Media Mix Modeling 2.

0
The Fine Art of Blending and Mixing Paid & Earned Media to Build Brands

ARF RE-THINK 2010 DRAFT VERSION

Pete Blackshaw, Executive Vice President Nielsen, Digital Strategic Services pete.blackshaw@nielsen.com

Blackshaw Media Mix Modeling 2.0 1

Summary
Improvements in online measurement and the growing adoption of consumer listening platforms are laying foundation for a new framework for maximizing brand value through mixing and weighting Paid and Earned media online that we might refer to as Media Mix Modeling 2.0. Paid and earned media complement and reinforce one another, especially with correct levels of brand readiness, agility, advocacy and latency working in the background. This paper will provide marketing and research stakeholders with a framework and decision-guide for thinking about paid, earned, and blended media inputs using select examples of cross-platform integration, from Super Bowl advertising to Twitters echo effect on primetime television shows and advertising. We will discuss how this should inform key operational choices such as resource, media spend, and indirect marketing (e.g. CRM, customer service, social media engagement) decisions. Nielsen learned through in-depth analysis of both paid and earned media inputs during the 2010 Super Bowl that the interplay of the two made a significant difference for participating brands. Coordinated activity increased brands overall level of conversation and primed the ad buy for an ongoing annuity of free media through search results and site indexing, among other things. We also learned that it mattered for brands to be prepared and primed we use the terms brand readiness and brand agility before a paid media investment. This paper will detail those learnings as well as high-level learning from the 2010 Olympics and Academy Awards.

Background: Marketing Dilemma in a Digital Blended Age

In recent months there has been a growing level of industry attention on Paid Media versus Earned Media. One common refrain is that brands should consider shifting more attention and resources from the paid to the earned side of the equation, ideally resulting in word-of-mouth conversation favorable to the brand. This might involve, for example, greater investment in brand experience or customer service as this has proven to trigger favorable conversation as displayed in high visibility search results against brand searches. Marketers lead toward paid media because it is predictable, baked into existing media processes, and increasingly targeted and precise and increasingly so in a digitally enabled world -- while earned media, most embodied in social media and online conversation, is far harder to guarantee and typically works in longer-term cycles. But is it that simple? Paid media often triggers online conversation via the webs echochamber, creating a form of blended media that can display either positively or negatively for the brand. Conversely, earned media can inform paid media opportunities, such as the growing spectrum of co-creation activities. Frito Lays increasing reliance on viewer crowd-sourcing for Super Bowl ads richly illustrates this

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 2

phenomenon. Earned media can also inform the shape and direction of paid media and paid media in turn should echo this learning. How to best optimize the media mix in a paid/earned digital universe especially against new dynamics of consumer control and leverage -- and resource accordingly remains a huge question for marketers. If we know that customer service is the top driver of high-trust/high-impact earned media, should a percentage of paid media be used to pay for it, or should you utilize a more blended approach? (e.g. Best Buy anchoring TV advertising to Twelpforce, as service messaging makes TV copy more persuasive and impactful.) Moreover, how do we think about the mix during different stages of the product life cycle? Earned media tends to incubate, grow, and spread around new news. This makes sense as the as new provides currency for social connectors, digital or otherwise, to tell others and to time-stamp their discovery.

Research Solution: Media Mix Modeling 2.0?

Whats emerging is something we might refer to as Media Mix Modeling 2.0. This research makes a first attempt at framing core assumptions around earned, paid, and blended media inputs. More specifically, we attempt to probe and answer the following questions: o Definitions: What are the definitional parameters of Paid, Earned and Blended media? How do we drive distinctions with the PR industrys historic use of the term earned media? What is the interplay between paid and earned media? Core Measurements: What are the most critical measurements in this type of environment? What role does the listening platform play as a price of entry for the new media mix modeling? 2010 Super Bowl Learning: How did advertisers overall increase net return on their $2.5 MM Super Bowl spot leveraging earned media? In what ways did Frito Lay exploit Paid/Earned media framework in the 2010 Super Bowl via the latest version of the Crash the Super Bowl campaign? Decision Making & Organizational: What are the key marketing considerations for senior officers in a paid/earned word? What decisions are within scope, or out of scope? What are the critical organizational considerations one must grapple with?

