What We Know About Word of Mouth

Numbers, The Downside of Digital Word of Mouth Please And the Pursuit of Media Quality How Social Sharing Is Disrupting Digital Models and Metrics

Gian M. Fulgoni INTRODUCTION from person to person. Word-of-mouth communi- comScore, Inc. When the mobile Internet was still in its early cation plays on the value of trust to persuade how [email protected] growth phase, around 2012, social-networking others think, feel, and act. We know this from his- usage shifted quickly in its direction. What fol- tory, long before “social media” entered the pub- Andrew Lipsman lowed was an unprecedented boom in viral content lic lexicon. Remember the 1982 Fabergé Organics comScore, Inc. that shook the digital world. Shampoo spot that made Heather Locklear the face [email protected] Social media became so powerful a medium for of peer-to-peer advertising? “She liked her sham- spreading content and ideas that it was inevitable poo so much, she told two friends, who told two there eventually would be attempts to exploit it. friends, who told two more friends, ‘and so on and Within a very short time, the environment became so on and so on’” (Precourt, 2014, p. 124). riddled with various forms of “digital pollu- Marketers in the early 2000s became further tion”—from spam to fraud and “fake news”— the enamored with the potential of word of mouth prevalence of which has been accelerated and exac- when popular books—like Malcolm Gladwell’s erbated by the rise of programmatic advertising. (2000) The Tipping Point and Ed Keller and Jon The digital-media environment has evolved so Berry’s (2003) The Influentials—entered the pub- quickly that the metrics infrastructure has been chal- lic discourse. By the middle of the first decade, as lenged to keep pace. Digital word of mouth, in the the Internet began to mature as a communication form of the rapid sharing of content via social and medium with the emergence of social networks, the mobile channels, both has democratized content and need to reconsider the original model of how word has created a system of incentives that has commod- of mouth works became clear. With the average itized the ecosystem. The short-term chase for bigger person no longer limited to influencing only those audience metrics and greater impression volume has in his or her immediate circle of friends and family led to significant challenges for media economics. through person-to-person interactions, the Internet Metrics have been a part of the problem but also suddenly was enabling influence at scale. People promise to be part of the solution. The fundamental could make (or solicit) recommendations instantly metrics of media planning and campaign measure- to the hundreds of people in their networks. ment—impressions, reach, frequency, and demo- No longer did the truly powerful ideas or prod- graphics—need not go away. The difference now is ucts require a long and sustained grassroots effort that these metrics need a higher level of validation to gain critical mass and enter the public conscious- to ensure that the inventory being bought is clean, ness. The density of communication and velocity legitimate, and appearing in environments condu- with which ideas could spread on social networks cive to effective advertising. meant that if something resonated with the public, it could gain mass exposure very quickly. THE RISE OF DIGITAL WORD OF MOUTH Social sharing was thought of as a panacea for There is no more powerful way of communicating both users and marketers, with the promise that the information, including marketing messages, than most powerful ideas and marketing messages had

DOI: 10.2501/JAR-2017-020 June 2017 JOURNAL OF ADVERTISING RESEARCH 127 What We Know About Word of Mouth

