% clicking a result when no summary appeared; 26% ended their browsing after a page with a summary, against 16%9. Click-through rate is the share of people shown a result who click it. Ahrefs (a vendor) compared Search Console data for 300,000 informational keywords and found that an AI Overview went with a 34.5% lower click-through rate for the first result in March 202568 and a 58% lower rate in December 202569. Both are correlations across keywords, not experiments.
How often AI summaries appear depends on the search and the source. Pew found one on 18% of searches in March 2025: on 8% of one- or two-word searches and on 53% of searches of ten words or more9. Seer Interactive, a marketing agency, found AI Overviews on 36% of informational queries, 8% of commercial queries (comparing options) and 5% of transactional queries (ready to buy or book), across 5.47 million queries from January 2025 to February 202670. First Page Sage, an SEO agency, reports AI answers on 81.6% of US queries, a figure far above the others with the least method published71. The searches in this report’s case do not follow Seer’s average: in one check of Google results on 29 October 2026, an AI Overview appeared on 14 of 15 searches tied to the practice’s services in the US and 13 of 15 in Tampa (Figure 6).
Ranking is still the way into AI Overviews. Ahrefs found that 76.10% of pages cited in AI Overviews ranked in Google’s top 10 for the same search, 9.50% ranked 11th to 100th and 14.40% ranked lower, across 1.9 million citations72; First Page Sage reports that fewer than four in ten cited pages rank in the top 1071, and Ahrefs publishes the larger sample. Google says a page needs no extra technical work to appear in AI features beyond being indexed and eligible for a snippet, the text excerpt shown under a result,73, and its 2026 guide adds that a site must be included in its generative AI features through a Search Console setting43. Seer found that a site cited in an AI Overview drew 120% more clicks per impression than an uncited site, and still 38% fewer than with no AI Overview at all, without showing that citation caused the gain70.
Chatbots send little traffic so far. Across 3,000 sites, Ahrefs found that 63% received any visits from AI chatbots, and these made up 0.17% of the average site’s visitors, about half of them from ChatGPT74. Similarweb (a vendor) reports that Google’s AI Mode sends visitors to a website on 1.6% to 2.5% of queries, against 17% to 19% for classic Google search75. Local use is rising: 45% of US consumers in BrightLocal’s 2026 survey said they used ChatGPT or other AI tools for local recommendations, up from 6% a year earlier (self-reported)76. In BrightLocal’s study of 200,085 local searches, Google Maps named a tracked business in 66% of searches against 32% to 38% for AI tools, and business websites made up 93% of the distinct sites that AI tools cited77. Google says Business Profiles help a business appear in its AI responses43.
Local results rest on the business profile’s category, distance and reviews, and Google’s name rule bars the keywords experts favor in a business name
Google states how it ranks local results: “Local results are mainly based on relevance, distance, and popularity”; links to a business and the number of its reviews feed the last of these, and “There’s no way to request or pay for a better local ranking on Google”35. Practitioners add detail. In Whitespark’s 2026 survey of 47 local search experts on 187 factors, an opinion survey by a vendor, the top factors for the local pack were the profile’s primary category (score 227), the distance from searcher to business (225), keywords in the business name on the profile (223), a physical address in the city searched (213) and being open at the time of the search (189)34.
The third factor conflicts with Google’s rules. Google requires that a profile’s name “reflect your business’s real-world name, as used consistently on your storefront, website, stationery, and as known to customers”48, so adding service words to the name breaks the rule; one of the posts on X warns that profiles get suspended for it78. Google also asks for as few categories as possible, chosen to complete the statement “This business IS a”, and requires a service-area business, one that serves customers at their location, to hide its address48. A home-based practice therefore lists as a service-area business. Whitespark’s experts rate a shown address above a hidden one (factor 7, score 176)34, which leaves such a business more reliant on its website and its reviews. Google rules out virtual offices, mailing addresses where the business does not operate, and asks that a service area stay within about two hours’ drive of the base48.
