NAMED OFFERINGS · NEW Fractional CMO from $2,500/mo. Plus the 30-day AI Adoption Sprint at $25k. See both paths
The short answer

Share of LLM (SoL or SoLLM) measures how often a brand is cited, recommended, or mentioned by ChatGPT, Claude, Perplexity, and Gemini when B2B buyers ask vendor-selection questions. In 2026, with 60% of B2B research mediated by AI agents and LLM-referred traffic converting at 14.2% versus 2.8% for organic, Share of LLM has replaced Domain Authority as the primary B2B visibility metric.

The new B2B metric

Share of LLM is the new B2B visibility metric.

In 2026, your buyer asks ChatGPT before they ask Google. The brands that get cited compound. The brands that don't disappear from the consideration set. Here's how to measure it.

Run the manual SoL checkHave IG measure yours
14.2%
LLM-referred B2B traffic conversion rate (vs 2.8% organic)
60%
of B2B vendor research now AI-mediated
34.5%
click reduction from Google AI Overviews
24 mo
window before citation moats fully entrench

What Share of LLM actually measures

Share of LLM (SoL, sometimes SoLLM) is the percentage of times a brand is cited, recommended, or mentioned by AI language models, ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, when a B2B buyer asks a vendor-selection or category-research question.

It replaces Domain Authority because Domain Authority measured what mattered when Google was the only buyer interface. In 2026, when 60% of B2B research is mediated by AI agents, the relevant question is no longer "do I rank?", it is "do I get cited?"

Citation rates are not a vanity metric. LLM-referred traffic converts at 14.2% versus 2.8% for organic, a five-times difference that compounds across the funnel.

How to measure Share of LLM (manual method, 2-3 hours)

SoL measurement is not yet automated by mainstream analytics tools. The practical approach takes 2-3 hours per quarter and runs across five AI surfaces.

StepWhat you doWhat you record
1. Pick 25 buyer queriesList the vendor-selection questions your ICP actually asks ("best fractional CMO for AI startups", "agentic marketing agency 2026")Spreadsheet column A
2. Run each query 5xAcross ChatGPT, Claude, Perplexity, Gemini, Google AI OverviewsColumns B-F: brand mentions per query
3. Score visibilityCited = 3 pts, mentioned = 1 pt, missing = 0Column G: total score per query
4. Sum + indexCalculate your % share vs total possible (25 queries × 5 surfaces × 3 pts = 375)Column H: your SoL %
5. Repeat quarterlyTrack delta, citation share compounds, so the trend matters more than the snapshotNew tab per quarter

Why the citation moat closes fast

AI engines re-cite sources they have already learned to trust. ChatGPT in particular shows strong recursive authority, sources cited by other cited sources rank disproportionately. This means early citation wins compound, and late entrants face an asymmetric uphill climb.

The Princeton GEO study found that adding expert quotes boosts visibility by roughly 41%, statistical density boosts citation by 30%, and embedded citations boost recommendation by 30%. These optimizations stack. They are not interchangeable with backlink quantity.

The brands that take AEO seriously in the next 24 months will compound a citation advantage that is hard to dislodge. The brands that wait will spend the rest of the decade trying to climb back.

What raises your Share of LLM

  • Answer-first formatting: 40-60 word direct answer at the top of every key page. AEO weights this at 19% of the citation signal.
  • FAQ schema quality: 20% of citation signal. Pages with FAQPage schema are dramatically more likely to be cited in AI Overviews.
  • Statistical density: 16% of citation signal. Quote real numbers from real sources. The Princeton GEO finding: stats boost citation by 30%.
  • Expert quotes: 41% visibility lift in Princeton testing. Quote your own named operators, link to outside experts.
  • Entity clarity: consistent brand definition across all pages so the LLM's knowledge graph has a stable node to cite.
  • llms.txt at root: tells AI crawlers how to use your content. Most B2B sites still don't have one.
  • Freshness: updated date stamps within the last year. The large majority of AI citations come from recently-updated pages.

The five sub-metrics that make up Share of LLM

A single number for SoL hides too much. The teams who win in citation graphs track five sub-metrics together and react to the one that's drifting.

Sub-metricWhat it measuresHealthy range (B2B category)
Citation ratePercentage of queries in the test set that name your brand at all.15% or higher on priority queries
PositionWhen cited, where in the answer your brand appears. Above the fold, mid-answer, or in a closing list.Above-the-fold on at least 30% of citations
SentimentIs the citation positive, neutral, or qualified by a caveat the model added.70%+ positive or neutral
Surface mixHow citation distributes across ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews.Cited on at least 3 of 5 surfaces
Query depthIs the citation triggered by broad category queries, mid-funnel comparisons, or only deep narrow queries.Triggers across all 3 query depths

Common SoL measurement mistakes

SoL discipline is new enough that most teams measure it badly. The errors below are the ones we see most often when we audit a competitor's program before quoting an engagement.

  • One-surface bias. Measuring only ChatGPT because it's the loudest. ChatGPT serves the most volume but Claude and Perplexity dominate research-mode queries where comparison decisions get made.
  • Vanity queries. Running queries the buyer would never ask. Test against the queries your sales team actually hears, not the ones your marketing team would prefer.
  • Snapshot thinking. Running SoL once and treating it as a permanent number. Citation share moves week to week as models retrain and as competitors push new content. Quarterly minimum, monthly for active programs.
  • Ignoring negative citations. A model citing you with a caveat ("X is solid but lacks Y") is signal. It tells you exactly what the next piece of content needs to address.
  • No control set. Tracking your citation share without tracking a comparable competitor's. The relative number is what matters when you walk into a board meeting.

