INFLUENCER VIDEO ADVERTISING IN TIKTOK

Key words. Advertising marketing

Online advertising: the impact of targeted advertising on advertisers

2 juin 2021 In this research paper we provide a comprehensive overview of ... The role of targeting in digital advertising ... 2020-DPA-Report.pdf.

A STUDY ON THE INFLUENCES OF ADVERTISEMENT ON

In their research paper. "The effects of advertising spending on brand loyalty in services" have discovered that advertising spending has impact on

Ipsos/Google Advertising Attention Research: Video ads and viewer

to TV advertising and YouTube mobile advertising in a natural viewing environment (Attention. In-Home Eye-Tracking Study). Ipsos/Google conducted two

The impact of influencers on advertising and consumer protection in

This document was provided by the Policy Department for. Economic Scientific and Quality of Life Policies at the request of the committee on Internal Market

A STUDY ON ADVERTISING EFFECTIVENESS OF A BRAND

exposure advertising campaign which would have greater consumer research to study the advertising effect of such ... of Classic Articles.

BEHAVIOURAL STUDY ON ADVERTISING AND MARKETING

Title: Behavioural study on Advertising and marketing Practices in Online Social Media PDF/Volume_01 EB-01-18-679-EN-N 978-92-9200-945-8 10.2818/290217.

Research report - Article 10: Expression and advertising of political

Article 10. Expression and advertising of political positions through the media/Internet in the context of elections/referendums. Page 2. Page 3. This report

Guidelines 05/2020 on consent under Regulation 2016/679 Version

4 mai 2020 The term “utmost account” in Article 7(4) suggests that special caution ... paper advertising by the magazine itself and is not to be shared ...

Catégorisation des revues en Économie et en Gestion

revue – Management Science – qui publie 140 articles par an est estampillée par un « g » mais International Journal of Research in Marketing. 0167-8116.

Measuring the Effects of Advertising: The Digital Frontier

1 The ease of data collection in online advertising has led to standard reporting of precise quan- titative data for advertising campaigns most notably the click-through rate (CTR) Of course the CTR of an ad is only an intermediate proxy for the real outcome of interest to the advertiser:1

Measuring the Effects of Advertising: The Digital Frontier

Advertising is a social practice which does not work in vacuum; it interlinks many things together like person and object symbol symbolism and power and communication and satisfaction (Jhally 1987) To be truthful we cannot separate the content of an advertisement from culture

TV Advertising Effectiveness and Profitability: Generalizable

research documenting ine?ective advertising in digital advertising markets (Blake et al 2015 Lewis and Rao) our work should motivate economists to further study the man-agerialandagencyissuesinadvertisingmarkets Therestofthepaperisorganizedasfollows InSection2wesurveytheliteratureon

INFLUENCER VIDEO ADVERTISING IN TIKTOK - MIT Initiative on

The platform’s advertising revenue is estimated at more than $16 billion a year Fig 1: A video’s m-score correlates with additional sales The TikTok platform also has approximately 3 million influencers — video ad creators who simultaneously entertain viewers while selling them products These influencers have quickly emerged as major marketers

THE ROLE OF ADVERTISING IN PROMOTING A PRODUCT - Theseus

The standard definition of advertising includes seven main elements: 1 Paid form of communication 2 The presence of an identified sponsor 3 Distribution through the media 4 The presence of a specific audience for treatment 5 Lack of personalization of distributed information 6 Aimed action

Searches related to research paper on advertising pdf filetype:pdf

The research literature on deceptive advertising spans economics marketing and consumer policy Much of it focuses on factors that alter firm incentives to engage in deceptive advertising (e g Posner 1973; Darby and Karni 1973; Nagler 1993; Kopalle and Lehmann 2006; Zinman and Zitzewitz 2012) and the impact of specific regulatory

What is an analysis of the impact of advertising?

Essentiallyany analysis of the impact of advertising has to make a judgment call on which time periods touse in the analysis. Often this is the campaign window" or the campaign window plus a choseninterval of time (typically 1-4 weeks).

How to improve the quality of advertising in a newspaper?

Another important parameter for advertisement is the volume of the newspaper. Reliable information about number of pages of the newspaper will certainly improve the quality of execution of advertising, as a consequence - increase its effectiveness. Classified and display advertising in newspapers. In accordance with the existing

What is the concept of advertising effectiveness?