Reviews shape the choice once a searcher sees the profile (Figure 7). In BrightLocal’s 2026 survey of 1,002 US adults (vendor, self-reported), 97% read reviews for local businesses, 47% would not use a business with fewer than 20 reviews, 74% cared only about reviews from the last three months, and 89% expected owners to reply, while templated replies put off 50%76. Google bans offering incentives for reviews and asking only happy customers for them79, and the US Federal Trade Commission’s 2024 rule bans fake reviews and incentives tied to a positive or negative review, with civil penalties80.
The local posts on X rest on anecdote. One reports a plumber’s profile reaching 276 calls a month within seven months, with no business named78; another reports a dumpster company growing from $100,000 to $750,000 a year and recommends an office near the wealthiest zip codes, a business name that includes the service, 20 or more reviews a month and a location page per profile81. Location pages carry their own risk. Google names “multiple domain names or pages targeted at specific regions or cities that funnel users to one page” as doorway abuse40, and a 2022 trade-press column reported sites with little authority losing their near-duplicate location pages from the index after that year’s updates (practitioner report)82. No study read measures how city pages for service-area businesses have fared since 2024.
Links still count, more for local searches, and brand mentions track AI visibility more closely than links
Google still runs PageRank as “one of our core ranking systems”2. How much links move rankings is measured only by correlation, how closely two measures rise and fall together, which cannot show cause (Figure 8). Ahrefs compared the top 20 results for 1 million keywords using a rank correlation, on a scale from 0 (no relation) to 1 (perfect agreement). Rank correlated with the number of referring domains at 0.255, with backlinks at 0.248 and with domain rating, Ahrefs’s own 0 to 100 score of a site’s link strength, at 0.131; for local searches, links and referring domains correlated at 0.3338. Experts rank links high: 54.8% of the 131 practitioners in the 2026 survey named backlinks among the top factors41.
For AI answers, mentions matter more than links in the same correlational data. A brand mention is a site naming a business, with or without a link. Across 75,000 brands, mentions on the web correlated with how often a brand appeared in AI Overviews at 0.664, against 0.218 for backlinks42; in a second Ahrefs study, mentions on YouTube correlated with AI mentions at about 0.73783. Whitespark’s experts put presence on expert-curated best-of lists first among factors for AI visibility (score 179) and coverage in newspapers, blogs, government sites and industry associations fourth (160)34. Google’s guide warns that inauthentic mentions do not help43, so the evidence favors mentions that come from real coverage.
Bought links break Google’s rules. Its link spam list includes “Buying or selling links for ranking purposes”, excessive link exchanges, “Low-quality directory or bookmark site links” and forum comments with optimized links40. The most-liked backlink post on X lists 75 directory and profile sites, each with an authority score84; whether a given listing is low quality is not something any source read tested. In BuzzStream’s 2026 survey of digital PR practitioners (vendor), about 80% never buy links, 95.9% pitch content built on data, and the most common known cost per link, among those who knew it, was $300 to $50085; in an earlier BuzzStream summary, only about 4.6% of the guest-post sites that link vendors offered were of high quality86.
The case practice’s site has one indexed page and almost no links, and its open searches are narrow trade terms
The site today
In Motion Automation’s site, agentsinmotion.ai, had one page in Google’s index on 29 October 2026: a home page of 194 words. Its privacy and terms pages were not indexed. The site has no page for any single service; the home page lists eight general services, none of which names call answering, intake, missed calls or insurance verification. Its title tag, the page title Google shows in results, is the brand name alone (20 characters), and the privacy and terms pages repeat the home page’s meta description, the summary shown under the title. The site has no sitemap, a file listing its pages for crawlers, and no robots.txt, a file telling crawlers what to fetch. It has no structured data, no canonical tag naming the preferred address of each page, and no Open Graph tags, which build the preview card when a link is shared. The bare domain redirects to the www address with a temporary (307) redirect rather than a permanent one (301 or 308). The home page shows a contact email but no phone number or service-area text beyond “Tampa Bay, Florida”, and its contact email differs by one character from the one on brand.agentsinmotion.ai, a second subdomain of the practice.
