2026 SEO & AI Ranking Signals Infographic

View the Full 2026 Ranking Signals Report

View the Full 2026 Ranking Signals Report

The infographic is the short version. The interactive report has 16 chapters, 13 figures and every source.

Every finding below carries a confidence tag so you can tell a reproducible result from an interesting one:

High confidence Moderate confidence Low confidence Search did not get replaced in 2026. It got a second layer stacked on top of it, and that layer now decides what a huge share of people actually read. The useful question stopped being “does SEO still work?” a while ago. The useful question is which of two outcomes you are measuring, because ranking in Google and getting cited inside an AI answer have stopped agreeing with each other.

In mid-2024, roughly 76% of AI Overview citations came from pages already sitting in Google’s organic top 10. By February 2026 that number landed somewhere between 17% and 38%, depending on whose methodology you trust [11]. Take the widest credible reading and somewhere between 62% and 83% of citations now point at pages that do not rank on page one for the query that triggered them. That single shift is why the infographic is built around two outcomes instead of one.

The signal mix: two outcomes, one foundation

The infographic opens with a wheel labeled The Signal Mix. It breaks estimated ranking influence into six categories and normalizes them to 100. Two outputs sit on either side of it: organic ranking and AI visibility. The point of the layout is simple. You are not feeding two different machines two different diets. You are feeding one foundation and reading it on two different scoreboards.

Bars are scaled 3x so the gaps between categories are easier to read.

Why authority still tops the wheel

Authority sits at 26%, the largest single block. That surprises people who have heard that links are dying. They are not dying. The category here means links plus citations plus mentions plus trust signals, and as a whole it still carries the most weight. What collapsed is the cheapest way to buy it. Raw link volume on its own now sits at 11% in the factor table (panel two), and its correlation with AI visibility is weak. Authority did not shrink. It redistributed toward being discussed, not just linked.

Content quality is right behind at 24%, and it remains the most stable relationship in every dataset this report has tracked across four editions. Brand and entity signals at 14% are the category to watch, because they are the one that moved the most. More on that in the next section.

Two outcomes, a shared foundation

The signals that feed both lists overlap heavily: content quality, topical depth, freshness, demonstrable expertise and clean structure show up on both. That is why the right answer to “do I still need SEO?” is yes, and why the right organizational answer is one program with two sets of numbers. Teams running SEO and AI visibility as a single integrated workflow reported increased traffic or leads 81% of the time. Teams running them as separate initiatives reported the same result only 36% of the time [1].

Read the evidence correctly

The bottom strip of panel one is the fine print most infographics skip. It matters more than any single number on the page.

What changed in 2026: brand rises, content stays essential

Panel two compares estimated factor weights from the 2025 edition against 2026. The 2025 table had ten rows. This one has twelve, because two things that used to hide inside other categories now behave differently enough to need their own line: brand and entity signals, and freshness measured against citation rather than ranking.

Brand and entity signals are the story of the year

From roughly 2% to 8% in a single year, and that understates it. Unlinked brand mentions correlate with AI visibility at 0.66 to 0.71, against 0.22 for raw backlink count. Brands in the top quartile by mention volume receive on the order of 10x more AI citations than the rest of the field [5]. If one line in this table should change a budget, it is this one.

The mechanism is coherent even if causation is unproven: AI systems resolve a question to entities before they resolve it to documents. A brand that is discussed widely, consistently and in credible places is easier to resolve, easier to retrieve and easier to cite. Mention volume is simply the cheapest available proxy for that recognition.

Backlinks became a threshold, not a slope

Analysis of 1,000 domains found the relationship between authority and AI mentions far stronger by rank correlation (Spearman 0.57) than by linear correlation (Pearson 0.23) [9]. That gap is the statistical fingerprint of a threshold. You need to clear an authority floor to be considered at all. Once you are over it, additional links buy comparatively little.

Two details from the same study are worth sitting with. Follow and nofollow links correlate almost identically with AI visibility (0.334 vs. 0.340), and image links slightly outperform text links [9]. Read together, AI systems appear to treat a link as evidence that a brand is discussed rather than as a vote that accumulates.