a path to maximum exposure. The Internet This growth was generated exclusively and world leaders have conversations on had an even greater potential to democra- by mobile audiences, which increased ,” Williams wrote in one of his blog tize ideas and break through traditional 127 percent during that time period (com- posts. “If that’s happening, I frankly don’t distribution barriers. This meant brands Score, 2016a; See Figure 1). [care] if Instagram has more people look- had to get comfortable ceding some con- Publishers gained the added benefit ing at pretty pictures.”4 trol over their message, but also that they of audience scale, a positive develop- At Medium, Williams advocated use of could find new and creative ways to reach ment that typically would translate into a “total time reading” metric to capture and engage their customers. improved business prospects. Larger scale the overall engagement of his audience On a commercial level, Dollar Shave means an improved ability to reach tar- base rather than relying solely on metrics Club broke through with a highly effec- get audiences and be included in adver- of audience scale. But in January 2017, tive viral video advertisement that helped tisers’ media plans. Yet publishers often Williams announced that Medium was an unknown brand reach millions of men had difficulty effectively monetizing retooling its business in search of a new tired of paying too much for razors.1 At a their mobile inventory and could not take monetization model not dependent on more societal level, there was the poten- full advantage of this increased scale. advertisements.5 It seems that audience tial for spreading positive messages for Their predicament suggested that the scale remains essential for advertising- important causes, such as the Ice Bucket sometimes-singular pursuit of audience driven business models, but Williams also Challenge for building awareness of scale was shortsighted. makes a good argument that it may not be amyotrophic lateral sclerosis (ALS), other- Although top-line metrics of scale more sufficient. wise known as Lou Gehrig’s disease. (The easily could be achieved, metrics dem- Ice Bucket Challenge eventually led to the onstrating the depth of engagement and “Fake News” and Misaligned Incentives discovery of a gene tied to the disease.2) quality of the media environment often The problems of identifying the right met- Social media also helped engage more were overlooked. Dmitry Shishkin, digital rics further were compounded by corrup- citizens in the political process, by encour- development editor of BBC World Service tive practices in digital media. Indeed, the aging voting and providing improved Group, in 2015 said that with mobile “you viral content boom created a powerful way means of participation and grassroots might get random traffic spikes, but you for those seeking to take advantage of the organizing. won’t get engagement. We must figure out platform in potentially damaging ways. how to monetize minutes of engagement, Suddenly, the promise of digital word of Grappling with Metrics: not just eyeballs.”3 mouth at scale also was attracting various The Primacy of Audience Scale Another prominent publishing execu- forms of digital pollution that made the By the end of this century’s first decade— tive elaborated on this point. Evan Wil- environment less attractive and harder for as digital media time-shifted to mobile— liams, who cofounded Twitter and later marketers to navigate. users were spending more time in front of launched the blogging platform Medium, Case in point: Donald J. Trump’s sur- their screens than ever before. As social- questioned the industry’s overreliance prising march to victory in the 2016 U.S. media platforms referred significant traffic on “monthly active users” in response to presidential election. In the wake of this to a variety of publishers, digital audiences a headline about Instagram surpassing shocking result, which belied the many consistently climbed. The top 1,000 digital Twitter on that metric. Reducing every polls predicting a comfortable victory for media properties’ average U.S. monthly digital-publisher conversation to a single Hillary Clinton, many placed blame at the audience rose to 16.8 million in December metric of audience scale fails to acknowl- feet of various media channels (See “What 2016 from 12.3 million in December 2013, edge the many other factors that make an Survey Researchers Can Learn from the a gain of 37 percent in just three years. audience unique. “[Twitter is a] realtime 2016 U.S. Pre-Election Polls,” page 182). information network where everything

1 J. P. Pullen, “How a Dollar Shave Club’s Ad Went Viral,” in the world that happens occurs on Twit- 4 E. Williams, “A Mile Wide, an Inch Deep,” Janu- October 13, 20102. Retrieved March 24, 2017, from Entre- ter — important stuff breaks on Twitter ary 5, 2015. Retrieved March 24, 2017, from Medium: preneur.com: https://www.entrepreneur.com/article/224282. https://medium.com/@ev/a-mile-wide-an-inch-deep 2 K. Rogers, “The ‘Ice Bucket Challenge’ Helped Scientists -48f36e48d4cb#.esbb9tb5a. Discover a New Gene Tied to A.L.S,” , 3 J. Davies, “Publishers Share Their Biggest Headaches in 5 M. Gajanan, “Medium Announced Major Layoffs,” January July 27, 2016. Retrieved March 24, 2017, from https://www Monetizing Mobile,” October 6, 2015. Retrieved March 24, 4, 2017. Retrieved March 24, 2017, from Fortune.com: http:// .nytimes.com/2016/07/28/health/the-ice-bucket-challenge 2017, from Digiday.com: http://digiday.com/uk/publishers fortune.com/2017/01/04/medium-layoff-announcment -helped-scientists-discover-a-new-gene-tied-to-als.html. -monetizing-mobile-headaches/. -ev-williams/.