The site has almost no links. Moz, an SEO data firm, scored it at Domain Authority 5 of 100 and found 4 live linking sites, which appear to be aggregator or scraped-directory pages; a second link index found none, and DataForSEO’s authority rank was 0. The practice has no Google Business Profile. Its names collide with other firms: a search for “In Motion Automation” returns Motion Industries’ “Motion Automation Intelligence”, based in Oldsmar in the same metro area, in 5 of the top 9 organic results and all 3 local-pack places, and “Agents in Motion”, the legal name, returns the Motion productivity app. The subdomain brand.agentsinmotion.ai is not indexed but carries no block against indexing, and links internal writing guides and sample client documents.
Searches open to a new site
The practice’s services map to keyword groups for each offering. Search volume is Google Ads’s estimate of average monthly searches, here for October 2025 to September 2026, rounded into bands; for local areas Google Ads reports any rare term as 10, so a value of 10 means very low. We measured volume for the US, for Florida and for the Tampa-St. Petersburg (Sarasota) designated market area, the media market Google Ads uses for local figures. Cost per click is the average price advertisers pay Google for one ad click on the term, a rough sign of what a customer from that search is worth. Keyword difficulty is DataForSEO’s 0 to 100 estimate of how hard a term is to rank for, computed from the links held by the pages already in its top 10. For each of the 15 searches in Table 2, we recorded Google’s top 10 from a US location and from Tampa, and the authority rank of each site, from 0 to 100, counting sites at 20 or below as small.
| Search | US | FL | Tampa | CPC | KD | Small sites US / Tampa | Median US / Tampa |
|---|---|---|---|---|---|---|---|
| Narrow trade searches | |||||||
| roofing answering service | 40 | n/a | 10 | 30.52 | n/a | 4 / 5 | 29 / 23 |
| hvac answering service | 390 | n/a | 10 | 69.25 | 0 | 3 / 4 | 26 / 22 |
| plumbing answering service | 480 | n/a | 10 | 74.08 | 0 | 2 / 5 | 43 / 9 |
| ai receptionist for hvac | 50 | 10 | 10 | 38.91 | 0 | 4 / 4 | 16 / 6 |
| missed call text back | 390 | 140 | 40 | 25.37 | 8 | 4 / 1 | 37 / 42.5 |
| dental insurance verification services | 210 | 90 | 10 | 163.34 | 0 | 1 / 1 | 42 / 40.5 |
| dental answering service | 110 | n/a | 10 | 181.68 | 0 | 1 / 1 | 41 / 39 |
| legal intake services | 480 | 20 | 10 | 148.06 | 0 | 0 / 1 | 40 / 40 |
| quote follow up | 140 | 10 | 10 | n/a | 0 | 0 / 1 | 46 / 35 |
| answering service tampa | 20 | n/a | 10 | n/a | 9 L | 1 / 2 | 39 / 47 |
| Broad terms held by national vendors and large sites | |||||||
| law firm answering service | 260 | 20 | 10 | 285.63 | 17 | 0 / 1 | 44 / 50 |
| after hours answering service | 1,000 | 40 | 10 | 143.54 | 0 L | 1 / 1 | 33 / 33 |
| answering service for small business | n/a | n/a | n/a | n/a | 2 L | 0 / 1 | 37 / 41 |
| speed to lead | 880 | 70 | 20 | 18.53 | 33 | 1 / 1 | 47.5 / 44 |
| email parser | 390 | 30 | 10 | 9.82 | 17 | 0 / 0 | 60 / 60 |
Narrow trade searches are open (Figure 9). For “roofing answering service”, small sites held 4 of the US top 10 and 5 of the Tampa top 10; for “hvac answering service”, 3 and 4; for “plumbing answering service”, 2 and 5, with a median Tampa authority of 9; and for “ai receptionist for hvac”, 4 in both, with a median of 16 in the US and 6 in Tampa. The volumes are small, 40 to 480 US searches a month, but the HVAC and plumbing answering terms carry a keyword difficulty of 0. “Missed call text back” has the largest Florida volume of any service-specific term in the map, 140 a month, and “dental insurance verification services” shows 90 in Florida, 43% of its US volume. Legal intake has no small sites in the US top 10, but one niche site with an authority of 30 held four of nine places for “legal intake services”. Broad terms are closed: no small site ranked in the US top 10 for “answering service for small business” or “law firm answering service”, where a Florida Bar member-benefit page (authority 69) ranked first, and “email parser” returned tool vendors with a median authority of 60.
