Social Short-Form Video Marketing Research 03/10 | Research article / work analysis
Imagine two Mother’s Day videos of equal length. The first spends twenty seconds on unspoken affection between a mother and daughter, then introduces the gift and brand in the final five. The second keeps the same relationship and emotional arc, but choosing, ordering, delivery, and use are what resolve the tension. Both may hold attention. Both may move people. The material difference is this: if the product were removed, could the story still happen in exactly the same way?
This is not a demand for more logo exposure. Our central claim is that emotion may bring a viewer into a story, but narrative becomes a possible reason to buy only when the product plays a non-substitutable role in the causal chain and the key promise receives observable support. To turn that claim into an inspectable creative method, this article proposes three bridges—entry, agency, and verification—and separates story liking, product understanding, claim credibility, and behavior.
Scope: not a contest between storytelling and hard sell
Research was checked through 5 October 2026. The material includes the original Elaboration Likelihood Model book, classic narrative-transportation experiments, recent social-video and brand-attitude studies, and one study that connects video imagery with e-commerce sales. We also read ONCE’s public solution and case pages. We did not have access to client media accounts, audience experiments, unpublished scripts, or campaign sales data.
The article can therefore compare mechanisms and evidence boundaries and propose falsifiable creative hypotheses. It cannot establish that one narrative structure universally increases sales. “Reason to buy” here does not mean purchase itself. It means that a viewer can connect a personally relevant problem, the product’s role in it, and a claim that can be checked. The previous study on attention, retention, and product understanding showed why stopping, continuing, and understanding are different outcomes. This article adds the next distinction: understanding the plot is not the same as believing the product.
What a story can change—and what it cannot prove
In four experiments, Green and Brock described narrative transportation as a state combining attention, affect, and mental imagery. More transported participants generated fewer counterarguments and reported beliefs or evaluations more consistent with the narrative. The materials, however, were not short-form social-commerce ads, and the outcome was not a purchase. Transportation helps explain temporary acceptance of a story world; by itself, it cannot establish that a product claim is true. The accessible PubMed record for the 2000 paper also makes the four experiment sample sizes and outcome scope visible.
Petty and Cacioppo’s 1986 book-length account of the Elaboration Likelihood Model should not be reduced to “rational information takes the central route, emotional stories take the peripheral route.” Motivation and ability affect the extent and character of elaboration, but emotion can operate as a cue, influence how arguments are processed, or constitute relevant evidence in context. Routes are not two fixed audience segments, and they are not two fixed film genres.
Read together, the theories reveal a missing step in much short-form creative reasoning. Transportation concerns entry into a narrative world. The ELM concerns how information is processed under different conditions. Neither automatically answers why a product deserves belief. A video can produce involvement while leaving the brand as a replaceable end-card sponsor. It can also contain many claims that remain inert because they have no relationship to what the character is trying to do.
Three bridges: entry, agency, verification
We separate the journey from story to product reason into three bridges that can fail independently.
| Bridge | Creative task | Observable question | What it cannot prove alone |
|---|---|---|---|
| Entry | Use goals, conflict, emotion, or curiosity to create relevance | Does the viewer continue, and understand what the character is trying to solve? | Product understanding or claim credibility |
| Agency | Let the product change a choice, capability, or event | If the product is removed, does the causal chain break? Does the action express a concrete function? | Whether the function holds in real use |
| Verification | Return emotion to an observable use, result, boundary, or piece of evidence | Can the viewer explain why this product mattered? Is the key result inspectable? | Incremental purchase |
“Non-substitutable” does not mean characters should repeat a brand name. A better test is counterfactual. If the product were replaced by any category substitute—or removed entirely—could the character achieve the same goal in the same way? If yes, the story may be building category emotion; the brand still needs a route to ownership. If no, there is another question: did a real capability change the plot, or did the script simply assign the product magical powers? Verification exists to police that difference.
