A claim becomes a publishing decision
“Our proven process doubles qualified enquiries within three months.”
That sentence might appear in a service-page heading, a campaign landing page or a proposal. It contains several objective claims. The process is proven. Enquiries double. They are qualified. The result happens within a defined period. A buyer may reasonably rely on every part of that message.
Before the sentence goes live, somebody needs to decide whether the business can support it. That decision requires the proposed wording, its likely meaning, the evidence and a person accountable for approval.
The ACCC guidance on false or misleading claims says business claims must be accurate, truthful and based on reasonable grounds. The guidance covers websites, social media and testimonials. It also says omissions can mislead and asks businesses to check the overall impression, note important limitations and be prepared to prove their claims.
This practical pre-publication workflow is general operational information for Australian service businesses, not legal advice. Claims involving health, safety, finance, guarantees, regulated services or material customer harm need review suited to their risk. Seek qualified legal advice when the consequence or applicable rule is unclear.
Strong substantiation also supports clear, useful content. It is one part of what makes content worth citing, though citations and sound evidence do not guarantee rankings, AI citations or recommendations.
Find every claim before you assess the evidence
Review the draft in the form a reader will encounter it. Start with headings and body copy, then include testimonial selections, charts, calls to action, image text, page titles, descriptions and adjacent design. A restrained sentence can create a stronger implied message when it sits beneath an impressive number or beside a customer quote.
Record each material claim separately. Capture the exact wording, page placement, intended audience and likely implied meaning. Include statements about numbers, performance, credentials, comparisons, prices, delivery times, guarantees and future outcomes. Also capture broad labels such as “proven”, “leading”, “secure” or “sustainable” when a reasonable buyer could read them as factual.
The US Federal Trade Commission's advertising guidance provides a useful comparison. It says reviewers consider express and implied claims, omitted information and the full context of words and images. Advertisers should hold evidence before an advertisement runs. This is US guidance rather than Australian law, but its distinction between literal wording and likely meaning is a useful editorial check.
For the opening example, create separate records for the process, the increase, the definition of a qualified enquiry and the three-month period. One case study or analytics screenshot should not silently stand in for all four.
This inventory happens while copy is still in review. If unsupported claims are already spread across the live site, use the separate process to audit claims already live on the website.
Let risk determine the depth of review
Two questions make review effort proportionate. First, what could happen if a buyer relies on the claim and it is wrong? Second, how quickly can the supporting evidence stop being true?
Buyer consequence rises when a claim affects money, safety, eligibility, rights or a hard-to-reverse choice. Evidence volatility rises when the source depends on prices, availability, platform rules, current credentials, changing performance or a future event. A stable office address has low volatility. A promise about delivery capacity next month has much more.
These axes create four distinct review areas. Seeing them together helps the team decide when an ordinary evidence check is enough and when approval needs a more accountable route.
Higher risk claims need deeper approval
Buyer consequence increases
Verify
Higher consequence with stable evidence. Confirm exact scope and accountable approval.
Escalate
Higher consequence with volatile evidence. Require deeper review and consider qualified legal advice.
Check
Lower consequence with stable evidence. Use a proportionate evidence check and named owner.
Qualify
Lower consequence with volatile evidence. Narrow the wording and set an early review trigger.
Evidence volatility increases
The matrix sets review depth. It does not determine whether a claim is legally compliant. A low-consequence claim can still be false. A high-consequence claim may be publishable when the evidence, wording and specialist review are appropriate.
Match the evidence to the exact claim
Open the underlying evidence and compare it with the proposed words. Check whether it supports the same service, audience, location, timeframe, conditions and level of certainty. The full guide to evaluating whether a source supports the exact claim covers authority, independence, methodology, freshness and directness in more depth.
Consider a web agency that wants to say, “Our redesign process improves lead quality.” Three kinds of support produce three different decisions.
- A CRM analysis with a defined qualified-lead measure, comparable periods and documented project changes may directly support a bounded result for the projects measured.
- A client's comment that enquiries felt better may support an attributed experience. It provides only partial support for a general performance claim.
- A report showing that website visitors value easy navigation discusses a related idea. It does not prove that this agency's redesign process improves lead quality.
Client outcome claims also need clear measurement, attribution and permission. Use the specialist guide to substantiate client outcome claims when the proposed proof comes from an engagement.
Evidence strength should limit wording strength. A record of five completed projects can support facts about those projects. It cannot establish that every future client will get the same outcome. If the support covers organic form enquiries, the claim should not expand to all leads. If the analysis shows an association, avoid causal words such as “delivered” or “caused”.