The concept of advertising effectiveness contains such different ideas as economic benefits, psychological and social effect expressed in a certain impact on the society in whole (in particular, the influence on the formation of taste preferences of people, their views and ideas about different moral and material values).

What is advertising & how does it work?

Advertisement is the information distributed in any way, in any form or by any means, addressed to an uncertain number of people and aimed at attracting attention to the subject of advertising, the establishment or maintenance of interest in it and its promotion on the market. Advertising reaches customers living far apart.

1 T ik Tok is among the major online platforms blurring the line between content and commerce. Some creators of TikTok videos, known as in?uencers, simultaneously entertain their audiences and sell them products. Our research team studied the di?erence between TikTok in?uencer video ads that drive many sales and ones that drive

fewer sales. We also studied whether it's possible to predict which TikTok inuencer video ads will drive more or fewer sales.

To answer these and other related questions, we developed an algorithm that predicts TikTok in?uence video sales lift using a new metric that we call motion score, or m-score, for short. This statistic is based on an algorithm that quanti?es the extent to which the product is advertised in the most engaging parts of the video. We conclude that a one standard deviation increase in a TikTok

video's m-score translates into an additional $4,000 in monthly sales, an average increase of approximately 12% (Figure 1). We

also ?nd that in?uencer videos are most e?ective for impulsive, hedonic, and inexpensive purchases. Finally, we ?nd that a video's sales lift can be curtailed by incentive mismanagement, which occurs when an in?uencer promotes themselves rather than their products.

THE RISE OF INFLUENCER VIDEO ADS

The impact of TikTok, an online platform for short videos, is huge and fast-growing. In 2020 this online platform was the

IN THIS BRIEF

SODWIRUPKDYHHPHUJHGDVDPDMRUPXOWLELOOLRQ

GROODUIRUFHLQPDUNHWLQJ:HH[SORUHZKDW

7KLVDOJRULWKPXVHVD&RQYROXWLRQDO1HXUDO

1HWZRUNWRTXDQWLI\WKHH[WHQWWRZKLFKWKH

SURGXFWLVDGYHUWLVHGLQWKHPRVWHQJDJLQJ

9LGHRVZLWKKLJKHUPVFRUHVOLIWPRUHVDOHV

HVSHFLDOO\IRUSURGXFWVWKDWDUHERXJKWRQ

LPSXOVHDUHKHGRQLFRULQH[SHQVLYH$YLGHR

WKDWRQO\HQJDJHVRUWKDWVLPSO\IHDWXUHVWKH

SURGXFWWKURXJKRXWZLOOQRWQHFHVVDULO\KDYH

6WDNHKROGHUVLQWKH7LN7RNHFRV\VWHPFDQXVH

PVFRUHVWRDLGYLGHRGHVLJQWRVHOHFWYLGHRV

IRUYDULRXVFDPSDLJQVDQGWRDOLJQLQFHQWLYHV

INFLUENCER VIDEO

ADVERTISING IN TIKTOK

MIT INITIATIVE ON THE DIGITAL ECONOMY 2021,

2 world's most frequently downloaded app, and now TikTok has more than 1.4 billion monthly users worldwide. The platform's advertising revenue is estimated at more than $16 billion a year. Fig. 1: A video's m-score correlates with additional sales The TikTok platform also has approximately 3 million in?uencers - video ad creators who simultaneously entertain viewers while selling them products. These in?uencers have quickly emerged as major marketers. One of TikTok's biggest in?uencers in China, a young man named Li Jiaqi, recently sold more than 7 million units of lipstick in just one day, worth an estimated $145 million. Such huge marketing impact has caught the attention of companies including Walmart, the world's largest retailer, which recently tested product-driven TikTok videos. It still remained unclear, however, why some TikTok in?uencer video ads like Li Jiaqi's drive many sales while others drive few. We set out to develop a way to predict the e?ect of in?uencer videos on product sales by drawing on the theory of bottom- up attention from cognitive psychology and neuroscience (for example, Milosavljevic and Cerf, 2008). Bottom-up attention is a rapid, automatic form of selective attention that depends on the intrinsic properties of the input. It's also known as saliency-based attention, indicating that the more salient, or outstanding, an object, the higher the probability of it being noticed. By contrast, top-down attention is volitional, focal, and task-dependent; it can be likened to a spotlight that enhances the processing of a selected item (Koch

2004).