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 3

Defining Terms: Paid versus Earned Media

In a Media Mix Modeling 2.0 world, marketer and researchers should be indifferent to how they receive media impressions. Marketers are always in relentless pursuit of maximizing favorable impressions for the brand. These impressions later influence awareness, trial, and ultimately purchase behavior. Today such impressions can source from either what brands pay for (e.g. TV views, online eyeballs) or what they earn (online conversation, PR). Both sides have grown more complex with the advent of digital media and with the proliferation of consumer expression venues. To put this in perspective, there are now over 400 million Facebook accounts globally; there are an equal number of blogs, and Twitter accounts are mushrooming. Consumergenerated media and social media infrastructure has been building exponentially for the past 15 years. Combined, these entities amount to a massive repository of media impressions, many of which implicate (or reward) brands. The challenge of such media is that it is difficult and some cases, impossible to control. Much of the brand-related chatter emanates directly from brand experience, and that can not be easily changed overnight. Indeed, solid foundations must be in place for word-of-mouth to have it full effect and impact. Brands are increasingly trying to maximize the play of the two as they combine online and offline ad buys with offline PR and social media. Thus, the term earned media does include the offline and oft-used PR input. Another term we sometimes hear in marketing circles is owned media, which might include the brand website, in-store advertising, and the like. For the purpose of this exploratory, we put the owned media somewhere between paid and blended.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 4

Spurned Media: When Earned Goes Negative

One advantage of paid media is it is predictable the marketer remains completely (or almost completely in the case of media such as Doritos co-creation ads) in control of the messaging. Earned media doesnt function that way, and quite often as hundreds of brands have learned the hard way it often leans negative. We call this spurned media -- earned media that goes horribly negative, invades otherwise pristine search results or bleeds into traditional media. Bad customer service is a top driver of "spurned media." Spurned media can include counter-claims by consumers, hostile reviews, viral waves of negatives (e.g. lack of transparency), off equity content, activists hijacking or co-opt ad messaging, or various other challenges such as to Green or Health themes. Paid advertising even TV spots are increasingly counter-balanced by spurned media.

Maximizing the Paid & Earned Synergy

The critical question for todays CMO, or media planner, or research assigned to boost advertising effectiveness is this: how does a brand maximize the interplay between paid media and earned media. How do we ensure the two are complimentary and not at odds with each other? Can the two be blended to maximize impact? If so, what variables or factors might make a meaningful difference in the blending process? We focus special attention on four key measurement considerations and approaches to both optimize and determine return on Earned versus Paid media frameworks.

Brand Readiness: Brand readiness reflects the degree to which the brand is prepped and primed for incremental media impressions and online conversation. This includes having the right listening platform in place. Indeed, a growing percent of earned media can be primed through such things as influencer identification, ensuring the brand website provides ample currency to spread the message, and collecting feedback through consumer relations (which signals respect for the consumer and nurtures advocacy). Brand readiness also primes earned media echo effects from paid media inputs e.g. should the ad copy also be placed on YouTube or the brand site? In a social and digitally enabled world, the readiness list continues to grow. Brand Agility: Brand agility is the degree to which the brand is primed to act in actual or near real-time in response to stimuli for the purpose of increasing exposure or media impression count. Does the brand, for instances, have resources (internal or agency) ready to nurture, propel, advance, or occasionally sandbag and defuse the conversation? If early buzz on an ad fixates on a certain aspect of the copy,

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 5

brands with high levels of brand agility are able to respond through ad tweaking, PR interventions, social media engagement, and more. The oft-used term in research circles, sense & respond ties closely to brand agility. Brands with growing reserves of employees on Twitter, Facebook fan sites, or even brand-hosted communities tend to score well on the agility front. Brands who staffed their CRM and Twitter accounts 24/7 during the Olympics displayed high levels of brand agility. Brand Advocacy: Brand advocacy is at the heart of earned media, and reflects the degree to which consumers recommend, endorse, praise, or publicly bear allegiance to the brand. (Facebook fans, typically reflect high levels of brand advocacy.) Brand advocacy is both the requirement and engine of word-of-mouth. It can be quantified and translated in many ways, including via Nielsens Brand Advocacy Quotient (BAQ). Moreover, platforms that lend themselves to brand advocacy are proliferating across the web. A high quotient of brand advocacy before a media buy generally increases odds of favorable viral or conversational lift. Brand Latency: Latency measures the degree to which the content sticks or how it appears in search results or other critical places in the online consumer purchase funnel. At Nielsen, we measure latency through an instrument called iShelf that assigns value to brand positioning in search results, not unlike how category management assigns value to shelf-space. Brand latency might suggest, for instance, that Nationwide insurance still reaps millions of dollars in earned media impressions for their highly conversational ad in the Super Bowl four years ago featuring Kevin Federline. Latency is also reflected in Wikipedia brand entries, YouTube search results and tags, and just about any social media discovery engine.