Every phrase that pairs AI with Tampa, St. Petersburg or Florida, such as “ai consultant tampa”, returned no Google Ads data. Local demand for AI help appears only in broad terms within the Tampa market area: “ai automation” at 140 searches a month, “ai tampa” at 140 and “ai consultant” at 110. Such searches tend to resolve through the local pack, which the practice cannot enter without a Business Profile. The local pack appeared in Tampa for “hvac answering service”, “plumbing answering service”, “roofing answering service”, “answering service for small business” and “answering service tampa”; for “roofing answering service” it showed roofers, which suggests Google reads part of that search in Tampa as a search for a roofer. Owners do not search for the audit service by name: no business phrasing about missed calls returned volume, and “missed calls” itself is a consumer search about phones.
Pages that rank with low authority share a pattern. Each names the trade and the service in its title and main heading, and the Tampa winners add the city. A roofing answering-service page from a site with an authority of 8, ranked second in the US, covers storm-surge overflow, routing of active leaks before routine inspections and handoff to insurance, and includes a lost-revenue calculator; a Tampa dental answering-service page from a site with an authority of 22, ranked first in Tampa, presents estimated Tampa figures labeled as drawn from Census and labor statistics; an HVAC answering-service page from a site with an authority of 13 ranks fifth in the US and second in Tampa with 508 words and the city in its title. AI answering vendors publish prices on these pages, from $14 to $149 a month.
Implications for a site in this position
This section draws implications from the evidence above; it is not a measured result. The evidence favors fixing what keeps the existing page from being understood and found, then adding one page per narrow search where small sites already rank.
Before new pages, the evidence supports these fixes: a Google Business Profile listed as a service-area business, with one primary category and the name the practice uses with customers and nothing added; a robots.txt file and a sitemap; structured data on the home page describing a professional service that carries both In Motion Automation and the legal name Agents in Motion LLC, which gives Google a way to tell the practice apart from the firms with similar names; title tags and descriptions that name the services and Tampa Bay; a permanent redirect from the bare domain; Open Graph tags; one contact email used everywhere; a block on indexing for the brand subdomain; and a verified Search Console with the setting for Google’s generative AI features turned on.
The searches in Section 8 support a page set of a home page and one page per offering (Figure 10): HVAC and plumbing answering; law firm intake, built on “legal intake services”; dental answering with insurance verification; a roofing storm answering line; missed-call text-back; a quote follow-up page written as a how-to with templates, since that search asks for email wording; a missed-call audit page with a calculator; and an email-to-table data entry page. Separate Tampa pages fit only the trades where Tampa results already reward city pages (HVAC and plumbing, dental and law), each with its own Tampa detail, since cloned city pages are a named doorway example. The audit, the email-to-table page and the Tampa pages have little or no measured search volume. They would earn their keep by converting visitors who arrive from the trade pages, the Business Profile and referrals.
Off the site, the evidence supports reviews from real clients, requested from every client without incentives, answered in the owner’s own words; links and mentions earned through local business groups, data the practice publishes and lists of providers that editors choose; and no bought links, link swaps, mass directory listings or parasite posts.
Discussion
The tips that spread on X share a shape: a single action, a large number and a short time. A file to publish, a trick with phones or ten thousand pages in two days is easy to retell, and a story with a big before-and-after figure needs no method to be believed. The incentives point the same way. Thirteen of the 30 recommending authors sell a product, course or service tied to the tactic, and the most-liked post in the sample showed no result at all. The pushback posts, which cite Google’s documentation and measured cases, drew a small fraction of the likes. On this sample, likes measure how well a tip travels, and the evidence behind it has to be read separately.
