Who Gets to Make the Claim? Matching People, Proof, and Trust in Advertising Films

Advertising Film Strategy Research 07/10

The same line—“it saves me time”—can be delivered by an actor, a real customer, a paid creator, an engineer, or a brand executive. The words may not change, but the audience is being asked to make a different judgment each time. Has this person actually used the product? Are they qualified to explain why it works? Are they saying this because they were paid? Can one person’s experience stand for other customers?

Casting conversations often begin with presence, camera performance, follower count, and budget. Those factors can help answer who will look convincing or attract attention. They do not answer who is entitled to carry a particular claim. This article’s central argument is that credibility is neither a trait that a person simply possesses nor a texture created by handheld camerawork, a domestic set, or an apparently spontaneous performance. Credibility depends on the fit among source, claim, and proof. The person provides a position from which something could reasonably be known; the film and its supporting materials must show why the claim deserves belief.

This is a research-based analysis, not a peer-reviewed paper, and ONCE work is not presented as experimental evidence. Research was refreshed through 5 October 2026. The review compares source credibility, endorser fit, sponsorship disclosure, and parasocial-relationship research, then checks current advertising guidance in mainland China and the United States. Attitudes, credibility ratings, engagement intentions, and purchase intentions are not treated as actual purchases or incremental sales.

Ask what must be proved before asking who should speak

People in advertising films perform at least three different jobs. The first is staging a situation: an actor makes a use case, conflict, or relationship visible. The second is reporting experience: a real customer or creator describes something that genuinely happened during use. The third is explaining facts: an engineer, specialist, or responsible executive explains a mechanism, process, test, or organizational capability.

All three jobs can appear in one film, but they cannot impersonate one another. An actor can accurately demonstrate an approved operating procedure. If reasonable viewers are likely to interpret the actor’s first-person line as a personal experience, however, a dramatization begins to function as a testimonial. A real customer can accurately describe an individual experience, but that experience alone does not prove that everyone will obtain the same result. An engineer may understand a technical mechanism without being able to guarantee a customer’s long-term outcome. A brand executive can state an organizational commitment, but the commercial relationship should be obvious rather than disguised as independent judgment.

The US Federal Trade Commission’s current Consumer Reviews and Testimonials Rule Q&A makes a useful distinction. Actors in an obviously fictional dramatization are not automatically giving testimonials. Yet when actors portray testimonialists, the underlying testimonials still cannot be fake or false, and the presentation may be deceptive. China’s Advertising Law defines an endorser as a party other than the advertiser who recommends or certifies a product or service in their own name or image; an endorsement must be based on facts, and endorsers may not recommend products they have not used or services they have not received. The precise legal treatment depends on market, category, and execution, so project counsel should assess the actual film.

Before a casting session, it is therefore useful to label every consequential line. Is it dramatic dialogue, personal experience, professional judgment, or a brand promise? Once those identities blur, natural performance cannot repair the confusion.

Credibility is not one score; it is four questions

Ohanian’s 1990 research developed a perceived-celebrity-endorser scale built around expertise, trustworthiness, and attractiveness. It remains influential, but the scale measures how audiences perceive endorsers; it does not establish that an endorser’s statement is true. Later reviews similarly suggest that high-credibility sources often have a persuasive advantage, while also showing interactions with the message, channel, receiver, and communication task. Source credibility does not operate alone in every setting.

For advertising-film production, “Is this person credible?” can be replaced with four more diagnostic questions:

These are not points to add into a universal score. A relatable person is not automatically qualified to assess a technical claim. A genuine expert still needs to disclose a material relationship. A customer’s experience can be wholly truthful without supporting a population-wide comparison.

Fit is more useful to test than who looks most like a user

Schouten, Janssen, and Verspaget compared social media influencers with traditional celebrities in two experiments (N=131 and N=446) and manipulated product–endorser fit. Participants reported greater similarity, identification, and trust toward influencers, and those variables were related to advertisement attitude, product attitude, and purchase intention. Product–endorser fit did not, however, consistently explain all differences across both experiments. The research suggests that a closer relational distance can help creators, but self-reported intention in an experiment is not sales evidence and does not show that an influencer will always outperform a celebrity.

A 2025 PLOS ONE study included four quasi-experiments with 473 participants in sporting contexts. Product–endorser fit, popularity, self-congruity, and similarity moderated the relationship between endorser type and attitudes toward the advertisement; likeability did not show the same moderating effect. Its sporting categories, quasi-experimental design, and attitude outcomes cannot provide a casting formula for every market. The more defensible lesson is that being likeable or popular does not make a person suitable for every claim.