A cleaning-device story, for example, can declare that the product resolved a stain. But if the edit removes contact, path, and before-and-after conditions, viewers receive a plot conclusion rather than a testable demonstration. The reverse is also true: a complete demonstration does not have to be emotionally empty. The character’s reason for caring is what gives the function consequence.
What one study close to real transactions can—and cannot—tell us
Jeremy Yang, Juanjuan Zhang, and Yuhan Zhang’s “Engagement That Sells: Influencer Video Advertising on TikTok” was published online on 20 December 2024 and appears in Marketing Science 44(2), 2025. The study uses Chinese Douyin influencer videos and Taobao sales data from May to November 2019. Its product–engagement, or PE, score uses computer vision to compare product placement with visually salient areas associated with engagement. The authors then exploit variation in posting time in a difference-in-differences design to estimate post-video sales changes for 259 treated products.
The useful result is a separation, not a viral formula. In this sample, aggregate likes, comments, and shares were not significantly related to the 30-day before–after sales difference. A higher alignment between the product and the image regions associated with engagement predicted the authors’ estimated sales lift. Yet the PE score alone explained only about 3% of the variance, and creative treatment was not randomly assigned. The research makes the relationship between “where engagement is generated” and “where the product appears” worth measuring. It does not establish that placing a product inside a visual high point causes sales.
The idea that the product should carry causal agency in the story is ONCE’s creative interpretation of that finding, not the paper’s original claim. We advance it because spatial co-occurrence may still be decorative placement. When a product action changes what a character can do, the emotional moment has a clearer path into product meaning. That proposition remains to be tested.
Frontier evidence does not supply a universal recipe
A 2026 study by Mónika Imreh-Tóth and colleagues, Driving Purchase Intentions Through Visual Storytelling: A Study of Social Media Platform Reels Sponsored Advertising, surveyed 393 active users in India’s National Capital Region and modeled relationships among interesting content, scenario, usefulness, participation, attitude, and purchase intention. It brings visual storytelling into a recent Reels setting, but the design uses self-report and purchase intention rather than randomized exposure and transaction behavior. It can locate candidate relationships. It cannot turn those paths into universal creative causes. The publisher page was not reliably readable in this run, so we use only the bibliographic and method information exposed by the formal search record and do not place an uncertain link in the article.
Madhurima Deb, Aruni Ghosh, and Yupal Shukla’s 2025 study, whose formal abstract is available from Wiley, connects narrative transportation and retrospective reflexivity with brand attitude across experiments, a survey, and a field study, and reports moderation by self-construal. The outcome is brand attitude, not purchase. A useful counterexample comes from Strick and colleagues’ experiments: the PubMed abstract reports that moving music increased transportation and some behavioral intentions but did not correspondingly improve memory, beliefs, or explicit attitudes. Emotion may shift intention without improving product understanding.
The literature is therefore not a vote on whether “storytelling works.” Stories alter the processing environment. The outcome measured, category risk, prior knowledge, platform setting, and identification strategy determine what a study has actually observed.
Involvement and risk change the load on each bridge
In a familiar, low-risk category, viewers may devote little cognitive effort to the decision. Affect, familiarity, and processing fluency can carry more weight, but brand ownership can still fail. A story everyone likes but attributes to a competitor has distributed content, not completed brand persuasion.
High-risk, high-ticket, or unfamiliar products often increase motivation to think, but they do not automatically give viewers the ability to evaluate. People may want to investigate while lacking vocabulary, comparison standards, or access to verification. These cases need a legible product action, clear conditions, and a next verification step. Splitting the work into an emotional film and an unrelated specification film leaves the bridges disconnected.
Where prior attitudes are strong, narrative can also support selective interpretation. Creators should treat reduced counterarguing as a mechanism that creates responsibility, not as an unconditional success metric. Lower resistance does not make a claim more truthful.
Case boundary: when the product enters the plot
ONCE’s public Huawei Mall Mother’s Day TVC case page describes a story structured around “choice”: a mother chooses food and clothing for her family, while the family chooses a gift for her. The page further describes online selection, ordering, delivery, and a new phone as the gift. We did not obtain a complete film in a form that allowed frame-by-frame verification in this run. Our analysis is therefore limited to the disclosed story structure; it does not assess editing, performance, sound, audience response, or campaign effectiveness.
