Can Buyers Verify Your Market-Leading Claim?

What public evidence can prove, what it cannot, and where a founder's claim fails before the sales call

Research date: August 17, 2026
Scope: B2B technology buying, with supporting evidence from multi-industry B2B studies and professional research standards. Software findings are labeled and do not stand in for every enterprise market.
Evidence standard: studies with a disclosed sample, population, and collection method; first-party research and ethics standards; dated public sources; visible limitations. Self-reported behavior is treated as directional. Vendor-sponsored research is never used as neutral proof of the vendor's own importance.

Executive summary

A founder writes “market leader” on the homepage. The buyer opens Google, asks an AI system for alternatives, reads two review pages, sends the shortlist to IT, and asks finance for a return case. The same three words now have to survive five different questions.

The public evidence supports that sequence in B2B technology. It does not support the usual leap from “buyers research independently” to “we know what your buyer believes.”

What holds up:

• B2B technology buyers use several sources before and during seller contact. G2's 2025 survey of 1,169 software purchase decision-makers reports review sites, AI search, Google, market-research firms, vendor sites, and peers in the enterprise research mix. Wynter's 2024 study of 100 B2B SaaS marketing executives reports that 97% visited a supplier website, 81% used online reviews, and 78% shortlisted three vendors for a demo. Both studies are vendor-sponsored and software-specific. They still support a bounded conclusion: a claim has to travel beyond the page where the company wrote it. (G2, 2025; Wynter, 2024)
• The buyer is usually a group. G2 reports participation from IT, security, finance, executives, operations, procurement, legal, and technical roles. Wynter reports a median buying group of five in its narrow SaaS sample. LinkedIn and Ipsos surveyed 887 B2B buyers across seven countries and found that 40% reported larger buying committees and that non-users participated 44% of the time. The exact mix changes by category and deal. The need for evidence that can cross roles is well supported. (G2, 2025; LinkedIn and Ipsos, 2025)
• AI is part of buyer research, with wide uncertainty around its weight. G2 reports that 79% say AI search changed how they research software. LinkedIn and Ipsos report 94% use AI somewhere in the purchase process. TrustRadius reports that 72% encountered Google AI Overviews and 90% clicked a cited source in its 2025 technology-buyer survey. These are self-reported, vendor-sponsored figures with different questions and samples. They confirm usage. They do not establish that AI has replaced search, peers, websites, or sellers. (G2, 2025; LinkedIn and Ipsos, 2025; TrustRadius, 2025)
• Source identity affects credibility. Forrester's 2023 survey reports that more than 90% of respondents thought buyers completely or somewhat trust industry peers, 85% thought buyers trust vendor customers, and more than 80% thought buyers trust analysts. Vendor salespeople ranked last among the tracked groups at 29%. The survey measured respondents' views of buyer trust rather than observed buying behavior, so the ranking is directional. (Forrester, 2023)
• Public evidence can test presence, consistency, support, contradiction, and freshness. A dated review can show whether a claim appears across owned pages, search results, AI answers, reviews, customer proof, competitor material, analyst commentary, and procurement documents. It can show whether the supporting evidence is findable and whether the same claim changes across surfaces.
• Public evidence cannot reveal private motive or predict a committee's decision. It cannot establish why a named buyer accepted or rejected a claim, how common a belief is across a market, which private conversation carried the most weight, or whether a repair will produce shortlist entry or revenue. Those questions require direct research, permissioned records, or observed decisions.
• Three comparison companies cannot prove category leadership. A comparison set can show how a claim fares against those named alternatives. A market-wide superlative needs a declared market universe, an inclusion rule, and adequate coverage of that universe.
• A neutral diagnostic needs a real path to “no action.” Professional research standards separate research from persuasion and require enough methodological disclosure for others to assess the result. A diagnostic that agitates pain, steers every answer toward a sale, or hides the scoring rule has stopped behaving like research. (Insights Association, 2025; AAPOR disclosure standards; ICC/ESOMAR Code, 2025)

The founder-level conclusion is simple: before you fund and repeat a consequential market claim, test whether a named buyer can find the proof, follow it across the surfaces they can inspect, and carry it through the roles that can stop the decision. Call the result a public-evidence judgment. Save claims about private buyer truth for research that actually observes buyers.

Claim-by-claim verification

Provenance flags

• G2, Wynter, LinkedIn Sales Navigator, and TrustRadius sell into the markets their studies describe. Their samples are useful. Their product recommendations and claims about their own channel require more skepticism.
• The studies ask different questions. “Used AI somewhere,” “AI changed research,” and “encountered an AI Overview” are three different measures. Combining them into one AI adoption number would be false precision.
• Software carries most of the quantitative weight. A founder in healthcare, industrial services, financial services, or public procurement should treat the source mix as a starting hypothesis.
• Public evidence has a hard ceiling. Findability and contradiction can be observed. Belief, motive, prevalence, causality, and a committee's eventual choice cannot be inferred from a screenshot.