RELATED MARKET RESEARCH

Our work is inspired by several streams of marketing research. First, we address a problem in in?uencer marketing (Avery and Israeli, 2020). In?uencer marketing - a large industry currently estimated at $10 billion and growing by 50% a year - involves the in?uence of key individuals or opinion leaders to drive brand awareness among consumers and to sway their purchasing decisions (Brown and Hayes, 2008). The main channel through which in?uencers wield their impact is social media. Previous work on in?uencer marketing has studied the e?ect of in?uencer attributes on self-reported purchase intent. For example, Lou and Yuan (2019) found that an in?uencer's trustworthiness, attractiveness, and similarity to their followers a?ect brand awareness and purchase intent. In related work, Schouten et al. (2020) showed that in?uencer endorsements are more e?ective than those from celebrities, and that the e?ect is mediated by a higher perceived similarity and trust. Overall, our area of focus, in?uencer video advertising, has been understudied. One exception is Rajaram and Manchanda (2020), who analyzed the relationship between YouTube in?uencer video ad content and video views, interaction rates, and sentiment. In contrast, we focus on sales conversion and sales lift from in?uencer video ads. The algorithm we developed takes unstructured data from actual TikTok videos and transforms it into structured information in the form of compact, intuitive, and interpretable summary statistics. Not only is this information predictive of sales lift, but it can also be applied before a video is published. We call this summary statistic motion score, or m-score. The motion concept comes from an analogy with Newton's ?rst law of motion: an object will remain at rest or continue to move at a constant velocity unless acted upon by an external force. Similarly, a video ad will not produce sales lift until it has both content engagement and product placement. More speci?cally, variations in engagement over the pixel space and time of a video create force ?elds with varying strength. Drawing further on our analogy, we ?nd that the overall motion (sales lift) is strong if the object (product placement) appears 3 where the force (content engagement) is also strong. Based on theories of bottom-up attention, our hypothesis behind m-score is, other things being equal, that the more salient and engaging an advertised product is in an in?uencer video ad, the more e?ective the video ad will be in lifting sales. To put this into operation, we de?ne a video's m-score as its average pixel-level engagement-weighted advertising intensity.

We compute this in three steps:

We constructed a 3D matrix that we call the

content-engagement heat map. The matrix's three dimensions are the height of each video frame in pixels, the frame's width (also in pixels), and the video's overall length in seconds. This content engagement heat map is a pixel-level saliency map that shows the gradient of video-level engagement (such as number of likes and shares) for each and every pixel. The engagement scores are estimates created by a

3D Convolutional Neural Network trained to work from the

video-level engagement data from some 30,000 video ads.

We constructed a 3D product placement heat

map (Figure 2), which shows whether the product being advertised is present at a given pixel in a given video frame.

Fig. 2: Construction of the m-score algorithm

an image of the product to each frame of the video. To do this, we use an object-detection algorithm called the scale- We estimate the product placement heat map by matching invariant feature transform.

We computed m-score using the two 3D matrices,

normalized by the total number of pixels of the video. This normalized inner product can be interpreted as the average pixel-level, engagement-weighted advertising intensity (Figure 2). In other words, it's the extent to which the product is shown in the most engaging parts of the video ad. We further hypothesize that a video ad with a higher m-score will be more e?ective in lifting sales than one with an m-score that's lower. It is important to note that an m-score measures the complementarity between content engagement and product placement. A video that only engages, or that only features the product throughout, will not necessarily have a high m-score. To earn a high score, a video must feature the product at the most engaging pixels and time (Figure 2, below). We tested our approach by analyzing a proprietary dataset. This set contained some 40,000 in?uencer video ads from the original Chinese version of TikTok, as well as their corresponding product sales revenue on Taobao, a website sometimes referred to as the Amazon of China. We examined sales revenue from 4 March through June of 2019. We focused on the Chinese version of TikTok because of its mature ecosystem around in?uencer video advertising. TikTok in China o?ers an online marketplace, known as Xingtu, which has attracted more than

330,000 in?uencers and 760,000 product sellers.

