Marketing

What makes for successful social media marketing? Is it messaging that creates excitement or leans towards the informative? Stunning visuals and/or stirring music? Celebrity influencers and their stamp of approval? And how does it all affect sales anyway?
Identifying the elements of effective social media messaging is critical, considering that 3.6 billion people use social media, and social media ad spend is on track to surpass US$200 billion by 2024. In a groundbreaking study, my co-authors* and I investigate the impact of different types of firm-generated content (FGC) on the product sales of a large consumer electronics multinational over two years. We found that hedonic messaging, which seeks to evoke fun, playfulness, enjoyment and excitement, increased sales more than utilitarian ones, which emphasise products’ functionality. Importantly, this effect came through only when the message was simple and of sufficient quality and consistency.
Our study, the first to study hedonic and utilitarian FGC approaches in the context of a large firm, underscores the importance of coherent and integrated marketing communication based on the principles of quality, consistency and simplicity.
The study
The subject of our study is a major global consumer electronics and manufacturing firm, which agreed to be involved on condition of anonymity. It provided us with its daily sales revenue, measured in US dollars, for each product in our analysis of its 2016–2017 marketing and sales programme database. During the two-year period, the firm engaged with customers on its Facebook page and Twitter account.
We retrieved all firm-generated content on both social media platforms as well as consumer comments, posts, shares, and “likes”. Then, we manually analysed if the posts were hedonic or utilitarian as well as their quality, consistency and complexity. We also analysed consumers’ reactions to the FGC. Below are our criteria for each metric:
Hedonic or utilitarian
Utilitarian messages emphasised the product’s practical benefits, such as completing specific tasks efficiently and effectively and informing users how to resolve specific problems with the product. Hedonic ones highlighted the product as bringing fun, enjoyment and excitement. They were experiential and multisensory as well as imaginative and aspirational.
Message quality
Message quality is measured by the average score of three independent experts’ rating of each message, from 0 to 10, in terms of uniqueness, novelty and believability.
Message goal consistency
Message goal consistency is measured by the average rating of three evaluators who independently assessed each message on whether it was compatible, consistent and congruous with the firm’s brand image of “cutting-edge, future-oriented, tech-savvy innovation, world-class and experiential”.
Message complexity
Message complexity was measured by readability of the marketing content. For each product, we used the mean readability score (calculated with the Dale–Chall formula) of the firm’s Facebook and Twitter posts to gauge the messages’ complexity.
Keep it simple, keep it consistent
Our analysis shows that hedonic messaging boosted product performance – measured in terms of daily sales revenue – more than utilitarian messaging. But what’s even more noteworthy is that message quality and message goal consistency played a decisive role in determining the marketing outcome.
Whether the messaging was hedonic or utilitarian, the posts had to reach a certain level of quality and consistency before they began to have a positive impact on sales. The minimum quality score for hedonic messaging was 4 (out of 10) and 6 for utilitarian ones. For both types of messaging, the minimum consistency score was 0.
Further, message quality affected the impact of hedonic messaging more than that of utilitarian ones.
Put simply, high-quality messaging boosted the firm’s sales more when the messaging was hedonic rather than utilitarian, while poor-quality messaging damaged sales more when the messaging was hedonic. Message consistency had similar moderating effect on the two types of marketing appeals.
How to make great content
While there’s a growing body of research on FGC, our study is the first empirical test of the effectiveness of quality, consistent and simple online marketing communication on product sales.
So how might firms achieve the holy grail of unique, consistent and easy-to-comprehend messaging? They could start by asking the following questions:
1. Is the content compelling?
FGC should provide value by helping users solve their problems, answer their questions and concerns, pique and expand their interests, or build trust in the product. The most successful FGC on social media also engages with users, including responding to customer comments.
2. Does the content flow well aesthetically and logically?
Users won’t click on content that is not entertaining, intriguing or informative, so use different types of media including photos, videos, music and infographics. But be careful not to neglect the basics – spelling and grammatical mistakes can damage credibility, no matter how good the messaging is otherwise.
3. Is the content aligned with the firm’s overall branding?
Consistent messaging is aligned with what the brand stands for, its value proposition and purpose. It also adheres to the brand DNA in terms of style and tone. Consistent FGC makes the brand more recognisable and credible to consumers.
Of course, all that hinge on whether firms have clear brand positioning and a set of branding guidelines for all employees as well as external partners. This will help ensure that the look, tone, terminology and framing of FGC across all products, themes, markets and platforms are consistent.
Take it from Apple. The “world’s most valuable brand” has few rivals in crafting meticulously curated messages that reinforce its central identity of creating sleek and elegant products that push the boundaries of innovation. Another role model is Walmart, which never fails to emphasise in its FGC its low prices and value creation for customers.
Conversely, FGC that diverges from the brand positioning sends mixed signals to consumers. Many automobile brands fall into this trap: While the brand’s own content usually conveys a slick, professional image, the local dealer’s messaging is often unprofessional and poorly executed.
4. Is the content concise and well-articulated?
Strong, memorable FGC is characterised by crisp, persuasive and clear writing. It evokes positive emotions and helps customers visualise how they might benefit from using the firm’s product or service. For this reason, it is a good idea to test content for readability before posting it on social media.
Remember, the more complex the content is, the more likely it will be ignored.
 