In a 2026 online experiment, Nguyen and Mero showed 338 Instagram users posts featuring a fictitious influencer and an insulated bottle. Influencer–audience fit significantly predicted perceived credibility, which partially mediated behavioral-engagement intention. Sponsorship disclosure did not significantly weaken the relationship between fit and credibility. Because the study used a fictitious account, one product, brief exposure, and mainly perceptual or intentional outcomes, it cannot establish return on ad spend. It does suggest that clear disclosure does not necessarily erase a relevant relationship that already makes sense to an audience. Concealing the commercial relationship is not a sound method for protecting authenticity.

Parasocial relationships also need to be interpreted carefully. Aw and Fernando’s 2026 systematic review included 76 peer-reviewed studies published between 2000 and 2023. The authors found that much of the literature examines relationship formation and non-transactional outcomes, while relatively few studies investigate transactional outcomes such as purchasing, impulse buying, or gifting. Familiarity can make audiences willing to listen; it is neither technical qualification nor causal sales proof.

Six kinds of on-screen source, each with limited authority

Actor-led dramatization is useful for making a problem, action, and outcome legible, particularly in narrative, lifestyle, and standardized demonstration films. An actor can show how something is used. They should not invent a personal history of long-term use. If the film uses the first person, reasonable viewers should be able to recognize the dramatized identity.

A real-customer testimonial is useful for describing an actual experience in a defined context. A responsible script preserves product version, duration of use, situation, and limitations rather than rewriting the customer’s account into a universal promise the brand cannot support. One case can show that something happened; it usually cannot establish that most users will obtain the same outcome.

Paid creator content can place a product inside a task and language already familiar to the creator’s audience. Payment, gifts, and content control need to be clear. Liu and Zhao’s 2025 meta-analysis synthesized 288 effect sizes from 37 studies involving 12,721 participants. Sponsorship disclosure, on average, increased persuasion knowledge and brand recall and reduced credibility, while overall effects on attitudes and conative responses were not significant. Disclosure effects varied with disclosure characteristics, stimuli, audiences, and controls. There is no fixed cross-context performance answer, and legal compliance cannot be decided by an average effect.

A professional or internal expert is useful for explaining methods, engineering, risk, and limitations. Expertise has to match the claim. An industrial designer can discuss space and interaction but is not automatically qualified to establish a medical outcome. An engineer can explain structure but does not automatically represent user preference. The expert is most useful when the film shows how a conclusion was reached, not merely what the person is wearing.

Organic user content has value because the original expression was not controlled by the brand. Once a brand pays to license it, re-edits it, adds a purchase link, or distributes it as advertising, the new context may turn that material into an ad or testimonial. Provenance, permission, material connections, and editing scope must be checked again. A favorable sentence cannot simply be removed from its context.

Brand-produced UGC-style film remains produced advertising. Direct address, a phone aspect ratio, pauses, and ambient sound alter the distance of expression; they do not turn an actor into a customer. The form can honestly use platform-native language, but it should not manufacture the impression of an independent recommendation.

Let claim strength determine the required level of proof

A person cannot substitute for evidence. The more objective, comparative, and consequential the claim, the more the burden should move from “this person said it” to “this is what viewers can inspect.” That extends the claim–evidence contract developed in Episode 04: the person carries the source role inside that contract.

• “I like how it feels” is a personal experience that a real user can express within a defined situation.
• “Installation takes three steps” requires a complete, uninterrupted procedure and its applicable conditions.
• “Faster than the previous model” requires a defined comparator, metric, sample state, and test protocol.
• “Suitable for a professional task” requires a qualified person to explain the task standard and show the critical process.
• “Most users are more satisfied” requires a defensible sample and aggregation method, not a montage of selected reactions.

China’s Administrative Measures for Internet Advertising require internet ads to be identifiable. When products or services are promoted through knowledge sharing, experience sharing, or consumer reviews and are accompanied by a purchase link or another purchasing method, the content must be prominently identified as advertising. The State Administration for Market Regulation’s 2024 enforcement guidance further explains that the presence or absence of an “advertisement” label is not the sole basis for deciding whether commercial information is an ad; information feeds, audiovisual content, and livestream settings still need to make the advertising nature recognizable. The official texts were checked on 5 October 2026.

In the United States, the FTC’s Endorsement Guides, revised in 2023, require clear and conspicuous disclosure of material connections that are not reasonably expected and could affect the weight or credibility audiences give an endorsement. The current Consumer Reviews and Testimonials Rule also prohibits buying or disseminating reviews and testimonials that a business knew or should have known were fake or false. These markets have different rules, but the production lesson is shared: a film cannot first create the impression of independent judgment and expect a marginal disclaimer to repair it.

A public case: a person can ground the creative context without proving every product claim

ONCE’s public SKYWORTH Wallpaper TV × Yang Huazhen case organizes its content around a Qiang embroidery practitioner, a natural environment, needlework details, and imagery on the screen. For this review, we read the case-page text and visible imagery. We did not play the complete film or audio and did not review the original brief or performance data.