What the primary sources actually say

The claim is checked before the sales call

Wynter's 2024 sample is small and narrow: 100 B2B SaaS marketing executives at companies with more than 200 employees. Inside that sample, 54% began with a category search, 41% began with vendors they already knew, and 58% used networks to build a shortlist. Ninety-seven percent visited the supplier website, 81% used online reviews, and 78% reported shortlisting three vendors for a demo. (Wynter, 2024)

The useful result is the sequence. Buyers define requirements, discover a category or known vendor, assemble a compact field, and inspect public proof. After the demo, the questions become more specific: integration, security, economics, support, compliance, operating history, references, scalability, and limitations. A founder's category claim sits near the beginning. It has to remain coherent when the questions become more demanding.

G2's 2025 sample is larger and still bounded to software. Its 1,169 decision-makers report review sites, AI search, Google, market-research firms, vendor sites, and peers in the research mix. Forty-five percent reported a shortlist of two or three vendors. Thirty-one percent reported four to seven. A seller may never get the chance to correct a claim that fails during independent research. (G2, 2025)

One claim meets several standards inside the same account

G2 reports participation from IT in 47% of software purchase decisions, security in 32%, finance in 31%, executive leadership in 30%, operations in 29%, procurement in 19%, and legal in 18%. These are participation rates in its sample, not a universal buying-group template. They show why one broad proof line rarely carries the whole decision. (G2, 2025)

The commercial champion may need category fit and business value. IT may need integration evidence. Security may need controls. Finance may need a defensible return case. Procurement may need terms and comparison. Executive approval may depend on risk, timing, and the cost of being wrong. “Market leader” needs a defined type of leadership, market, period, comparison rule, and proof source.

LinkedIn and Ipsos add a broader multi-industry sample. Their 887 respondents came from seven countries and included technology, professional services, finance, manufacturing, and other industries. Forty percent reported larger buying committees. The report says buyers arrive with compact shortlists and value seller input most during comparison and pilot stages. Proof, trade-offs, and internal defensibility become concrete there. It is vendor-sponsored self-reporting. It still matches the multi-role pattern found in the software studies. (LinkedIn and Ipsos, 2025)

AI increases the number of claim-checking surfaces

G2 reports that 79% say AI search changed how they conduct software research. LinkedIn and Ipsos report that 94% use AI at some point in a B2B purchase. TrustRadius reports that 72% of 2,058 technology buyers encountered Google's AI Overviews and that 90% clicked a cited source. (TrustRadius, 2025)

The three figures describe different actions. None earns a claim that all buyers now start with AI. Together they show a new verification path: a buyer can ask a system to name the category, compare vendors, identify risks, and surface proof sources before opening a vendor page. The AI answer matters, and its citations matter too. A company can be present in one and absent from the other.

Trust attaches to the source as well as the statement

Forrester's 2023 study asked respondents which sources they thought buyers trusted. Peers, vendor customers, analysts, existing providers, and colleagues ranked above unfamiliar salespeople. That does not mean a peer is always right or a vendor is always discounted. It means identical words can carry different weight depending on who says them and how close that source sits to the buyer's circumstances. (Forrester, 2023)

A verification review should record the source class, date, access conditions, relation to the claim, and any commercial interest. A customer case and a founder statement can agree and still carry different evidentiary weight. An analyst report and an AI answer can disagree even when they cite the same thin source. Counting mentions misses the question the buyer is actually asking: “Can I defend this claim if somebody challenges it?”

Research has to preserve the possibility of an adverse answer

The 2025 Insights Association code defines research as systematic gathering, analysis, and interpretation used to support decisions. It distinguishes research from activity intended to persuade an identified person. AAPOR requires enough disclosure about collection, instruments, population, sampling, and processing for others to assess public findings. The 2025 ICC/ESOMAR code centers legality, honesty, transparency, fit-for-purpose design, accountability, and human oversight. (Insights Association, 2025; AAPOR; ICC/ESOMAR, 2025)

Those standards change the design of a public claim check. Intake can define the claim, buyer, decision date, and known evidence. It cannot turn the founder's answers into an invented buyer-trust score. The tool can show which questions remain unanswered. It has to allow a clean result such as “the claim is supported,” “the evidence is insufficient,” “primary research is required,” or “no Petrichor work is warranted.”

What is measured versus what is inferred

The transfer rule is strict: observations can travel only as far as the method permits. A public-surface review can show the evidence available to inspect. Interviews can show what eligible participants say. Observed decision records can show how a particular decision unfolded. None should borrow the authority of the others.

Dated timeline

The steelman

The strongest case against this report is that public evidence may be the wrong object of study.