*Jifeng Mu, Alabama A&M University; Jonathan Zhang, Colorado State University College of Business; and Jiayin Qi, Shanghai University of International Business & Economics
is an INSEAD Assistant Professor of Marketing and a Visiting Assistant Professor at the Wharton School of Business. His research expertise is in generating customer and firm insights from unstructured data using computational linguistics and machine learning, and in examining returns to marketing strategy. 
“Creative Appeals in Firm-Generated Content and Product Performance” is published in Information Systems Research.
Anonymous User
21/07/2022, 06.08 pm

Dear Abhishek
Thanks a lot for going through this analysis with your colleagues and sharing these insights (especially for the manual analysis of the posts; having done something similar in the past I know it is very tedious to say the least…).
It would be interesting to know how you dealt with the relationship between posts and daily sales. To begin with, did your analysis include both online and offline sales? Also, what assumptions on sales attribution did you make? E.g. straight correlation between posts on a given day with sales on that day? For online sales, did you make the standard google hypothesis on attribution based on latest point of contact? Would you agree that assumptions on this have an impact on the results? E.g. a customer can buy on day N mainly influenced by a single post seen many days/weeks earlier, or a customer can buy based on a set of different messages seen both online and offline that have each contributed to develop his/her awareness, interest, intention over time until final purchase.
Best regards
Matteo

Dear Abhishek
Thanks a lot for going through this analysis with your colleagues and sharing these insights (especially for the manual analysis of the posts; having done something similar in the past I know it is very tedious to say the least…).
It would be interesting to know how you dealt with the relationship between posts and daily sales. To begin with, did your analysis include both online and offline sales? Also, what assumptions on sales attribution did you make? E.g. straight correlation between posts on a given day with sales on that day? For online sales, did you make the standard google hypothesis on attribution based on latest point of contact? Would you agree that assumptions on this have an impact on the results? E.g. a customer can buy on day N mainly influenced by a single post seen many days/weeks earlier, or a customer can buy based on a set of different messages seen both online and offline that have each contributed to develop his/her awareness, interest, intention over time until final purchase.
Best regards
Matteo
Anonymous User
03/08/2022, 02.07 pm

Great questions Matteo.
First answer is that this was only online sales.
As for the assumptions regarding the effect, we used the prior day online posts and then looked at its effect on the next day online sales. We tried different lags such as t-1, t-2, t-3; the results were similar but the best fit was the t-1 lag. We also used Vector Autoregressive models (VARX) and the results were similar.
Yes, we just used the last point of contact but note that we controlled for about 15 other variables like advertising, media, quality, sentiment of the posts, price, etc.
Yes, we cannot definitely say what is the exact process of how the consumer buys. It could be either seeing the post days earlier or seeing a medley of messages and then buying. What we do find empirically after accounting for endogeneity, reverse causality, a horde of control variables, is that the t-1 lag works best.
Hope this answers your questions. Thanks much for your interest and these superb questions.
Abhi

Great questions Matteo.
First answer is that this was only online sales.
As for the assumptions regarding the effect, we used the prior day online posts and then looked at its effect on the next day online sales. We tried different lags such as t-1, t-2, t-3; the results were similar but the best fit was the t-1 lag. We also used Vector Autoregressive models (VARX) and the results were similar.
Yes, we just used the last point of contact but note that we controlled for about 15 other variables like advertising, media, quality, sentiment of the posts, price, etc.
Yes, we cannot definitely say what is the exact process of how the consumer buys. It could be either seeing the post days earlier or seeing a medley of messages and then buying. What we do find empirically after accounting for endogeneity, reverse causality, a horde of control variables, is that the t-1 lag works best.
Hope this answers your questions. Thanks much for your interest and these superb questions.
Abhi
Operations
H Cho, M Sosa, S Hasija
Marketing
Abhishek Borah
Marketing

This website is part of INSEAD Digital Ecosystem
© 2022 INSEAD
INSEAD does not discriminate on the basis of race, gender, religion, age, sexual orientation, disability, colour, or national or ethnic origin.

source

Leave a Reply

Your email address will not be published.