Within that boundary, the mall and the product do more than sign the ending. Selection and delivery give the family a way to answer the mother’s everyday care. Remove those actions and the second half of the causal chain must be rewritten. That is an example of the agency bridge. It does not prove brand attitude or sales. Verification would still require asking whether viewers can explain what the platform enabled and whether delayed memory assigns that meaning to the correct brand.
The same production logic appears in ONCE’s TVC and commercial production solution: define the character’s objective, the product’s dramatic role, and the conditions that need to be shown during script development, instead of attempting to repair weak brand linkage with an end card after the film is complete.
A minimum test designed to be allowed to fail
If a budget permits only one focused test, create two versions that are as similar as possible in talent, duration, music, offer, audience, and first-frame brand cue:
- Version A preserves the emotional arc but introduces the product only at the end.
- Version B lets one specific, observable product action change the resolution without increasing the number of claims.
Under randomized paid exposure, preregister two primary outcomes: unaided explanation of why the product mattered, and correct brand attribution. Completion, story liking, click, and immediate intention are secondary. Incremental purchase requires transaction linkage or a holdout design and should be estimated separately. If B improves product comprehension but sharply damages viewing, the agency bridge may have been written as an instruction manual. If comprehension does not differ, the supposed causal action may not be legible. If B improves clicks without improving brand attribution, persuasion is not complete.
Recommendation dynamics, interference between users, statistical power, and multiple comparisons remain constraints. This is an experiment proposal, not an automated media recommendation and not an ONCE result.
Conclusion: if the product disappears, does the story break?
The dangerous shortcut in short-form narrative is to merge “finished, felt, and liked” with “believed, remembered, and bought.” The three bridges keep the outcomes separate. Emotion creates entry. Product action changes the event. Evidence makes the promise inspectable. Even if all three hold, price, distribution, alternatives, and timing still shape purchase. But if the second bridge can be removed without changing anything, the brand should at least know that it is primarily buying emotional adjacency, not product persuasion.
Leave three questions in the script review: What is the character trying to solve? Which step does the product actually change? What allows the viewer to believe that change? For a project that needs one narrative system across brand proposition, product action, and platform versions, ONCE’s commercial-production solution above is the practical starting point. The first working session should bring real product constraints, not only a mood board.
Sources and reading scope
- Richard E. Petty and John T. Cacioppo, Communication and Persuasion: Central and Peripheral Routes to Attitude Change, Springer, 1986.
- Melanie C. Green and Timothy C. Brock, “The Role of Transportation in the Persuasiveness of Public Narratives,” Journal of Personality and Social Psychology 79(5), 2000, DOI 10.1037/0022-3514.79.5.701. PubMed abstract and disclosed samples read.
- Jeremy Yang, Juanjuan Zhang, and Yuhan Zhang, “Engagement That Sells: Influencer Video Advertising on TikTok,” Marketing Science 44(2), 2025, DOI 10.1287/mksc.2021.0107. Open full methods, results, and limitations read.
- Mónika Imreh-Tóth et al., “Driving Purchase Intentions Through Visual Storytelling: A Study of Social Media Platform Reels Sponsored Advertising,” Cogent Business & Management 13(1), 2026, DOI 10.1080/23311975.2026.2639685. Bibliographic and method record only.
- Madhurima Deb, Aruni Ghosh, and Yupal Shukla, “Storytelling Ads, Narrative Transportation, Retrospective Reflexivity, and Attitude Towards Brands,” Psychology & Marketing 42(7), 2025, DOI 10.1002/mar.22204. Formal abstract read.
- Iris Strick et al., “Striking the Right Chord: Moving Music Increases Psychological Transportation and Behavioral Intentions,” Journal of Experimental Psychology: Applied 21(1), 2015, DOI 10.1037/xap0000034. PubMed abstract read.