Enterprise choices can turn on private references, previous experience, a board relationship, procurement policy, product performance, security review, an internal champion, or a conversation that never becomes public. A company could have a thin public footprint and still win through an existing buyer relationship. It could have impeccable public proof and lose at a technical implementation test. B2B International and other established research firms use interviews and surveys to address perception and prevalence. Public artifacts cannot answer those questions.

The quantitative sources lean heavily toward technology. G2 and TrustRadius benefit when reviews matter. Wynter benefits when buyer research matters. LinkedIn benefits when seller expertise matters. Self-reported channel use does not prove actual influence. The most dramatic AI figures come from different instruments and cannot be pooled.

That challenge changes the conclusion. A public review should never promise “buyer truth,” market prevalence, or predicted revenue. Its valid job is narrower: reconstruct the inspectable evidence around one consequential claim, expose where it is supported or contradicted, and name the private questions that remain. When a private question could change the decision, the work should stop or add a properly disclosed primary-research method.

The authoritative sources

1. AAPOR Disclosure Standards, revised 2021. The clearest operational standard for disclosing sponsor, instrument, population, sampling, collection, processing, and limitations in public research.
2. ICC/ESOMAR International Code, 2025. A current global professional code covering transparency, fit-for-purpose design, AI, accountability, participant protection, and human oversight.
3. Insights Association Code of Standards and Ethics, updated September 2025. Defines research and non-research activity and sets duties to participants, clients, and the public.
4. G2 Buyer Behavior Report, 2025. The largest disclosed buyer sample in this review, with 1,169 software purchase decision-makers. Strongest for technology-specific source use and role participation; vendor-sponsored and incomplete for population inference.
5. LinkedIn and Ipsos, The Trust Advantage, 2025. A 887-person, seven-country, multi-industry B2B sample. Useful for the relation among AI, comparison, committees, and seller expertise; vendor-sponsored and self-reported.
6. Wynter B2B buying study, 2024. A small, narrow study with enough question-level detail to map the sequence from category discovery to final risk checks.
7. Forrester, Who Do B2B Buyers Trust?, 2023. A useful source-class ranking, with the limitation that respondents reported their view of buyers rather than observed buyer choices.
8. The AQuA Book, 2025. Government guidance for analytical specification, assurance, records, sign-off, limitations, and review. It supports process design, not a commercial outcome claim.

Frequently asked questions

What is buyer verification?

Buyer verification tests whether one consequential market claim survives the evidence a named buying group can inspect before a decision. It follows the claim across public and permissioned sources. Support, contradiction, and unobserved buyer questions stay separate.

Can public evidence show what buyers believe?

No. It can show presence, findability, consistency, support, contradiction, and freshness. Private belief, motive, informal influence, and unrecorded objections require direct research or permissioned decision records.

Can three competitor comparisons prove market leadership?

Only when those companies exhaust a declared market universe. A category-wide claim needs a defensible universe, an inclusion rule, and adequate coverage. Three named alternatives support a three-company comparison.

When does buyer verification require primary research?

Use primary research when motive, prevalence, private objections, message interpretation, or influence among buyer roles could change the decision. State the population, recruitment method, instrument, field dates, completed counts, analysis method, and limitations.

What decision should the research produce?

The decision owner should choose to carry, revise, verify, pause, or stop the claim. The record should retain the accepted evidence, dissent, first action, owner, due date, and the event that reopens the judgment.

Open questions

• Which role owns the risk of an unsupported market claim inside a post-Series A B2B company?
• Which event creates a real decision date: enterprise launch, category change, campaign lock, analyst briefing, board approval, or sales kickoff?
• Which public surfaces actually enter that category's buying process, and in what order?
• When public evidence and private win-loss records conflict, which source changes the decision?
• What minimum market-universe coverage makes a leadership superlative defensible?
• Does a buyer use a Decision Record in the meeting or workflow named at intake?
• Which repairs change a live decision, and which merely improve the appearance of the evidence?
• When does the work require interviews, a survey, technical assurance, legal review, or a different specialist?

Research notes and limitations

This report draws on four commercial buyer studies, one analyst trust survey, professional research codes, and decision-quality guidance. The buyer studies are cross-sectional and mostly self-reported. G2, Wynter, LinkedIn Sales Navigator, and TrustRadius all have commercial interests in the behaviors their reports feature. Their samples, question wording, and published denominators differ. No combined percentage has been calculated.

The report's public-evidence claims are method claims. They establish what a dated, source-traceable review can observe and where that observation ends. They do not show that a Buyer Verification engagement causes shortlist entry, revenue, category ownership, or a more favorable buyer response. Any such outcome requires its own study.

The operating question: Can this claim survive the evidence your enterprise buyer can inspect before you fund and repeat it?

Petrichor Projects | Category Gravity | petrichorgrowth.com/research