128 JOURNAL OF ADVERTISING RESEARCH June 2017 The Downside of Digital Word of Mouth And the Pursuit of Media Quality thearf.org

inventory and media environment tends to Total Digital Desktop Mobile 20 take a back seat. 18 In this environment, the incentive 16.8 16 changes for content producers. Instead 14 of investing in building a brand through 12.7 12 12.3 high-quality content that attracts loyal m (MM) 10 readership, cobbling together cheap web- 8.1 8 sites promoting sensationalist headlines 6 6.3 5.6 that easily can generate clicks becomes by Platfor 4 an easier path to generating advertising- verage of Unique Visitors

A 2 0 impression volume. In many cases, that takes the form of clickbait designed to find its way into

Dec-2013Feb-2014Apr-2014Jun-2014Aug-2014Oct-2014Dec-2014Feb-2015Apr-2015Jun-2015Aug-2015Oct-2015Dec-2015Feb-2016Apr-2016Jun-2016Aug-2016Oct-2016Dec-2016 people’s social-media feeds and gener-

Source: comScore ate quick eyeballs. In more extreme cases, this clickbait comes in the form of fake news. In yet more extreme cases, there are Figure 1 top 1,000 Properties: Average Monthly Audience (MM) websites built purely for the purpose of generating fake impressions through bot There was criticism of which stories were was an easy way to generate clicks, which traffic, siphoning off advertising dollars and were not covered by certain media generated advertising impressions, which along the way. outlets, the attention given to particular generated revenue. Increasingly wary of In December 2016, digital-advertising stories, and, most notable, the role that fake news, brands have taken a stronger security company WhiteOps publicized “fake news” might have played in influ- stand about being placed alongside this the existence of “Methbot,” a bot network encing the electorate. and other objectionable content, with with roots in Russia that was generating A BuzzFeed analysis suggested that in greater attention being placed on blacklist- fake impressions on a massive scale.9 the lead-up to the general election, the total ing sites known to be peddling these types comScore research corroborated that, number of engagements on social media of content.8 although this particular bot network was for fake news stories surpassed the num- a large one, it was not especially unique in ber of mainstream news stories during the Programmatic Pressure its magnitude—representing just 0.86 per- period of August through Election Day.6 And the Problem with Bots cent of global invalid traffic observed by Some purveyors of fake news—such as a Although many pollutants of the digital- comScore on December 19, the day before network of teenagers in Macedonia—were media ecosystem were inevitable, one could the threat was made public by WhiteOps. motivated primarily by economics, real- argue that their prevalence was exacerbated Threats of this nature will continue to izing that exploiting social-media users’ by the rise of programmatic advertising. persist as long as they are able to exploit willingness to believe and share fake news Automated trading platforms, through the a digital-advertising ecosystem that lacks stories that reinforced their ideological promise of improved efficiency in reaching transparency—allowing for rampant infla- bent was a lucrative business.7 Fake news the right audiences, have become important tion of fake advertising impressions that

6 C. Silverman, “This Analysis Shows How Viral Fake Elec- new gatekeepers in how digital media are go undetected—because verification ser- tion News Stories Outperformed Real News on Facebook,” bought and sold. When the impressions on vices were not used. November 16, 2016. Retrieved March 24, 2017, from Buzz- advertising exchanges get packaged and Feed News: https://www.buzzfeed.com/craigsilverman/ viral-fake-election-news-outperformed-real-news-on sold like pork bellies, the quality of the -facebook?utm_term=.ho65OVOxG#.rmEE0x0nl. 7 A. Smith and V. Banic, “Fake News: How a Partying Mac- 9 J. Marshall, “Russian Hackers Stole Millions from edonian Teen Earns Thousands Publishing Lies,” Decem- 8 J. Nicas, “Fake News Sites Inadvertently Funded by Video Advertisers, Ad Fraud Company Says,” The Wall ber 9, 2016. Retrieved March 24, 2017, from NBC News: Big Brands,” The Wall Street Journal, December 8, Street Journal, December 21, 2016. Retrieved March 24, http://www.nbcnews.com/news/world/fake-news-how 2016. Retrieved March 24, 2017, from http://www.wsj. 2017, from http://www.wsj.com/articles/russian-hackers -partying-macedonian-teen-earns-thousands-publishing com/articles/fake-news-sites-inadvertently-funded-by-big -stole-millions-from-video-advertisers-ad-fraud-company -lies-n692451. -brands-1481193004. -says-1482272717.