The advice that holds up is slower and less novel. A page that answers one buyer’s exact search, a complete Business Profile, recent reviews and links from real sources are what Google’s documents describe and what the measured studies, mostly correlational, point to. The AI-era tips that work reduce to ordinary SEO: Google says its AI features rest on its core ranking systems43, three in four AI Overview citations come from pages already in the top 1072, and Google says Business Profiles help a business appear in its AI responses. For a local service business, AI answers so far add little traffic on their own, while the local pack and organic results for narrow trade searches remain open to small sites.
The evidence is thin in three places. No current independent study was found on how small business owners find and choose automation providers or other business services, so the value of case-study and comparison pages for such a practice is unmeasured; the professional-services research found dates from 2014 and 2015. No source measured referral traffic from Reddit posts to small business sites, although Reddit is prominent in Google and AI answers. And no study isolates how city pages for service-area businesses have performed since Google’s 2024 updates; the evidence is policy text and a 2022 practitioner column.
The site’s own data can test what the literature cannot. Once the trade pages are indexed, Search Console would show which searches each page appears for, its impressions and its average position, so the share that reaches the top 10 within months could be set against Ahrefs’s base rates of 1.74% to 6.11% within a year. Queries where a page sits 8th to 20th would identify the pages worth refreshing. Business Profile reports of calls, website clicks and direction requests would show whether the local pack or the organic pages bring more contacts, and whether review count and recency move them. Search Console’s figures for Google’s AI features, if reported, would show whether pages cited in AI Overviews draw clicks.
The study has limits. Most measured studies come from vendors that sell SEO, marketing or analytics products, and several (on links, mentions and AI citations) report correlations only. The X sample is a set of high-engagement posts found through keyword searches, not a random sample, and likes were counted on one day. The case search results are a single desktop snapshot on 29 October 2026 from a US and a Tampa location; rankings change daily, and Tampa city results and the Tampa market area cover different areas. Google Ads volumes are rounded into bands and floored at 10 for local areas, so small local demand cannot be told apart from none. Keyword difficulty and authority are models from SEO data firms, not Google’s measures. Some AI phone terms show a rise in Google Ads volume from July to September 2026 that inflates their 12-month averages.
Methods and sources
Sources were 41 posts on X from January 2024 to October 2026, retrieved with their likes, replies, reposts and views on 29 October 2026, with the threads and articles they linked; Google’s Search Central documentation, Business Profile help pages and Search Status Dashboard; published studies and surveys, labeled by publisher; Google Ads keyword estimates of average monthly searches for October 2025 to September 2026 in the US, Florida and the Tampa-St. Petersburg (Sarasota) market area, with keyword difficulty for 242 terms; Google search results for 15 priority searches from a US location and from Tampa on 29 October 2026, with authority ranks for the 146 domains in those top 10s; an on-page crawl of the case site and eight competing pages; and the Moz and SE Ranking backlink indexes. All were read on or before 29 October 2026.
Sources
- BrightLocal, "Consumer search behavior," 2025. brightlocal.com
- Google Search Central, "A guide to Google Search ranking systems," 2025. developers.google.com
- Tim Soulo, "96.55% of content gets no traffic from Google," Ahrefs, 2023. ahrefs.com
- Patrick Stox, "How long does it take to rank in Google?" Ahrefs, 2025. ahrefs.com
- Ahrefs, "Long-tail keywords," 2026. ahrefs.com