A New Measurement Framework: The Blended Media Score (BMS)

With such variables in mind, Nielsen recently embarked upon an effort to see whether there might be an opportunity to combine both paid and earned media metrics to provide a more complete holistic view or campaign, launch, or event effectiveness to marketing stakeholders related. And so we developed and tested a new metric, the Blended Media Score (BMS), with the goal of giving brands and content providers a more complete view of ad effectiveness. This BMS metric tracks the impact from traditional Paid Media (TV ads, banner ads) but also blends data from online buzz and social media what were calling Earned Media.

How is the Blended Media Score calculated?

The BMS is the sum of an ad/brands paid media efforts and the earned media gained online through consumer conversation and action. Our approach is an empirically-derived algorithm which takes into account key metrics for each paid and earned media ensuring that both components are equally represented. All metrics are normalized to the same scale, summed, and then indexed about the mean to derive the overall score.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 6

The BMS is indexed to the mean which is equal to 100; therefore, we consider indices between 90-110 as average performers. Ads/brands that receive a score below 90 signify under-perform relative to competitors, while scores above 110 indicate high performance.

What are the paid and earned indices?

To add context we provide independent indices for the paid and earned media components to illustrate how an ad/brand performs for each element. Similar to the overall blended score, these metrics are indexed about the mean. Accordingly, the overall score is not an average of the earned and paid scores, but rather a composite of all metrics. A weighting scheme is applied to the metrics to reflect their value in the broader scope of advertising effectiveness. We find in both paid and earned media that recall, appeal and reach are the most important metrics and therefore give them slightly more weight than variables such as sponsorships/other in-program placements or change in Twitter followers/Facebook fans.

Super Bowl Case Study

There is no better environment to road test this Blended Media Score than the Super Bowl, where the online conversations is already prevalent and can impact brand perception in new and meaningful ways both before and after a traditional advertisement hits the airwaves.

Unique Considerations
The metrics considered were focused to accommodate Super Bowl advertising evaluation, and the paid media variables limited to only TV metrics to simplify the analysis. Earned variables are limited to consumer responses to those TV ads; therefore, customer service and product performance do not play a role in ad evaluation, and influencers/advocates do not play a part in immediate post Super Bowl reactions. Below is a simplified version of the metrics included in the Super Bowl analysis:

PAID

EARNED

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 7

Blended Score Super Bowl Results

The top Blended Media Score performers for the 2010 Super Bowl managed to perform well in both earned media (buzz volume, sentiment, increased engagement through social media channels) and paid media (recall, likeability, audience/reach). Brands that succeeded in only one area fell in the middle of the pack for their BMS ranking. For example, Focus on the Family over-indexed on earned media due to their very high volume of conversation (122) but under-indexed in paid media (80) causing them to fall 16th out of 43 advertisers.

Paid and Earned Media Scoring by Brand

Brands that offered free products or trial incentives (Dennys, Dockers) over-performed in earned media relative to other advertisers. The offers not only spurred higher levels of buzz, but also appeared to provide unique pass along currency to consumers. Specifically, Dennys succeeded largely due to the earned media its advertising gained because of the free factor. Offering free breakfast won over consumers and led to very high likability scores as well as a large amount of online discussion. Their frequent updates and teasers on Facebook led to many individuals becoming fans of the brand online. The top six brands all over index on both paid and earned media components and not surprisingly showed a rounded offering. Brands seven and eight on the blended score rating as you can see in the chart below. Coke and Bridgestone, had very strong paid components but slightly less than average earned media.

Biggest Bowl and Biggest Buzz

Nearly 107 million tuned into CBS telecast of Super Bowl XLIV, making it the most watched TV program ever in the United States. The game also generated the highest volume of online conversation and Earned Media, with conversation spikes on Facebook and Twitter playing a disproportionate role the buzz. Key Data from The Super Bowl: 14% of home Super Bowl viewers with Internet access browsed the web at least once during the big game, up slightly from last years 12%. This compares favorably to the Olympic Opening Ceremonies where 13% of viewers multitasked. These multitaskers average time spent online also increased from 24 to 29 minutes. This total was less than the 32 minutes spent by Olympic viewers. 36% (38% Olympics) of Super Bowl users visited Google.com and 34% (41% Olympics) visited Facebook.com. Facebook was visited during the game by 1 in 20 of all at-home Super Bowl viewers with Internet access, averaging 19 minutes per user.
*Data from Nielsen Convergence Panel and select National People Meter homes

