AI-powered marketing, lead response & automation for growth-focused teams. Jessica is online · answers in ~7s📞 (661) 219-7549
We're a marketing agency, and we run our own research on how search competition actually works. It keeps turning up things that surprise us — including the inconvenient findings most companies would quietly bury. We publish all of it here, with the real numbers and how we got them, because we'd rather be the people who showed their work. We even pointed the engine at our own website and published everything it caught.
If you write about SEO, small-business marketing, or AI in marketing, this page is for you. Everything below is free to use.
Preregistered studies with published methodology, aggregate downloads and a canonical page for every finding.
You may quote or reproduce any data on this page with attribution to AgencyFix and a link to this page. No need to ask us first.
Questions, or want the raw numbers or a chart? Email Jared directly: jared@agencyfix.com
| Measure | Value |
|---|---|
| Sites | 28 |
| Affected % | 25% |
| Affected Sites | 7 |
“In a check of 28 small-business websites, 7 had no written summary for Google to show in search results, so the search engine made one up itself — the line of text that often decides whether a customer clicks.”
In precise terms: 25% of Small Business Homepages Have No Meta Description — 7 of 28 sites audited (25%) were missing a meta description, leaving Google to auto-generate the search snippet that decides whether people click.
Aggregated over the most-recent audit per DISTINCT host in afx_tool_requests (excluding agencyfix.com + internal/test runs). Each percentage/median is recomputed directly from the stored per-site reports. n = 28 distinct sites.
Dataset: afx_tool_requests — 28 distinct external sites audited, US, computed 2026-09-30
| Measure | Value |
|---|---|
| Sites | 28 |
| Affected % | 28.6% |
| Affected Sites | 8 |
“New research found 8 of 28 local-business websites lack the behind-the-scenes code that lets Google and AI assistants like ChatGPT reliably identify and cite the business, meaning those firms can be overlooked when people ask an AI for recommendations.”
In precise terms: 28.6% of Local Business Sites Are Invisible to AI Assistants — 8 of 28 local-business sites audited (28.6%) have no structured-data (schema) markup — so Google and AI assistants like ChatGPT can't reliably identify or cite the business.
Aggregated over the most-recent audit per DISTINCT host in afx_tool_requests (excluding agencyfix.com + internal/test runs). Each percentage/median is recomputed directly from the stored per-site reports. n = 28 distinct sites.
Dataset: afx_tool_requests — 28 distinct external sites audited, US, computed 2026-09-30
| Measure | Value |
|---|---|
| Models | 3 |
| Questions | 10 |
| Answers Checked | 410 |
| Distinct Firms Named | 797 |
| Small Provider Mention Rate % | 0% |
“Across 410 answers from three leading AI assistants to 10 common buyer questions, the recommendations named 797 agencies and firms, but small or independent providers did not surface in a single answer.”
In precise terms: AI Assistants Recommended 797 Agencies Across 410 Answers — Almost None Small or Local — Across 410 AI-assistant answers (10 common buyer questions run through 3 leading models), the recommendations concentrated on 797 named agencies — overwhelmingly large, established firms. A small/independent provider did not surface in a single answer. AI search is consolidating discovery around incumbents, and smaller businesses are effectively invisible in it.
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Keywords | 237 |
| Multiple | 6.5 |
| Low Cpc Median Referring Domains | 8,704 |
| High Cpc Median Referring Domains | 1,346 |
“Among 237 local-marketing search terms, the ones businesses pay the most to advertise on needed a median of 1,346 links from other websites to rank, about 6.5x fewer than the 8,704 demanded by the cheapest terms — suggesting small businesses may face a lower barrier on the most commercially valuable searches.”
In precise terms: The Most Valuable Local-Marketing Clicks Need 6.5x FEWER Backlinks Than the Cheapest — Sorting 237 local-marketing keywords by advertiser value, the top-quartile-CPC terms — the ones businesses pay the most to advertise on — need a median of just 1,346 referring domains to rank, about 6.5x FEWER than the 8,704 the cheapest clicks demand. Counterintuitively, some of the most commercially valuable organic real estate is the LESS authority-gated. (Referring-domain and CPC figures are point-in-time snapshots.)
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Median Cpc Usd | 35.46 |
| Winnable Keywords | 74 |
“Across 74 local-marketing search terms that a smaller website could realistically rank for, the typical advertiser pays $35.46 per click, suggesting these terms attract buyers even without an elite profile of links from other sites.”
In precise terms: The Winnable Local-Marketing Keywords Carry a Median $35.46 Click Value — The keywords a lower-authority site can realistically win aren't cheap scraps: across 74 winnable local-marketing terms, the median advertiser click (CPC) is $35.46 — real commercial intent that's reachable without an elite backlink profile. (CPC is an advertiser-cost snapshot and shifts over time.)
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Sites | 28 |
| Median Score | 83 |
“An audit of 28 local-business websites found the typical site scored 83 out of 100 on technical setup, meaning many are only a few simple repairs away from being easier for customers to find.”
In precise terms: The Median Local Business Website Scores 83/100 on Technical SEO — Across 28 local-business websites we audited, the median technical-SEO score was 83/100 — most are leaving easy, fixable wins on the table.