Authority did not shrink, it redistributed. The change is not that authority stopped mattering. It is that the cheapest way to buy it stopped working.

2026 Ranking Signals Report

Freshness hardened into a maintenance requirement

Pages updated within 30 days earn 3.2x the AI citations of older equivalents, and pages left untouched for three months or more are 3x more likely to lose citations they already held [18]. Roughly 65% of AI Overview citations come from content under a year old and 89% from content under three years old [1].

Exact-match keywords keep losing precision

Title tag weight slid from 14% to 11%. Query fan-out is the reason. When one prompt gets decomposed into a dozen sub-queries you never see, tuning a title for the exact string the user typed is tuning for one input out of many. Cover the intent and the adjacent questions around it instead.

Technical is still hygiene, with a new way to fail

Mobile-friendliness, HTTPS and fast loading remain baseline expectations where failing costs more than excelling gains. The new failure mode is client-side rendering. Several AI crawlers execute little or no JavaScript, so content that only exists after hydration may never be retrieved at all. That problem did not exist in the 2024 version of this list.

Before you rewrite anything, read your traffic report correctly

The most common diagnostic mistake of 2026 goes like this: organic sessions are down year over year, so the content must be the problem. For a large share of sites it is not. Across publishers, impressions rose roughly 49% after AI Overviews launched while click-throughs fell about 30%, with positions holding steady [1]. Behavioral research measured an 8% click rate on queries with an AI Overview against 15% without, only about 1% of users clicking a link inside the overview, and 26% of sessions ending right after reading one vs. 16% without [13].

The traffic that still arrives is worth more, and that part gets lost in the doom headlines.

Someone who arrives after an assistant already answered their comparison questions is a different visitor from someone who clicked the fourth blue link. A channel report that only counts visits will tell you to abandon the most valuable traffic you have.

Two outcomes, two signal sets

Panel three puts two bar charts side by side, and it is the panel most worth printing out for a client meeting. The left chart shows what correlates with organic ranking position. The right chart shows what correlates with AI search visibility. Same industry, same year, very different ordering.

How to read the organic side

How to read the AI side

YouTube shows the strongest single correlation at 0.737, and YouTube supplies about 23.3% of all AI Overview citations [16]. Part of that is a Google-owned-property effect and deserves a discount. Enough survives the discount that publishing video versions of your highest-value written assets is a defensible AI visibility tactic, not a brand vanity project.

Domain Rating (0.296) and raw backlink count (0.218) sit at the bottom. That does not mean links are useless. It means that past the authority floor, link volume stops telling you much about who gets cited. Meanwhile earned media supplies roughly 82% of AI citations, with journalism alone accounting for 20% to 30% [17], and the same story syndicated across multiple publications generated up to 325% more citations than publishing it only on the owned site [1].

Five caveats that apply to every coefficient above:

How information becomes an answer

Traditional SEO optimizes for one event: a position in a ranked list. An AI answer gets assembled in four separate stages, and a page can fail at any one of them while looking perfectly healthy at the others. Panel four draws those stages as a lab pipeline running from a prompt to a visible citation. Almost every credible finding in generative engine optimization research only holds for one of these stages, and most of the bad advice comes from ignoring which one.

Why the stage matters more than the tactic

The widely repeated “40% visibility gain” from the original GEO paper was measured on documents that had already been retrieved [3]. It describes stage three. End-to-end testing that includes retrieval and reranking found that body-only rewrites can cut top-20 presence by roughly 9% and top-10 presence after reranking by roughly 16% [2][4]. Optimizing for the generator while ignoring retrieval can make a page harder to find, not easier.

AI Mode is a different product, not a bigger overview

Content written for a four-word head term is structurally mismatched to a conversational, comparison-shaped prompt. The register is different, the expected answer length is different, and the sub-questions the system fans out to are different. That is the clearest argument there is for writing comparison and consideration content instead of one more keyword-targeted page.

Position-one CTR impact estimates range from -15.5% in one 700,000-keyword study to -34.5% in a 300,000-keyword study, widening to -37% when a featured snippet also appears. The range is wide because the methods differ. The direction is not in dispute.