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 8

Scoring and Blending Brands

MEAN

242 225 189 161 140 134 129 124 121 115 Blended Score

Budweiser Beer Doritos Tortilla Chips Denny's Restaurants Dockers Clothing Snickers Candy Google Coca-Cola Beverages
Bridgestone Tires

147 162 164 136 101 111 129 153 160

195

117 117 93 94
136

130 121 Earned 110 105 Paid 150 200 250

E*Trade Online Financial Emerald of California Nuts 0 50

100

100

E*Trade leveraged an integrated social media campaign to gain attention of their Facebook page and YouTube channel. They integrated their Facebook page across their full advertising strategy; as a result, they experienced a large influx of Facebook fans and experienced the second largest increase in fans among all advertisers.

Game Recap Learning: Does Timing Matter?

Until Super Bowl 2010, the season finale of M*A*S*H in 1983 aired the highest commercial minute, when an estimated 108.9 million viewers watched the second half hour of the program. However, a Doritos commercial featuring two men attacked in a gym for stealing someone elses Doritos was seen by an estimated 116.2 million viewers in the last Super Bowl which made it the most watched television commercial of all time. The ad ran in the games fourth quarter at 9:30pm ET. Audis Green Police ad earned the title of second most viewed ad with 115.6 million watching. Electronic Arts spot for its new game Dantes Inferno drew 115.1 million. Focus on the Familys ad featuring Tim Tebow tied for the least viewed ad of Super Bowl XLIV, despite the heavy pre-game buzz going into the game.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 9

There is a positive correlation between seconds of advertising, buzz volume and ad appeal. Many of the brands featuring only 30 seconds of paid time receive lower appeal ratings and minimal buzz while the more appealing ads receive more buzz. Budweiser is a clear leader with both high levels of paid time and subsequent online discussion. Doritos also is a strong performer and receives more bang for their buck they paid for less time than Budweiser, had consumers make the creative (ads), yet received more buzz and higher appeal rating.

Budweiser:
Budweiser garnered points for buying the most seconds of advertising (300) and placing several of their ads during Q1, leading to high recall, as well as Q4 when viewership was at its max. Five out of the six Bud ads were both highly memorable and likeable based on next-day viewer response. The ads, combined with the brands sponsorship of the aerial coverage of the game, gave Budweiser a heavy paid presence during the telecast. On the earned side, the Clydesdale ads generated a large amount of discussion. The brand also leveraged a Facebook campaign leading to a significant increase in Facebook fans as fans were asked to vote on which ads should be aired during the game.

Doritos:
Doritos succeeded again this year with co-creation of commercials in an online contest. Three out of the five most effective ads of 2010 were from Doritos. Although one ad, Snack Attack Samurai , was polarizing based on TV viewer response, online chatters loved the Doritos ads overall and stated they were the funniest of the night. Doritos purchased fewer seconds of advertising than Budweiser but they appeared in strong timeslots and rounded out their Super Bowl presence by sponsoring the Halftime Report. Post-game fans flocked to follow Doritos on Twitter. Additionally, ccording to Nielsen Netview, Doritos saw almost 430,000 unique visitors to the website during the pre- Super Bowlad voting period.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 10

The Role & Impact of Brand Readiness

High readiness ratings did not consistently convert into high BMS scores; nevertheless, among the top 10 according to BMS, both Doritos and Bridgestone received high readiness scores. (Doritos has the second highest BMS and the highest readiness, while Bridgestone ranked 8 by the BMS with the second highest readiness score)
Brand Buzz Volume Brand SB Buzz Twitter Linkage FB Fans Sentiment Followers Buzz

Appeal

IPP/ Sponsor

Paid Ad

Audience Recall

Budweiser Doritos Denny's Dockers Snickers Google Coca-Cola Bridgestone E*Trade

n/a n/a n/a n/a n/a n/a n/a

Brands could succeed without having all the boxes checked off, but then they needed to over compensate in another area. For instance, both Budweiser and Doritos are in the bottom tier for one social media platform. In Budweiser's case, the brand far outperforms all others on the paid variables that the deficit in Twitter is balanced out. Doritos, on the other hand, compensates for the lack of lift in Facebook fans by over performing on buzz volume as well as gaining slight incremental value through the paid variables second of advertising and recall. All top brands were successful in creating appeal toward the ad(s), generating buzz linked to the super bowl, and receiving a social media lift in Facebook fans OR twitter followers.