Aggregated over the most-recent audit per DISTINCT host in afx_tool_requests (excluding agencyfix.com + internal/test runs). Each percentage/median is recomputed directly from the stored per-site reports. n = 28 distinct sites.
Dataset: afx_tool_requests — 28 distinct external sites audited, US, computed 2026-09-30
| Measure | Value |
|---|---|
| Keywords | 372 |
| Max Top3 Referring Domains | 480,410 |
“One search term's top results were backed by 480410 links from other websites, showing that in some corners of local search no amount of good content alone can break through.”
In precise terms: The Toughest Local-Marketing SERP Fronts 480,410 Referring Domains — At the top of the competition curve, a single keyword's top-3 fronted 480,410 referring domains — some SERPs are simply unreachable at any content quality.
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Analyzed | 498 |
| Winnable | 91 |
| Winnable % | 18.3% |
“Of 498 local search terms studied, just 91 (18.3%) were judged realistically winnable by a website with weak standing, meaning most are effectively off-limits for small businesses trying to get found.”
In precise terms: Only 18.3% of Local-Marketing Keywords Are Winnable for a Low-Authority Site — Of 498 keywords analyzed, 91 (18.3%) were graded winnable for a low-authority site; the rest were graded out of reach, dead terms (almost no searches), or had insufficient data.
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Keywords | 242 |
| Min Monthly Volume | 50 |
| Floor Referring Domains | 32 |
“A study of 242 search terms people actually use found that the least competitive one still needed links from 32 other websites to reach the top results, a useful reality check for small businesses hoping to rank without building any links.”
In precise terms: Even the Easiest Real-Demand Keyword Needs 32 Referring Domains — Among keywords with real demand (50+ monthly searches), the LEAST competitive top-3 still had 32 referring domains — the practical floor for a keyword worth targeting.
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Keywords | 372 |
| Median Top3 Referring Domains | 1,512 |
“Across 372 local-business marketing search terms, the median business in the top three results has about 1,512 links from other websites, a useful benchmark for small-business owners sizing up how much link-building it takes to compete.”
In precise terms: The Median Local-Marketing SERP Wants ~1,512 Referring Domains — Across 372 local-business marketing keywords analyzed, the median top-3 competitor has about 1,512 referring domains — the typical authority bar to rank.
Computed directly from srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22 — reproducible by re-running the same aggregate (median/min/max/rate) over the stored srp_keyword_competition rows.
Dataset: srp_keyword_competition — 372 contestable local-marketing keywords with top-3 referring-domain data (502 platform-dominated SERPs excluded), US, computed 2026-10-22
| Measure | Value |
|---|---|
| Location | US |
| Scan Month | 2026-07 |
| Keywords Analyzed | 203 |
| Min Top3 Median Referring Domains | 474 |
“Across 203 local-business marketing keywords studied, even the least competitive searches had top-ranking pages backed by an average of 474 links from other websites — meaning there is no easy shortcut into this space for small businesses trying to get found on Google. (This is our earlier 203-keyword sample, so it measures a different slice than the demand-filtered figures elsewhere on this page — not a contradiction.)”
In precise terms: The 474-Referring-Domain Floor — Across 203 local-business marketing keywords, even the LEAST competitive top-3 averaged 474 referring domains — the practical floor to rank in this niche.
DataForSEO SERP (top-3 organic) + Backlinks API referring-domain medians across 203 keywords, US, July 2026.
Dataset: srp_keyword_competition (203-kw scan, 2026-07)
| Measure | Value |
|---|---|
| Basis | equal-volume keyword pairs |
| Max Competition Multiplier | 7,000 |
| Min Competition Multiplier | 15 |
“Two keywords searched the same number of times each month can differ by up to 7,000 times in how many competing sites link to the pages already ranking, meaning popularity alone tells a small-business owner almost nothing about whether a topic is actually winnable.”
In precise terms: Same Demand, Up to 7,000x Harder — Two keywords with near-identical monthly search volume can face a 15x–7,000x gap in top-3 referring domains — demand tells you nothing about winnability.
Paired equal-volume keywords compared on top-3 median referring domains, same 203-keyword dataset.
Dataset: srp_keyword_competition (203-kw scan, 2026-07)
| Measure | Value |
|---|---|
| Bottleneck | referring domains / domain authority |
| Starting Point | bottom of winnability scale |
| Tracked Over Time | Yes |
“Research shows that for a new site, the main barrier to ranking on Google isn't how hard you work but how many other websites link to yours — and that reputation, while slow to build, does grow measurably over time.”
In precise terms: Zero to Climb: A New Site's Winnability Trajectory — A brand-new site can start at the bottom of the winnability scale and the climb is measurable — authority, not effort, is the bottleneck, and it moves over time.
Tracking our own site's winnability + referring-domain profile over time from launch.
Dataset: srp_keyword_competition (203-kw scan, 2026-07)
This page updates automatically as new findings clear review. Newest first. 13 published findings, last updated Sep 30, 2026.
Email Jared directly — a real person, usually same-day. He'll send the underlying data, build you a chart, or get on a call. And to say it once more: you don't need permission to use anything on this page. Just attribute it to AgencyFix with a link here.