Every surface cites from a different pool

URL-level overlap between Google organic, AI Overviews and Gemini runs a Jaccard similarity of just 0.11 to 0.18. ChatGPT and Perplexity share only about 11% of cited domains. The top 15 domains hold roughly 68% of 680 million analyzed citations [16], and only 36 brands show up in the top 100 on all four major platforms every month [8]. There is no global AI ranking to optimize toward. Pick the one or two surfaces your buyers actually use and measure those properly.

Retrieval starts with access

Bots now account for about 57.5% of HTML traffic, and AI crawlers make up roughly 26.7% of that [15]. Whether the right ones can read you is the first gate in the pipeline, and the economics are lopsided:

As of May 2026, 51.8% of AI crawler activity was training-focused, 35.7% mixed and 9.3% search-only [15]. The emerging publisher consensus is to disallow training crawlers while allowing the answering and retrieval bots that can actually cite you, decided bot by bot. A broad wildcard rule written years ago for scrapers can quietly remove you from answers without recovering the bandwidth you care about.

Context changes the mix

Aggregate weights are a starting point, not a plan. Panel five shows how sharply emphasis shifts by vertical and by query type, and the AI layer has widened those gaps rather than smoothing them out.

Coverage expanded, then diverged. Education jumped from 18% to 83%. Restaurants went from 10% to 78%. B2B technology from 36% to 82%. E-commerce went the other direction, from 29% down to roughly 4%, as Google steered commercial intent toward Shopping surfaces and ad units. If you sell things, the AI Overview is mostly not your problem right now. If you explain things, it is the whole problem.

Three patterns worth acting on

Concentration matters too. In news and media the top three brands hold 82.9% of AI visibility, and in consumer electronics it is 76.9%. Finance sits at 41.4% and industrial at 42.2% [8]. In a concentrated vertical the winnable game is long-tail and comparison prompts. In a distributed one there is real room to move, and effort goes further.

Device still matters, just less interestingly. Mobile carries roughly 60% to 65% of search traffic, and 77.2% of mobile searches now end without a click, against about 60% overall [14]. The useful mobile question in 2026 is not whether your pages work on a phone. It is whether the zero-click rate on your mobile queries has quietly turned a channel you still report on into one that no longer moves revenue.

Evidence that gets noticed: seven signals

Panel six ranks the signals behind AI citations by measured strength, not by how often the tactic gets recommended on LinkedIn. Each one carries the report’s confidence rating, which reflects how consistently the finding reproduces across independent datasets, not how big the effect is [1]. This ordering is the most useful thing in the whole infographic, because it flips the priority list most SEO teams are still working from.

Sources: intent alignment and evidence effects [2][3]; mentions and YouTube [5][16]; recency [18]; position within the document [14]; heading hierarchy [7].

Answer position beats answer length

What the controlled benchmarks say

Two findings deserve more attention than they get. First, only 3 of 54 method-and-domain combinations in one benchmark showed a significant positive effect [2]. Generic advice like “add tables and bullets everywhere” generalizes poorly. Second, controlled multi-actor experiments show individual GEO gains eroding toward zero as adoption spreads. Early movers capture a redistribution of existing visibility, not new visibility. The tactical edge has a half-life. Genuine expertise, first-party data and earned coverage do not.

Where optimization turns into manipulation

The research literature lands on four cumulative tests. Pass all four and it is optimization. Fail any one and it is manipulation, carrying the same class of risk as every other tactic that works right up until it gets detected.

Foundations and limits

Panel seven continues the signal list from 8 to 13, and it splits into two moods: useful structure you should ship because it is cheap, and diminishing returns you should stop overpaying for. Its tagline says it plainly: solid foundations, realistic expectations.

The technical checklist for AI retrieval

If your platform generates an llms.txt file, leave it. Do not build a workflow around it, do not bill a client for it, and do not report it as an AI visibility initiative.

2026 Ranking Signals Report

Turn evidence into action

Everything before this panel is diagnosis. Panel eight is the part you can act on next week. It plots 18 actions by estimated impact against implementation effort, then closes with a measurement plan you can repeat every quarter. The top-left of the map is where a quarter should start. The bottom half is where a lot of SEO retainers quietly burn their hours.