Within that observable scope, the source fits the subject. Yang Huazhen can give the craft patterns, materials, and cultural context a grounded origin; her hand movements and work are not merely a famous face placed next to the product. This does not mean that the person herself proves the television’s color accuracy, durability, or sales effect. The artist can ground a cultural and spatial proposition. If product performance is claimed, demonstration, technical records, or testing must carry that separate burden.

The same division of responsibility applies to corporate films. A founder can explain a choice and commitment; an engineer can explain a technical route; a delivery lead can explain process; and a customer can report an actual collaboration. ONCE’s guide to corporate-film interviews and verifiable business facts puts people, processes, and records in one production structure. It is a production method, not experimental evidence of credibility effects.

Run a source-switch test before production

The following is an experiment suggestion—a relationship to be tested, not an ONCE client experiment.

Select one claim that materially affects the buying decision. Produce three animatics or low-cost pilots while holding information, product action, duration, brand presence, and approximate production quality as constant as possible. Version A uses an actor and is clearly dramatized. Version B uses a real user speaking within the limits of an actual experience. Version C uses a professional whose expertise matches the claim. Commercial relationships remain clearly disclosed; concealment is not used as an experimental treatment.

First measure whether viewers correctly identify the source. Then measure claim comprehension, proof recall, perceived qualification, counterarguments, and the next relevant behavior. If the real-user version feels more relatable without improving comprehension, the weakness may be missing proof rather than performance. If the expert version appears competent but raises questions about commercial motives, increase method transparency rather than merely hiring a more fluent expert. If none of the three versions allows viewers to explain the basis of the claim, reduce the claim or reshoot the proof.

The point is to put “who has the most reach?” back in its proper place. Reach affects how many people may see the film. Fit among source, claim, and proof affects whether those people have a reason to believe it.

Conclusion: assign the burden of proof before assigning the center of the frame

One advertising film does not need one person to carry every kind of trust. An actor can stage the situation. A customer can provide a bounded account. A creator can connect to an existing audience. An expert can explain a method. The brand can own its promise. The product and supporting materials can provide inspectable facts.

The danger is not using actors or paid creators. It is presenting them as a different kind of source. The danger is not polished cinematography. It is letting polish replace the proof that should be visible. The decisive planning question is not “who looks most like a real user?” but who is qualified to make this claim, and what else must the camera show?

If your project needs to put source roles, claims, proof, disclosure, and multi-version delivery into one production plan, see ONCE’s TVC and campaign production approach. Start with the communication task, then assign the people—rather than starting from a talent list and inventing the rationale later.

References and access scope

1. Roobina Ohanian (1990), “Construction and Validation of a Scale to Measure Celebrity Endorsers’ Perceived Expertise, Trustworthiness, and Attractiveness,” Journal of Advertising, 19(3), 39–52. DOI 10.1080/00913367.1990.10673191. The formal abstract and bibliographic record were read; the scale is not treated as proof of truth.
2. Alexander P. Schouten, Loes Janssen, and Maegan Verspaget (2020), “Celebrity vs. Influencer Endorsements in Advertising,” International Journal of Advertising, 39(2), 258–281. DOI 10.1080/02650487.2019.1634898. The open full article was read for design, samples, results, and limitations.
3. Chen-Yueh Chen et al. (2025), “Social Media Influencer or Traditional Celebrity?” PLOS ONE, 20(7): e0326911. DOI 10.1371/journal.pone.0326911. The open article’s four quasi-experiments, sample, and results were read; the findings are not extrapolated into a cross-category sales rule.
4. Chuong H. B. Nguyen and Joel Mero (2026), “The Role of Influencer–Audience Fit in Shaping Influencer Credibility Across Disclosure Types,” Journal of Marketing Communications, published online 19 August 2026. DOI 10.1080/13527266.2026.2714915. Visible full-text methods and results were read; stimuli were fictitious Instagram posts.
5. Xia Liu and Hong Zhao (2025), “A Meta-Analysis of the Effects of Sponsorship Disclosure in Influencer Marketing,” Marketing Letters, 36, 519–532. DOI 10.1007/s11002-024-09757-z. The formal abstract and meta-analytic scope were read; the full article was restricted.
6. Eugene Cheng-Xi Aw and Angeline Gautami Fernando (2026), “Consumers’ Parasocial Interactions and Relationships with Online Influencers: A Systematic Review,” Journal of Consumer Marketing, 43(5), 653–674. DOI 10.1108/JCM-02-2024-6611. The formal abstract and methodological scope were read.
7. China’s Advertising Law, the Administrative Measures for Internet Advertising, and SAMR’s Guidelines for Enforcing Internet-Advertising Identifiability; the US FTC Endorsement Guides and Consumer Reviews and Testimonials Rule Q&A. Official pages were checked on 5 October 2026. Markets, categories, and fact patterns differ.