- Google Search Central, "What web creators should know about our March 2024 core update and new spam policies," 2024. developers.google.com
- Google, "Google Search Status Dashboard: Ranking history," 2026. status.search.google.com
- Google Search Central, "Latest documentation updates: Deprecating the FAQ rich result feature," 2026. developers.google.com
- Pew Research Center, "Google users are less likely to click on links when an AI summary appears in the results," 2025. pewresearch.org
- @aribk24, "Rip SEO: we are getting more and more clicks from LLMs everyday," X, 2025. x.com
- @fba, "If you run a SaaS and you haven’t published an llms.txt file yet, it’s time," X, 2026. x.com
- @kensavage, "Google just made it official. They added llms.txt as a Lighthouse audit," X, 2026. x.com
- @lennysan, "My biggest takeaways from @ethan_l_s on how to win at AEO," X, 2025. x.com
- @gregisenberg, "Understanding GEO, AI search and how to get traffic (and money) from it," X, 2025. x.com
- @wespreadjam, "Granola, a $1.5B company, increased monthly new users from AI search by 4,600% in JUST 7 months," X, 2026. x.com
- @irabukht, "Railway hacked AI SEO, and it took them from $20M to ~$100M ARR in 1 year," X, 2026. x.com
- @mikestrives, "Why is this guy driving around with 100s of phones in his car?" X, 2026. x.com
- @borjafat, "Jev is WILD for SEO audit," X, 2026. x.com
- @marclou, "Yes, you can grow a startup with 0 followers and $0. This is how I get 300 visitors a day from Google, for free, using programmatic SEO," X, 2024. x.com
- @jakezward, "How I built 13,000+ pages in 3 hours and grew SEO traffic +466% in 60 days," X, 2026. x.com
- @startupideaspod, "10,000 SEO pages in 48 hours. Here’s how programmatic SEO actually works," X, 2026. x.com
- @gregisenberg, "200,000+ new vibe coding projects get created every day yet almost NONE of them get customers," X, 2026. x.com
- @JulianGoldieSEO, "This Perplexity SEO hack is WILD. One site went from 0 to 87,000 visitors in days," X, 2025. x.com
- @ConnorShowler, "Easiest SEO traffic and LLM citations going into 2027," X, 2026. x.com
- @indexsy, "Here’s what’s actually working in SEO in 2026 vs what’s getting people cooked," X, 2026. x.com
- @JespernissenSEO, "Google’s own properties you can use for parasite SEO right now (October 2026)," X, 2026. x.com
- @denohawari, "Reddit is the EASIEST way to rank #1 on ChatGPT," X, 2026. x.com
- @timsoulo, "In the last 6 months at @Ahrefs, we analyzed over 1 billion data points across 14 studies," X, 2026. x.com
- @tibo_maker, "Google just officially proved every ‘AEO expert’ wrong," X, 2026. x.com
- @nichepursuits, "At least 1,446 Google manual actions have been applied since March 5th," X, 2024. x.com
- @rustybrick, "Looks like Google slapped Forbes Advisor with a manual action over the site reputation abuse policy," X, 2024. x.com
- @aleyda, "The 2026 Google ranking factors expert survey report is here," X, 2026. x.com
- Google Search Central, "Creating helpful, reliable, people-first content," 2026. developers.google.com
- Darren Shaw, "2026 local search ranking factors," Whitespark, 2025. whitespark.ca
- Google Business Profile Help, "Tips to improve your local ranking on Google," n.d.. support.google.com
- @Charles_SEO, "HUGE SEO tip (USING AI) that costs ZERO $ and takes UNDER 30 minutes," X, 2026. x.com
- @bloggersarvesh, "SEO pro tip #39: DON’T publish new content, instead," X, 2024. x.com
- Patrick Stox, "Google says “links matter less”: We looked at 1,000,000 SERPs to see if it’s true," Ahrefs, 2025. ahrefs.com
- Louise Linehan, "Programmatic SEO, explained for beginners," Ahrefs, 2026. ahrefs.com
- Google Search Central, "Spam policies for Google web search," 2026. developers.google.com
- Cyrus Shepard, "Google ranking factors expert survey 2026," Zyppy Signal, 2026. signal.zyppy.com
- Ahrefs, "An analysis of AI Overview brand visibility factors (75K brands studied)," 2025. ahrefs.com
- Google Search Central, "Optimizing for generative AI features on Google Search," 2026. developers.google.com
- Louise Linehan, "We tracked 1,885 pages adding schema. AI citations barely moved," Ahrefs, 2026. ahrefs.com
- Google Search Central, "Updating our site reputation abuse policy," 2024. developers.google.com
