- The shortlist forms before first contact. 6sense puts 61% of the buying journey behind the buyer before a vendor hears from them, and 95% of eventual winners came from the four-vendor list drawn on day one.
- Your own site is rarely the source that gets cited. One study of 118,067 B2B SaaS citations found no vendor brands at all in the top ten most-cited domains.
- The two biggest studies of this disagree, and we say so. Review platforms are either a top-three citation source or under 9% of citations, depending on whose measurement you read.
- Google says no special markup is required. Its own documentation states there are no additional requirements to appear in AI Overviews or AI Mode. We quote it below, because it costs us work to say it.
None of the local playbook applies here
Most of what an agency sells as local SEO has no surface to run on for a software company. Google's own guidelines state that a Business Profile is available if a business “has a physical location that customers can visit, or travels to customers where they are.” An online-only SaaS vendor meets neither condition. No profile, no map pack, no proximity ranking, no review velocity on a listing that does not exist.
What replaces it is a category you do not control. G2 requires a minimum of ten products before it will open a new category, adds roughly five to ten categories a month, and states that final categorisation decisions sit with its own market research team. Vendors can petition with demos and documentation. They cannot decide. A plumber in Marietta competes inside a category Google defined and will not move. A SaaS company competes inside a taxonomy a third party revises around sixty to a hundred times a year.
The third difference is who is reading. B2B software buying runs about 10.1 months and involves a group Gartner sizes at five to sixteen people across as many as four functions. Two thirds of those buyers now say they would prefer to complete the purchase without talking to a representative at all. The page has to answer the question a sales call would have answered, or the sales call never happens.
Six sources, ranked by what the research supports
Every figure here is sourced at the bottom of the page. Where the sources conflict, both are given.
Third-party “best of” pages, by a wide margin
Analysis of 1,263 solution-aware prompts across 250 B2B categories found 70.8% of all citations pointed at pages with “best”, “top”, “leading” or “popular” in the title, and 51.6% carried the current year. The single most-cited asset in this category is a listicle somebody else wrote.
Review platforms, at a share nobody agrees on
This is the honest part. One study of 118,067 B2B SaaS citations put G2 second only to Reddit. Another, run a year later across 250 categories, put the entire review-aggregator group at 8.6% of citations. Both are real measurements. They cannot both describe your category, and neither of them was measured on it.
Peer discussion, with volatility to match
Reddit topped the citation table in one dataset and sat seventh at 1.4% in another. Semrush observed Reddit's share of ChatGPT citations fall from roughly 60% to roughly 10% in a single week in September 2025, with no matching drop on other platforms. Anyone selling you a Reddit strategy as a durable channel is selling you a snapshot.
Your documentation, changelog and pricing page
Transparent pricing has been the number one thing software buyers say they want changed about the buying process for four consecutive years. It is also the page most vendors gate. Documentation is usually the largest indexable surface a SaaS company owns and the one least often treated as a search asset.
Analyst content, which is fading
The share of software buyers consulting analyst reports fell to 13% in 2026, a 63% decline since 2022. Analyst relations is a shrinking input to this decision while reviews have held roughly flat.
Your own site, mostly as corroboration
Vendor domains make up around a quarter of the top hundred cited sources on ChatGPT and Perplexity, and around three quarters on Google's AI Overviews and Gemini. Where you rank depends heavily on which engine the buyer opened.
Three things this industry is currently being charged for
Special AI markup. Google's documentation on AI features states, in its own words: “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” It goes further and says you do not need to create machine-readable files or AI text files, and that there is no special structured data to add. The stated condition is that the page is indexed and eligible to show with a snippet.
An llms.txt file, sold as an AI visibility measure. Google's John Mueller stated flatly in June 2025 that no AI system currently uses it. A study of 137,210 domains published in June 2026 found 97% of published llms.txt files had never been requested by anything at all, and that of the requests which did occur, more came from SEO audit tools than from AI retrieval bots. This site publishes one. We put it there before that evidence existed, it has cost us nothing to leave it, and on the current evidence it is doing nothing for us either. The glossary entry says the same thing.
Review volume as an AI visibility lever. An analysis across 500 G2 categories, published by G2 itself, found review count explained under 2% of the variance in AI citations. More reviews are worth having for buyers who read them. Buying them to move an AI answer is not supported by the only study we can find on it.
What we actually do here
The same six layers as every other engagement, weighted for a market with no local surface.
Category and entity resolution
Which categories you are listed in across the review platforms, whether they match the language buyers use, and whether your entity resolves consistently across your site, your listings and your funding and press records.
The comparison surface
The “best X” pages and comparison queries that decide the shortlist, who currently owns them, and which of them are reachable through work rather than through spend.
Documentation as an indexable asset
Treating docs, changelog and integration pages as retrievable content rather than as a support cost. This is usually the largest surface a SaaS company owns and the least deliberately built.
Pricing legibility
Whether a buyer, or a model summarising your category, can state what you cost. Four consecutive years of survey data say this is the thing buyers most want changed.
Measurement across engines
The same Share of Answer protocol we run everywhere, with the engine split reported separately, because vendor sites are cited at roughly three times the rate on Google surfaces as on ChatGPT.
Review-platform hygiene
Accurate categories, current screenshots, answered reviews. Worth doing because buyers read them. Not sold as an AI ranking lever, for the reason given above.
Fair questions
Do you specialise in B2B SaaS?
No, and the page would be more persuasive if we claimed otherwise. Our documented results are in home services. What we bring here is the measurement protocol and the six-layer method, applied to a market where the local surface is absent and the category is owned by a third party. If you want an agency with ten SaaS logos, we are not it, and we would rather you found that out now.
Can you get us cited in ChatGPT?
Nobody can promise that, and the research explains why. Generated answers are probabilistic, the citation mix shifts week to week, and the most-cited assets in this category are pages other people control. What we can do is measure where you stand, name the sources that decide it, and work on the ones that are reachable.
Is llms.txt worth adding?
On current evidence, no. It costs almost nothing, so keep one if you want it. We will not invoice for it, and we will not describe it as an AI visibility measure. The study is in the section above.
Do you work with pre-revenue or seed-stage software companies?
Rarely, and it is usually the wrong call for you. This work compounds over quarters and the fees only make sense against a customer worth enough that one won account pays for months of it. Before product-market fit, the money is better spent elsewhere. Our pricing is public so you can make that judgement without a call.
How do you report on this?
The same way as every other engagement: a Work Ledger of what was done, a visibility score against a frozen question set, and the numbers that did not move alongside the ones that did.
Start with a measurement, not a pitch
A free Visibility Check shows where you actually stand across Google and AI answers in your category, before any conversation about scope or price.
Get my free Visibility Check