How SoL fits alongside traditional SEO

SoL doesn't replace organic SEO. It sits on top of it. The two metrics correlate but the playbooks diverge in important ways.

Organic SEO optimizes for click-through from a results page. SoL optimizes for inclusion in a synthesized answer. Both reward authoritative content with structured data, but SoL adds new requirements: answer-first formatting, statistical density, FAQPage schema, llms.txt, and explicit attribution-ready phrasing.

The brands winning in SoL run both playbooks in parallel. We measure organic rankings quarterly and citation share quarterly, and we let the leading indicator be SoL because it moves first when the underlying authority compounds. If SoL is rising and organic is flat, the moat is forming. If SoL is flat and organic is falling, the moat is eroding faster than the dashboard suggests.

A six-week SoL launch plan

If a B2B company is starting from zero on Share of LLM, here's the engagement we run with portfolio companies. Six weeks, three deliverables, one named operator.

WeekWhat shipsOwner
Week 1Query set defined (50 queries across category, comparison, and decision tiers). Baseline measured across 5 surfaces.IG SoL lead
Week 2-3Citation-magnet content built. 6 to 8 pieces with answer-first format, statistical density, FAQ schema.Ashlesha Khond
Week 4Technical AEO sweep. Schema, llms.txt, sitemap, canonical, robots, structured data validation.David Garcia
Week 5Distribution and citation seeding. Partner content, mention placements, third-party validation.Chris Salazar
Week 6Re-measure baseline. Citation lift reported. Quarterly cadence handoff. The program is now in steady state.IG team
FAQ

Frequently asked questions

What is Share of LLM?
Share of LLM (SoL) is the percentage of times a brand is cited or recommended by AI language models like ChatGPT, Claude, Perplexity, and Gemini when buyers ask vendor-selection questions. It is the B2B visibility metric that has effectively replaced Domain Authority in 2026, because over 60% of B2B vendor research now happens through AI tools rather than traditional search.
How is Share of LLM different from SEO?
SEO measures whether you rank on a search engine results page. Share of LLM measures whether you get cited when an AI generates an answer. SEO's objective is "rank #1 for the keyword". SoL's objective is "be the source the AI uses when a buyer asks a question". Both matter, but with AI Overviews cutting organic clicks by 34.5%, SoL is now the more leveraged metric.
How do you measure Share of LLM?
Manually. Pick 25 buyer queries that match your ICP. Run each one across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Score 3 points for citation, 1 point for mention, 0 for missing. Sum the score, divide by the possible total (375), and you have your SoL %. The process takes 2-3 hours per quarter. Innovative Group runs this for clients as part of the SEO + AEO workstream.
What boosts Share of LLM the most?
Per Princeton GEO research: expert quotes boost visibility by approximately 41%, statistical density boosts citation by approximately 30%, and embedded citations boost recommendation by approximately 30%. AEO weighting research from 2026 shows FAQ schema quality contributes 20% of citation signal, answer-first formatting contributes 19%, and statistical density contributes 16%. These optimizations stack.
How fast does Share of LLM compound?
AI engines re-cite sources they already trust. ChatGPT specifically shows recursive authority: sources cited by other cited sources rank disproportionately. This means the first 24 months of citation are when the moat is built. Brands that establish citation share now will be cited at higher rates indefinitely. Late entrants face an asymmetric climb.
How is Share of LLM different from Share of Voice?
Share of Voice traditionally measured ad spend or media mentions across a category. Share of LLM measures citation share inside AI-synthesized answers. The two metrics overlap when traditional PR coverage compounds into LLM training data, but SoL is downstream of authority signals that PR alone doesn't generate (schema markup, structured data, statistical density). A high SoV does not guarantee a high SoL.
Can we improve our SoL without changing our website?
Partially. Third-party content (analyst mentions, podcast appearances, citations in industry reports) lifts SoL because LLMs train on the open web. But the durable gains come from on-domain changes: answer-first formatting, FAQPage schema, llms.txt, statistical density, and citation-ready phrasing. Off-domain alone is a moving target; on-domain compounds.
How often should we re-measure?
Quarterly minimum. Monthly for active programs in competitive categories. AI models retrain frequently and citation share is responsive within weeks of meaningful content shifts. The fastest signal: when a competitor publishes a deep AEO-optimized piece, your SoL on adjacent queries can move within 14 days.
Do you offer Share of LLM as a standalone service?
Yes. The 6-week SoL launch plan is a fixed engagement that ships with a named operator and a baseline-to-quarter-2 deliverable. Companies that run the full IG operating model get SoL measurement as part of the analytics workstream at no separate cost.
Related at Innovative Group
Pillar
Agentic marketing operating model
Services
AEO services for B2B
Blog
Share of LLM metric explainer 2026
Vertical
Fractional CMO for AI startups
POV
Customer Zero marketing playbook
Audit
The 2026 website audit checklist

Measure your SoL before the moat closes.

IG runs the SoL audit + the optimization stack as part of the SEO and content engine workstream. You get a baseline by next Friday, a 90-day improvement plan, and weekly tracking against ChatGPT, Claude, Perplexity, Gemini, and AI Overviews.

Book a SoL audit →

Thank you for reaching out!

Our team has received your message and will be in touch shortly.