June 2017 JOURNAL OF ADVERTISING RESEARCH 129 What We Know About Word of Mouth

dissemination of counterfeit goods. As ImprImpressions-Wessions-Weightedeighted Brand Brand Lift LiftResults Results Based Based on 15 on Lar15ge Lar U.S.ge U.S. Display Display of May 2016, more than 30 leading andand Video Video Campaigns Campaigns 2015–2016 2015–2016 digital-advertising companies and agen- cies had agreed to participate in TAG’s antifraud initiative.16

0.890.89 In July 2016, comScore conducted research 0.530.53 highlighting the value of quality adver- tising environments. The advertising effectiveness study examined the value DCNDCN Publishers Publishers Non-DCNNon-DCN Publishers Publishers of premium media environments, which Note:Note: DCN DCN publishers publishers are membersare members of the of Digital the Digital Context Context Next Nexttrade trade group. group. were defined as member companies of SourSource: comScorce: comScore e the Digital Content Next trade group. It showed that premium publishers deliv- Figure 2 average Brand Lift of Premium versus Nonpremium ered 67 percent higher branding lift than Publishers nonpremium publishers (comScore, 2016b; See Figure 2). Some of this difference could be A RETURN TO MEDIA QUALITY Kellogg’s, and Nestlé, had stated publicly explained by higher viewability rates Digital media stakeholders increasingly they would be pulling their advertise- (50 percent versus 45 percent) and lower are standing up for quality: ments from the alt-right-leaning Breit- incidence of invalid traffic (2.2 percent bart News site and other sites they deem versus 3.5 percent) on those publisher • Facebook and Google both have said inconsistent with their brand values.13,14 sites. The halo effect of the premium, they would take action to remove fake • Media investment firm GroupM in 2015 well-lit, contextually relevant environ- news from their newsfeeds and search demanded that its publishing part- ment, however, was the more significant results.10 ners get TAG certified in order to get cause of improved advertising effective- • Publishers such as Slate and The New included on their media buys.15 TAG is ness.17 Research that helps isolate the Yorker stopped hosting “around the the Trustworthy Accountability Group, value of the environment—beyond its web” advertisements on their websites.11 which in February that year announced ability to deliver an impression against • The Economist has begun selling adver- its antipiracy program. Its goal is to help a certain type of audience—can better tisements on the basis of attention met- prevent the placement of advertise- inform media buyers about the inven- rics, such as time in view.12 ments on websites that facilitate distri- tory that most likely will work for them. • As of late March 2017, more than 1,500 bution of pirated content or the illegal What is more, the use of supplementary advertisers, including 3M, Allstate, Avis, metrics of engagement—whether time in Bed Bath and Beyond, Charles Schwab, 13 T. M. Andrews, “Kellogg’s, Citing ‘Values,’ Joins Grow- view, interaction rates, completion rates, ing List of Companies That Pledged to Stop Advertis- or a variety of other metrics—also gives ing in ,” The Washington Post, October 10 N. Wingfield, M. Isaac, and K. Benner, “Google and Face- 30, 2016. Retrieved March 24, 2017, from https://www. the media buyer better information to use book Take Aim at Fake News Sites,” The New York Times, washingtonpost.com/news/morning-mix/wp/2016/11/30/ to allocate media spend. November 14, 2016. Retrieved March 24, 2017, from https:// kellogg-citing-values-joins-growing-list-of-companies www.nytimes.com/2016/11/15/technology/google-will-ban -that-pledged-to-stop-advertising-in-breitbart-news/ -websites-that-host-fake-news-from-using-its-ad-service ?utm_term=.3836269e1e4f. 16 “Group M Proudly Supports TAG’s Launch Anti-Fraud .html. 14 Sleeping Giants, “Sleeping Giants Confirmed List Certification Program,” Group M press release, May 23, 11 S. Maheshwari and J. Herrman, “Publishers Are Rethink- Updated 3.17.17,” Retrieved March 24, 2017, from Sleeping 2016. Retrieved March 26, 2017, from https://www.groupm. ing Those ‘Around the Web’ Ads,” The New York Times, Giants Twitter page: https://docs.google.com/spreadsheets/ com/news/groupm-proudly-supports-tags-launch-anti October 30, 2016. Retrieved March 24, 2017, from https:// d/1i9o8CR_kjJ6mBd44k6CRZEhlXuZqq-XCCOoj-e8RJ7Q/ -fraud-certification-program. www.nytimes.com/2016/10/31/business/media/publishers edit#gid=0. 17 M. Swant, “Ads on Premium Publisher Sites Are 3 Times -rethink-outbrain-taboola-ads.html. 15 T. Peterson, “Group M Wages War on Piracy Sites More Effective at Boosting Brand Favorability,” Adweek, 12 J. Davies, “How The Economist Plans to Grow Atten- with Anti-Fraud Group TAG, Demands Publishers Join,” July 14, 2016. Retrieved March 24, 2017, from http://www. tion-Based Ad Sales,” February 4, 2016. Retrieved March September 23, 2015. Retrieved March 24, 2017, from adweek.com/news/technology/ads-premium-publishers 24, 2017, from Didigday.com: https://digiday.com/uk/ AdvertisingAge: http://adage.com/article/digital/groupm -sites-are-3-times-more-effective-boosting-brand-favorability economist-plans-scale-time-based-ad-sales/. -wages-war-piracy-sites-anti-fraud-group-tag/300510/. -172509.