Consistent with our hypothesis, we found that m-score positively predicts the sales lift of a video ad. Speci?cally, as noted in Figure 1, a one standard deviation increase in m-score is associated with a 12% increase in sales revenue of the advertised product. Notably, neither engagement nor advertising intensity - the latter being the total number of video pixels in which the product appears - alone had an e?ect on sales lift. Rather, it is the complementarity between the two components that drives sales, highlighting the unique predictive power of m-score. To understand the applicability of our algorithm to di?erent product categories, we conducted a supplementary survey to classify advertised products in our data. We found that our algorithm is more e?ective in product categories associated with impulse purchases, hedonic consumption, or lower prices. These kinds of products are also popular choices for the growing new ?eld of entertainment commerce.

CONCLUSIONS

Our algorithm can be put into practice in several ways. First, in?uencers can use it as an automated tool to test and modify their videos for better sales lift. Second, the content engagement heat map can serve as a useful tool to understand video engagement. Third, m-score introduces a new contractual instrument to the entertainment commerce space. Product sellers can use m-score to screen candidate videos or directly write a contract. One key practical advantage of m-score is that it can be computed before a video ad is released, without relying on in- consumption user data such as eye tracking or live comments. This means the algorithm is highly scalable; it can be used to evaluate a large number of candidate videos very quickly. More speci?cally, in?uencers can use m-score to aid video content development in real time, and product sellers can use m-score as a novel contractual instrument. For example, product sellers can compensate in?uencers based on the m-score of their video ads. In comparison, the current industry practice of engagement-based compensation has been shown to be ine?ective, whereas sales-based compensation makes in?uencers accountable for product sales but exposes them to various factors beyond their control (such as perceived product quality, which is di?cult to put into a contract). In this sense, m-score can serve as a metric to help clarify the attribution of sales outcome between product sellers and in?uencers. Finally, entertainment commerce platforms can leverage m-score to launch various features to improve transaction e?ciency. For example, a platform could highlight its m-score as a key performance index of in?uencers. Providing an m-score alongside engagement metrics can also help product sellers choose in?uencers with richer information.

REPORT

THE FULL RESEARCH PAPER CAN BE FOUND HERE

ABOUT THE AUTHORS

is an Assistant Professor of Business Administration in the Marketing unit of Harvard Business School. He recently defended his doctoral dissertation at the MIT Sloan School of Management. His research focuses on optimizing managerial decisions by developing algorithmic products that turn unstructured data into actionable insights. XDQ=KDQJ is the John D. C. Little Professor of Marketing at the MIT Sloan School of Management. An expert in quantitative modeling, she combines economic theory with data science to optimize various business decisions. Her research covers industries including consumer goods, social media and healthcare, and includes product management, pricing, and sales.

KDQ=KDQJis a Lecturer of Marketing at the Beijing

Technology and Business University. She recently received her PhD from the School of Economics and Management at Tsinghua University. Her research focuses on facilitating 5 managerial decision-making in high-tech industry through quantitative analysis and case study.

REFERENCES

Avery, J., Israeli, A. (2020). In?uencer marketing. Harvard

Business School Case, N9-520-075.

Brown, D., Hayes, N. (2008). In?uencer Marketing. Routledge. Koch, C. (2004). The quest for consciousness. Engineering and

Science, 67(2), 28-34.

Lou, C., Yuan, S. (2019). In?uencer marketing: how message value and credibility a?ect consumer trust of branded content on social media. Journal of Interactive Advertising, 19(1), 58-73. Milosavljevic, M., Cerf, M. (2008). First attention then intention: insights from computational neuroscience of vision. International Journal of Advertising, 27(3), 381-398. Rajaram, P., Manchanda, P. (2020). Video in?uencers: unboxing the mystique. SSRN 3752107. Schouten, A. P., Janssen, L., Verspaget, M. (2020). Celebrity vs. in?uencer endorsements in advertising: the role of identi?cation, credibility, and product-endorser ?t. International

Journal of Advertising, 39(2), 258-281.

SUPPORTERS

3M | AB InBev | Accenture | Autodesk | BASF | Bene?tfocus

Tencent

And many generous individuals who prefer to remain anonymous.