Choose one recorded approval outcome
Every material claim should leave review with one named outcome.
- Approve when current evidence supports the exact wording, implication and placement, and an accountable reviewer accepts the limits.
- Qualify when the claim remains useful after a condition, measurement window, audience or limitation is made clear.
- Rewrite when the evidence supports a narrower or different statement.
- Escalate when the potential consequence, regulatory context or disagreement exceeds the team's competence or authority.
- Remove when support is absent, the risk is unacceptable or the claim adds little value.
Apply those outcomes to the opening sentence. If project records show different results and timeframes, remove the promise to double enquiries. A supportable rewrite might say, “We agree how qualified enquiries will be measured before redesign work begins, then compare the result against the recorded baseline.” It makes a useful process claim without inventing an outcome.
A smaller claim is often more informative. Scope, conditions and measurement tell the buyer what the business actually does. Empty confidence asks them to supply the missing proof.
Put qualifications where readers can use them
A qualification belongs close enough to shape the claim's meaning when the reader encounters it. The ACCC says fine print and qualifications must not conflict with the overall message of an advertisement. Its broader guidance asks businesses to note important limitations and check the overall impression.
The FTC's US guidance similarly says a disclosure should be clear and conspicuous. Treat that as a comparative editorial principle. It does not state Australian law. Review prominence, proximity, wording and the page's visual hierarchy together.
“Guaranteed launch in six weeks” cannot be repaired by a footer that excludes delayed approvals, missing content and third-party integrations. If those conditions define the promise, put them beside it and change the main wording. “Six-week target after content approval for projects within the agreed scope” conveys a different and more bounded offer.
Test the completed page on mobile as well as desktop. Accordions, truncated cards, sticky calls to action and chart labels can separate a limitation from the claim it governs. Ask a reviewer who did not write the copy to explain the offer and its conditions. Their answer reveals the impression the page creates.
Keep a claim record another person can follow
Approval needs a durable record because the author, subject expert and future page owner may be different people. The W3C PROV-O Recommendation models provenance through entities, activities and the people or organisations responsible for them. Its Data Quality Vocabulary describes ways to express quality measures, annotations, provenance and policies.
You don't need a semantic-web implementation to use that logic. A lightweight editorial record can connect the final claim, supporting evidence, responsible people, decision and later review activity. Record evidence limits in words so the judgement stays visible beyond a pass or fail label.
At minimum, save the final wording and implied meaning, page and placement, evidence link or internal record, evidence limit, owner, reviewer, approval outcome, approval date and review trigger. Keep confidential project data in an access-controlled location while giving authorised reviewers a reliable path to it.
Google's people-first content guidance asks publishers to consider clear sourcing, demonstrable expertise, accurate authorship and information about how content was produced. A traceable claim record helps an editorial team answer those quality questions. Rankings and inclusion in an AI-generated answer remain separate decisions.
Hand approved claims into maintenance
Approval describes a claim at a point in time. Give every volatile or consequential claim a date-based trigger, an event-based trigger or both.
Reopen the decision when a source changes, a credential expires, the service scope moves, pricing changes, a platform updates its rules, a new campaign places the wording in a different context, a complaint identifies a misleading impression or an evidence error comes to light. Regulator and platform pages cited in this article should be checked by 7 October 2027, and earlier if their guidance changes.
The full method to keep approved evidence accurate after publication covers monitoring, corrections and refresh decisions. When a team inherits claims without approval records, start with the published website evidence audit and rebuild the missing trail.
Make substantiation part of the publishing system
Put the claim check before final copy approval. Give the writer responsibility for identifying claims, the subject expert responsibility for evidence and scope, and the publisher responsibility for recording the decision. Define who can escalate a claim and who has authority to approve it.
Start with pages closest to a buying decision. Review service promises, outcome statements, comparisons, credentials, guarantees and time-sensitive facts. Save the approved wording with its evidence and trigger. Refuse publication when a material claim has no owner or decision.
Repeated evidence gaps may show a wider content strategy problem. Evidence-led search strategy can connect buyer questions, source standards and measurement before pages are drafted. If records, ownership and review dates cannot travel through the CMS, a governed website publishing system can make those controls part of templates and workflow.
The publishing decision should be explainable in one sentence. We approved this claim because this evidence supports this wording under these conditions, and this person will review it when this trigger occurs. If the team cannot complete that sentence, the claim is not ready to publish.