Every day, consumers in the United States encounter somewhere between 4,000 and 10,000 marketing messages. With that kind of saturation, it is no surprise that two-thirds of consumers say they want to receive fewer of them. For the people writing those messages, this creates an urgent problem: how do you get someone to actually pay attention, and then be persuaded?
One of the oldest debates in marketing is whether persuasive messages work better when they present logical arguments and hard facts or when they wrap information inside an engaging story. A new multimethod investigation, published in the Journal of the Academy of Marketing Science, set out to resolve this question at scale. Led by Davide C. Orazi of Monash University in Australia, a team of seven researchers synthesized decades of persuasion research and found that stories generally outperform facts alone. But their most striking finding may be this: messages that blend storytelling with factual arguments appear to be the most persuasive of all.
Two roads to persuasion
Before getting into the results, it helps to understand two concepts that sit at the center of this research. When people encounter a persuasive message, they tend to process it in one of two ways. The first is called analytical processing, which involves rationally evaluating the arguments presented. Think of reading a product spec sheet and weighing whether the features justify the price. The second is called narrative processing, which involves becoming absorbed in a story, identifying with its characters, and experiencing emotions along the way. Think of watching a commercial that follows someone training for a marathon and feeling inspired to buy the shoes they are wearing.
Decades of research have examined each pathway independently, but the field has lacked a comprehensive answer to a practical question: which one actually works better, and does the answer change depending on the format of the message or the kind of product being promoted? Orazi and his colleagues, including Anne Hamby of Boise State University, Dennis Herhausen of Vrije Universiteit Amsterdam, Tom van Laer of the University of Sydney, Stephan Ludwig of Monash University, Chahna Gonsalves of King’s College London, and Dhruv Grewal of Babson College, set out to fill that gap.
How the team investigated the question
The research unfolded in three stages. In the first stage, the team conducted a meta-analysis, which is a statistical method for combining results from many independent studies to identify overall patterns. They searched multiple academic databases, cross-referenced citations, and combed through unpublished dissertations and working papers. After screening thousands of articles, they arrived at a dataset of 227 published and unpublished works, spanning from 1978 to 2023. These works contained 767 distinct statistical relationships drawn from 308 individual studies involving a combined total of 54,715 participants.
The team organized the factors that influence persuasion using a well-known communication framework called the SMCR model, which stands for Source, Message, Channel, and Receiver. Under this framework, they categorized 15 different factors, or antecedents, that prior research had linked to how people process persuasive information. These ranged from characteristics of the person receiving the message (like their prior familiarity with a product or their personal need for mental stimulation) to characteristics of the message itself (like the quality of its arguments or whether it used figurative language).
In the second stage, the researchers surveyed 198 marketing professionals recruited through the online platform Prolific. These managers had an average of eight years of experience in marketing or sales. The survey asked them to predict how various factors would relate to analytical and narrative processing, and it also asked them to identify their most pressing unanswered questions about persuasion. The purpose was to compare what practitioners believe with what the accumulated evidence actually shows.
In the third stage, the team ran a controlled experiment with 450 U.S. residents. Participants were shown an advertisement for a pair of sneakers, a product chosen because it carries both practical and pleasure-related qualities. The ad was presented in one of three message formats: purely analytical (objective facts about the shoes), purely narrative (the same information delivered as a first-person story), or blended (a narrative introduction followed by factual claims). Each format was delivered either as written text or as audio. Participants then reported their willingness to buy the shoes.
What the data revealed
The meta-analysis produced a clear headline result: narrative processing was associated with a stronger overall link to persuasion outcomes than analytical processing. Across all studies, narrative processing showed an average correlation of 0.39 with persuasion, compared with 0.24 for analytical processing. In plain terms, storytelling-based messages were more consistently associated with favorable responses like positive attitudes and purchase intentions.
But the picture became more interesting when the team examined how context changed the equation. Communication format mattered. Analytical processing was associated with stronger persuasion when messages were delivered as audio, such as podcasts or radio ads. Narrative processing, on the other hand, was associated with stronger persuasion when messages were delivered in written form, such as blog posts or emails. The researchers proposed that reading requires more cognitive effort than listening because readers must decode text into sounds and then into meaning. That extra effort, when applied to a story, may deepen the reader’s engagement and immersion.
Product type also played a role. For hedonic products, those purchased primarily for pleasure or sensory enjoyment, narrative processing was linked to significantly stronger persuasion. For utilitarian products, those purchased mainly for practical function, analytical processing showed no significant advantage over narrative processing. This means storytelling appears to hold up well across product categories, but it is especially linked to favorable outcomes when promoting things people buy for enjoyment.
The experiment in the third stage added a new dimension. Written messages that blended storytelling with factual content were associated with the highest willingness to buy. In the written condition, blended messages outperformed purely analytical messages to a statistically significant degree. For audio messages, the pattern was different; analytical and blended messages performed similarly, and both directionally outperformed purely narrative audio messages.
Managers’ intuitions: right about half the time
One of the study’s most notable findings came from comparing the meta-analytic results with the predictions of the 198 marketing professionals. The managers’ intuitions aligned with the data about half the time. They correctly anticipated, for example, that narrative processing would be more effective for hedonic products and that media richness (the sensory variety of a communication channel) would be more closely associated with narrative processing.
But on several points, the managers got it wrong. A large majority, 72%, predicted that analytical processing would be more persuasive in written formats. The data showed the opposite: narrative processing was more strongly linked to persuasion in written formats. Managers also expected that factors like a consumer’s prior familiarity with a product would be more tightly linked to analytical processing. The meta-analysis showed that familiarity was actually more strongly associated with narrative processing.
When asked to name their most pressing question about persuasion, the most common response, from 28% of the managers, was about whether analytical and narrative elements could be combined in the same message. This question directly inspired the experiment that ultimately found blended messages to be the most persuasive written format tested.
What this means for business
The findings offer several actionable directions for marketing teams. First, companies that rely heavily on written channels like email newsletters, blog content, or social media captions may want to prioritize narrative elements. The data suggest that stories are especially effective in written form. Second, for audio channels like podcasts, radio, or voice assistants, a more fact-driven, argument-based approach appears to hold an advantage.
Third, and perhaps most practically, the experiment suggests that the choice between facts and stories may be a false one. Written messages that open with a narrative hook and then deliver factual arguments were associated with the highest purchase willingness. Several senior marketing managers who reviewed the findings during a follow-up validation exercise expressed interest in testing this blended approach in their own campaigns.
There are important caveats worth noting. The meta-analysis captures correlational patterns across hundreds of studies with varying designs, and while the experiment does allow for causal interpretation, it tested only one product in one market. The researchers themselves noted that the mechanisms driving the blended message effect remain unclear. It is not yet established why combining stories and facts in writing is associated with stronger persuasion outcomes, whether it is because the blend feels more credible, more balanced, or something else entirely. The team’s statistical models also captured 38% of the variance in original study results, meaning a substantial portion of what drives persuasion outcomes in any given study remains unexplained by the factors examined.
Still, for practitioners working to cut through the noise of thousands of daily marketing messages, the findings offer a useful starting point. Rather than defaulting to gut instinct about whether to tell a story or present the data, marketers now have a large body of synthesized evidence suggesting that the answer depends on the channel, the product, and, quite possibly, whether they can do both at once.