MIT Initiative on the Digital Economy

MIT Sloan School of Management

245 First St, Room E94-1521

Cambridge, MA 02142-1347

617-452-3216

Our Mission: The MIT IDE is solely focused on the

digital economy. We conduct groundbreaking research, convene the brightest minds, promote dialogue, expand knowledge and awareness, and implement solutions that provide critical, actionable insight for people, businesses, and government. We are solving the most pressing issues of the second machine age, such as de?ning the future of work in this time of unprecendented disruptive digital transformation.

Contact Us: David Verrill, Executive Director

MIT Initiative on the Digital Economy

MIT Sloan School of Management

245 First St, Room E94-1521

Cambridge, MA 02142-1347

617-452-3216

dverrill@mit.edu

Become a Sponsor: The generous support of

individuals, foundations, and corporations help to fuel cutting-edge research by MIT faculty and graduate students, and enables new faculty hiring, curriculum development, events, and fellowships.

Contact Devin Wardell Cook (devinc@mit.edu) to

learn how you or your organization can support the IDE.quotesdbs_dbs17.pdfusesText_23

[PDF] INFLUENCER VIDEO ADVERTISING IN TIKTOK - MIT Initiative on

What is an analysis of the impact of advertising?

How to improve the quality of advertising in a newspaper?

What is the concept of advertising effectiveness?

What is advertising & how does it work?

THE RISE OF INFLUENCER VIDEO ADS

IN THIS BRIEF

SODWIRUPKDYHHPHUJHGDVDPDMRUPXOWLELOOLRQ

GROODUIRUFHLQPDUNHWLQJ:HH[SORUHZKDW

7KLVDOJRULWKPXVHVD&RQYROXWLRQDO1HXUDO

1HWZRUNWRTXDQWLI\WKHH[WHQWWRZKLFKWKH

SURGXFWLVDGYHUWLVHGLQWKHPRVWHQJDJLQJ

9LGHRVZLWKKLJKHUPVFRUHVOLIWPRUHVDOHV

HVSHFLDOO\IRUSURGXFWVWKDWDUHERXJKWRQ

LPSXOVHDUHKHGRQLFRULQH[SHQVLYH$YLGHR

WKDWRQO\HQJDJHVRUWKDWVLPSO\IHDWXUHVWKH

SURGXFWWKURXJKRXWZLOOQRWQHFHVVDULO\KDYH

6WDNHKROGHUVLQWKH7LN7RNHFRV\VWHPFDQXVH

PVFRUHVWRDLGYLGHRGHVLJQWRVHOHFWYLGHRV

IRUYDULRXVFDPSDLJQVDQGWRDOLJQLQFHQWLYHV

INFLUENCER VIDEO

ADVERTISING IN TIKTOK

MIT INITIATIVE ON THE DIGITAL ECONOMY 2021,

2004).

RELATED MARKET RESEARCH

We compute this in three steps:

We constructed a 3D matrix that we call the

3D Convolutional Neural Network trained to work from the

We constructed a 3D product placement heat

Fig. 2: Construction of the m-score algorithm

We computed m-score using the two 3D matrices,

330,000 in?uencers and 760,000 product sellers.

CONCLUSIONS

REPORT

THE FULL RESEARCH PAPER CAN BE FOUND HERE

ABOUT THE AUTHORS

KDQ=KDQJis a Lecturer of Marketing at the Beijing

REFERENCES

Business School Case, N9-520-075.

Science, 67(2), 28-34.

Journal of Advertising, 39(2), 258-281.

SUPPORTERS

3M | AB InBev | Accenture | Autodesk | BASF | Bene?tfocus

Tencent

MIT Initiative on the Digital Economy

MIT Sloan School of Management

245 First St, Room E94-1521

Cambridge, MA 02142-1347

617-452-3216

Our Mission: The MIT IDE is solely focused on the

Contact Us: David Verrill, Executive Director

MIT Initiative on the Digital Economy

MIT Sloan School of Management

245 First St, Room E94-1521

Cambridge, MA 02142-1347

617-452-3216

Become a Sponsor: The generous support of

Contact Devin Wardell Cook (devinc@mit.edu) to