- Reddit Help, "Spam," n.d.. support.reddithelp.com
- Originality.ai, "Can Google detect and does it penalize AI content (March 2024 manual action list)," 2024. originality.ai
- Google Business Profile Help, "Guidelines for representing your business on Google," n.d.. support.google.com
- Google Search Central, "Do you need an SEO?" 2026. developers.google.com
- Google Search Central, "Local business (LocalBusiness) structured data," 2026. developers.google.com
- Google Search Central, "Changes to HowTo and FAQ rich results," 2023. developers.google.com
- @lilyraynyc, "Working on a presentation about AI content case studies that succeeded at first, but backfired," X, 2026. x.com
- Pedro Dias, "You’re not scaling content. You’re scaling disappointment," The Inference, 2026. theinferenceio.substack.com
- Google Search Central, "Google Search’s guidance about AI-generated content," 2023. developers.google.com
- Google Search Central, "Google Search’s guidance on using generative AI content on your website," 2026. developers.google.com
- Google, "New ways we’re tackling spammy, low-quality content on Search," 2024. blog.google
- Ahrefs, "Google doesn’t punish AI content; it punishes bad content (331k pages studied)," n.d.. ahrefs.com
- @noelcetaSEO, "I published 200 AI articles in 90 days. Zero EEAT signals added," X, 2025. x.com
- Barry Schwartz, "Did Google hit Forbes Advisor with a search penalty over site reputation abuse?" Search Engine Roundtable, 2024. seroundtable.com
- @lilyraynyc, "Google in 2024: talks a lot about “Site Reputation Abuse” and creates a new manual action for it," X, 2026. x.com
- Google, "Expanded Reddit partnership," 2024. blog.google
- Lily Ray, "IndexWatch: SEO winners in Google US search 2024," SISTRIX, 2025. sistrix.com
- Margarita Loktionova, "We analyzed 248K Reddit posts: What drives visibility in AI search," Semrush, 2025. semrush.com
- Reddit, Inc., "Transparency report: July to December 2025," 2026. redditinc.com
- Google Search Central, "SEO starter guide," 2025. developers.google.com
- Google Search Central, "Google Search’s guidance on using third-party SEO tools, services, and advice," 2026. developers.google.com
- @marclou, "Here’s how my startups got 100,000 visitors on autopilot for free using programmatic SEO," X, 2025. x.com
- Ryan Law, "AI Overviews reduce clicks by 34.5%," Ahrefs, 2025. ahrefs.com
- Ryan Law, "Update: AI Overviews reduce clicks by 58%," Ahrefs, 2026. ahrefs.com
- Seer Interactive, "AIO impact on Google CTR: 2026 update," 2026. seerinteractive.com
- First Page Sage, "Google click-through rates (CTRs) by ranking position," 2026. firstpagesage.com
- Louise Linehan, "76% of AI Overview citations pull from the top 10," Ahrefs, 2025. ahrefs.com
- Google Search Central, "AI features and your website," 2025. developers.google.com
- Louise Linehan, "63% of websites receive AI traffic (new study of 3,000 sites)," Ahrefs, 2025. ahrefs.com
- Limor Barenholtz, "Zero-click marketing: The SEO strategy for a world where most searches skip your site," Similarweb, 2026. aisearch.similarweb.com
- BrightLocal, "Local consumer review survey 2026," 2026. brightlocal.com
- Mike Hawkes, "Study: AI local search data," BrightLocal, 2026. brightlocal.com
- @noahiglerSEO, "We set up a Google Business Profile for a plumbing company and within 7 months it was doing 276 calls a month," X, 2026. x.com
- Google Maps, "Maps user-generated content policy: Fake engagement," n.d.. support.google.com
- Federal Trade Commission, "Federal Trade Commission announces final rule banning fake reviews and testimonials," 2024. ftc.gov
- @irentdumpsters, "I’ve run local SEO for 700+ home service companies across 40 states," X, 2026. x.com
- Search Engine Land, "Location pages in Google’s crosshairs after October spam update," 2022. searchengineland.com
- Louise Linehan, "AI brand visibility correlations (75,000 brands)," Ahrefs, 2025. ahrefs.com
- @hridoyreh, "75 places to get backlinks for startup," X, 2026. x.com
- Vince Nero, "State of digital PR report (2026)," BuzzStream, 2026. buzzstream.com
- Vince Nero, "Is link building dead? (15 data studies behind the clickbait)," BuzzStream, 2025. buzzstream.com
In Motion 