—— § Industries

The benchmark for
your category.

Six categories, each with its own citation economics. What earns a mention in a legal query is close to useless in ecommerce, and the gap between the best and worst performer in a category is usually wider than the gap between categories. These pages carry the benchmark numbers, the queries buyers actually type, the brands currently winning them, and a playbook written for that vertical alone.

—— § How to use this

Start with your own category, but do not stop there. The benchmark table below is the fastest way to see what the ceiling looks like where you compete — and the numbers are deliberately uncomfortable. In most SaaS categories the most-cited tool is cited roughly four times as often as the fifth, and in fintech the top twenty domains absorb the overwhelming majority of all citations. Those distributions are not a ranking you climb gradually; they are a threshold you either clear or you do not, which is why the playbooks are ordered by leverage rather than by effort.

The second thing to do is read one category that is not yours. The tactics transfer more than you would expect, because the underlying filter is the same: a model deciding whether a source is trustworthy enough to quote. Legal learned that a named, credentialled byline roughly triples citation rate; health tech found the same effect with clinician review. If you publish anything a reader could act on, that finding applies to you regardless of vertical. Each page ends with the queries buyers in that category actually ask, which is the single most useful artefact to take into a content planning meeting.

—— § Benchmarks

One number per category

The headline finding from each vertical analysis, with the brands the Observer sees cited most often in that category. Full sourcing sits on each page.

Headline AI citation benchmark and most-cited brands by industry
Category	Headline finding	Most cited today
SaaS	4.2× citation-rate gap between the #1 and #5 tool in most SaaS categories.	Linear — Wins on extractable feature pages. Notion — Dominates 'best X for Y' category queries. Stripe — Docs are the LLM retrieval gold standard.
Shopify Stores	73% of ecommerce AI citations trace back to Reddit threads or product schema.	Allbirds — Textbook product schema + review aggregation. Gymshark — Owns every Reddit thread in its category. Bombas — Unique-claim writing on social impact.
Law Firms	2.9× higher citation rate on bylined articles vs anonymous 'Firm Blog' posts.	Cooley — Docs-grade content on their /go startup hub. Gunderson Dettmer — Named attorney bylines on every article. Fenwick — Dominates pre-IPO legal queries.
Fintech	91% of fintech AI citations come from the top 20 source domains.	Mercury — Wins the 'startup banking' category outright. Stripe — API docs are the gold standard for citation. Wise — Price-comparison tables pull heavy citations.
Dev Tools	40-60% higher Claude citation rate for dev tools that ship llms-full.txt.	Stripe — Docs win is eternal. Vercel — Owns 'deploy X' queries via aggressive llms.txt. Supabase — Open-source + great docs = citation flywheel.
Health Tech	3.6× citation rate lift from clinician-reviewed bylines in health content.	Headspace — Clinician-reviewed content authority. Oura — Published peer-reviewed research on sleep data. Hims & Hers — Wins DTC men's health queries.

—— § Index

Playbooks

6 categories

Each page opens with the queries buyers in that category type into an AI assistant, then works through a playbook of concrete moves ordered by leverage.

Every playbook assumes the fundamentals are already in place. If they are not, the how-to guides cover llms.txt, FAQ schema and citation tracking first, and your first scan tells you which of them you are missing.

—— § Frequently asked

Where do these industry benchmarks come from?

From the AIRank scan cohort and the Observer query panel, segmented by category. Each page names its own sample — the SaaS figure is drawn from roughly 1,200 sites, the Shopify figure from about 2,400 stores, fintech from around 600. Sample sizes differ, so treat the numbers as the shape of a distribution in that vertical rather than as a precise industry average.

My category is not listed. Do these still apply?

Mostly. The six here were chosen because they show the clearest divergence from each other, so between them they cover most of the patterns. Pick the one whose buying process resembles yours — long considered B2B purchases behave like SaaS, regulated advice behaves like legal or health tech, and anything transactional behaves like ecommerce. The playbook steps transfer; only the query examples need swapping.

Why does the same tactic work in one industry and not another?

Because AI assistants apply different trust thresholds by topic. Health and financial queries are filtered hardest, which is why credential signals dominate those playbooks, while ecommerce answers lean heavily on third-party discussion because there is no credential to check. Matching the tactic to the filter your category faces is most of the work.

How current are the brand examples?

The most-cited brands listed on each page reflect what the Observer saw in the most recent analysis for that vertical, and they move. Treat them as worked examples of what a winning pattern looks like — Stripe's documentation, a bylined legal article, a Reddit thread that will not die — rather than as a live leaderboard. The leaderboard itself is a separate, continuously updated page.

Benchmarks tell you the ceiling. The leaderboard tells you where you sit against it right now.

Open the leaderboard