130 JOURNAL OF ADVERTISING RESEARCH June 2017 The Downside of Digital Word of Mouth And the Pursuit of Media Quality thearf.org

With the promise of improved metrics, marketers many innovative technological methods of measuring consumer behavior and advertising effectiveness. He can be assured that a cleaner, more transparent is a regular contributor to the Journal of Advertising Research. environment and improved economics are within reach. Andrew Lipsman is vice president of marketing and insights at comScore, Inc., covering multiple industries and overseeing the company’s marketing communications, insights, and thought-leadership CONCLUSION • Is it reaching an engaged and relevant initiatives. He began his career at The NPD Group, With the promise of improved metrics, audience? where he worked with clients such as Kraft Foods and marketers can be assured that a cleaner, Johnson & Johnson. Lipsman specializes in social more transparent environment and Metrics addressing media quality and media, e-commerce, online video, digital advertising, improved economics are within reach. audience attention are available and and multiplatform marketing and has contributed to Traditional metrics of media planning and should be used to supplement standard the Journal of Advertising Research. campaign measurement—impressions, media-planning variables. reach, frequency, and demographic— Finally, achieving a more hospitable remain the pillars of the media buying and advertising environment for both buyer REFERENCES selling ecosystem. A higher level of valida- and seller demands that each player is tion of these metrics should ensure that willing to play a role in defending against comScore. (2016a, September 13). “The the inventory being bought is pollutant challenges: 2016 U.S. Mobile App Report” (white free, legitimate, and appearing in effective paper). Retrieved March 24, 2017, from advertising spaces. • Advertisers and agencies must avoid https://www.comscore.com/Insights/ The layers of validation needed include shopping only for bargain-basement Presentations-and-Whitepapers/2016/ CPMs (i.e., cost per thousand), because The-2016-US-Mobile-App-Report. • the measurement of viewable these impressions often come from the

impressions; worst-offending sites as far as traffick- comScore. (2016b, June 14). “The Halo Effect: • sophisticated detection and removal of ing in fake news, low-quality content, or How Advertising on Premium Publishers invalid traffic; wholesale bot fraud. Drives Higher Ad Effectiveness” (white paper). • brand-safety protections (using both • Publishers must resist the urge to chase Retrieved March 24, 2017, from http://www. black lists and white lists). audience scale through clickbait or third- comscore.com/Insights/Presentations-and party traffic buying that perpetuates the -Whitepapers/2016/The-Halo-Effect-How These protocols ensure that advertise- deterioration of digital advertisement -Advertising-on-Premium-Publishers-Drives ments can reach actual human eyeballs buying. -Higher-Ad-Effectiveness?cs_edgescape_cc=US within an environment that does not tar- • Programmatic exchanges can build in nish the brand. more safeguards and filters to surface Gladwell, M.  The Tipping Point: How Little Measurement also can go a step fur- quality inventory that produces not only Things Can Make a Big Difference. New York: ther to improve what is known about reach and impressions but engagement Little, Brown, 2000. media quality as it pertains to advertising, and attention. however: Keller, E., and J. Berry. The Influentials: One in Ten Americans Tells the Other Nine How to Vote, About the Authors • Is the advertisement in view and reach- Where to Eat, and What to Buy. New York: Simon

ing humans? Gian M. Fulgoni is cofounder and chief executive officer and Schuster, 2003. • Is it in a contextually relevant of comScore, Inc. Previously, he was president and chief

environment? executive officer of Information Resources, Inc. During a Precourt, G. “What Do We Know About Peer- • Does it appear adjacent to a brand that career of more than 40 years at the c-level of corporate to-Peer Marketing?” Journal of Advertising delivers a halo effect? management, he has overseen the development of Research 54, 2 (2014): 124–125.

June 2017 JOURNAL OF ADVERTISING RESEARCH 131