The same four quadrants, in plain English

Where the budget should go

A repeatable measurement plan

AI answers are stochastic, and a big share of the reporting built on top of them is measuring noise and calling it progress. These four numbers are why a screenshot is not a measurement:

The adoption gap is the opportunity

Nearly everyone intends to do this. Under a quarter can tell whether it worked. Most competitors are running GEO programs they cannot evaluate, which means they cannot separate a real gain from drift and will keep optimizing toward whatever their last screenshot showed. A mediocre strategy with honest measurement beats a sophisticated one without it, because only one of them can correct itself.

Where this is heading: six bets for 2027 and 2028

Forecasts in this field age badly, so the report frames these as bets with stated reasoning, and each one names what would prove it wrong [1].

Fix measurement first, because without it every other decision is a guess. Invest in earned mentions and first-party data second, because those are the inputs with no half-life. And treat any tactic that works only because a model can be tricked as a liability with a countdown on it.

Jourdan Rombough, Optimization Theory

Questions people actually ask

The questions that come up in nearly every strategy conversation this year, answered with the data instead of reassurance.

How these numbers were built

Sources

Every figure in this breakdown traces to the 2026 report and the studies it aggregates. Many of the largest AI citation datasets come from companies that sell AI visibility tooling, so treat them as directionally useful and discount the magnitudes.

1. Jourdan Rombough, Optimization Theory. SEO Ranking Signals 2026: What Actually Moves Rankings and Citations. September 15, 2026. The full 16-chapter report this infographic summarizes.
2. arXiv. Optimizing Visibility in Generative Engines: A Critical Survey of Generative Engine Optimization (2023-2026). July 2026. Effect sizes, failure modes, the manipulation boundary and reliability bounds.
3. Aggarwal et al., Princeton, Georgia Tech, IIT Delhi. GEO: Generative Engine Optimization. KDD 2024. Origin of the widely quoted 40% visibility figure.
4. arXiv. C-SEO Bench: Does Conversational SEO Work?. 2025. Competitive and end-to-end GEO benchmarking.
5. Ahrefs. An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied). Spearman correlations for mentions, YouTube, anchor text, Domain Rating and backlinks.
6. Ahrefs. llms.txt study. 137,210 domains. The 97% zero-request finding.
7. Semrush. Technical SEO and AI search study. 5 million cited URLs across ChatGPT Search and Google AI Mode.
8. Semrush. 2026 AI Visibility Index. 126 million US prompts, 22 industries, four platforms.
9. Semrush. Backlinks and AI search visibility. 1,000 domains. Threshold effect, follow vs. nofollow, image vs. text links.
10. Semrush. Google AI Mode: early adoption and SEO impact. Query length, session depth and the 92% to 94% zero-click finding.
11. BrightEdge. Research reports. AI Overview coverage tracking, February 2026. Coverage by vertical and the organic overlap collapse.
12. Seer Interactive. Insights. AI Overview CTR longitudinal study, 2.43 billion impressions. The CTR curve and platform conversion rates.
13. Pew Research Center. Google users are less likely to click on links when an AI summary appears in the results. July 2025. 68,879 searches.
14. SparkToro and Datos. SparkToro blog. Zero-click search research. US zero-click rates and citation position within documents.
15. Cloudflare Radar. Bots. AI crawler telemetry. Crawl-to-referral ratios, crawler purpose split, blocking rates.
16. Surfer SEO. Surfer blog. AI Overview citation analysis, 46 million citations. Source concentration by domain.
17. Muck Rack. Muck Rack blog. AI citation and earned media study, 25 million cited links. The 82% earned media share.
18. ConvertMate. ConvertMate blog. Citation freshness analysis, 80 million citations. The 3.2x recency multiplier and depth advantage.
19. TechCrunch, reporting Adobe Digital Insights. AI traffic to US retailers rose 393% in Q1, and it’s boosting their revenue too. April 16, 2026.
20. Zyppy. Google ranking factors expert survey. 131 practitioners, 13,665 data points, 103 factors.
21. Google. Search Status Dashboard. Confirmed ranking update log.
22. Search Engine Land. Google algorithm updates library. Rollout reporting and corroboration.