Ad Recall
Super Bowl had a record audience size with viewership growing throughout the game. Ad recall, however, declines throughout the game. Nevertheless, many ads airing in the third and fourth quarters receive high recall levels when the audience size is at its peak.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 11

Other Results and Learning

Nielsen learned through in-depth analysis of both paid and earned media inputs in the 2010 Super Bowl is that the interplay of the two made a significant difference for the brand. It increased their overall level of conversation and buzz, and primed the spot for an ongoing annuity of free media through search results and site indexing. We also learned that brand readiness and brand agility played a key role. Preliminary learning from the 2010 Winter Olympics and the 2010 Oscars seem to echo these learnings.

Opportunities & Guidance from Super Bowl Analysis

1. Think hard about a compelling call to action. Advertisers leveraging a distinct online call to action (Dockers, Dennys) earned high levels of buzz, and greater interaction with their social media touch points. Brands offering free samples also sustained online buzz for a longer period of time than other advertisers. As social media expression venues proliferate, this may well become an even greater opportunity. 2. Brand Readiness really matters so line-up as many success factors and variables as possible. Go Daddy ads drove a higher number of survey respondents to their website), but didnt score in the top 20 Blended Media Score. This occurs because the ad has low appeal ratings and relatively low movement among Twitter followers and Facebook fans with only average recall. To be a top performer, the ad/brand had to succeed on more than one variable. (Additionally, website visitation is not a variable in the Super Bowl BMS and an issue that will be addressed in the future.) 3. Humor and free trials continue to drive highest levels of buzz and earned media. These talk drivers tended to drive significant activity as both a call to action and a unique pass along for online currency. 4. Steer away from half-time ads in favor of half-time conversation drivers. Game day buzz is highest during the half-time show when ad recall is at its lowest. Since consumers already spend more time talking online during half-time, find creative ways to encourage conversation, with or about your brand. Since Simultaneous Visitors are largely active on social media, while they are not focused on the game, try capturing their attention online. Consider having a conversation with your fans/followers during half-time or host a half-time contest on your website related to the game 5. Brand Latency is an ongoing game. Brands cannot determine true impact of all adds immediately as continued exposure occurs from the latency effect in search results. Pepsi, not an advertiser this year, still continued to receive views and placement from previous efforts.

2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 12

Brand Readiness Check-list for Maximum ROI

Readiness Check-list Brand

The Super Bowl (as well as the Olympics) provided data proving the importance of Brand Readiness to enhance customer feedback and commentary as well as Brand ROI. Readiness doesnt necessarily mean direct brand engagement in the conversation but instead mostly refers to ensuring the consumers have the means to easily discuss brand related topics such as the commercials. A Super Bowl Readiness Checklist can be seen to the right. Key points include ensuring mention of and video of the ads on the brand website to drive traffic and searchability for the ad via internal brand search and external search. Additionally, a presence on key social networking sites is a must to truly empower consumer discussion and enhance ROI from the paid ad purchase.

Website Preparedness Ad Mentioned on Brand Front Page Ad Video on Brand Site Separate Website for Ad Availability to Provide Feedback for Ad Ad-Related Mobile App Search Ability to search for ad on Brand Site Brand Sponsored Google Ad Links (SB) YouTube in Ads Google Results (10) Ad Presence on Brands Wikipedia Page Social Networking Official Brand Facebook Page Official Brand Twitter Handle Official Brand YouTube Channel Readiness Rating:

Future Exploration, Conclusions and Drivers for Organizational Action

BMS and the entire theory behind Media Mix Modeling 2.0 is not a finished product. The industry continues to evolve and the score must evolve with it. We continue to work to answer (among many others) the following questions to better model consumer behavior and their interactions with media be it online, offline, digital, paper, etc: What drives the latency effect? Is the impact from free giveaways/contests lasting? What readiness metrics truly impact long term performance? Is there added earned media or conversational value in having a spokesperson? How do brands fare that leverage only an online campaign compared to brands that leverage only paid TV advertising? Is one form more successful than the other? Regardless of the questions that still to be answered, it is apparent that media leaders cannot afford to look at paid and earned media in isolation. Media must function as part of an integrated campaign which works towards the desired consumer action. Failure to for the left hand to know what the right hand is doing will not just result in less value for marketers, a big concern in itself, but could cause negative reactions and spurned media, damaging the brand long term. Simple but integrated tactical steps can drive this synergy within organizations and increase success of campaigns. Stay on equity, stay transparent and stay agile to respond to consumer needs. The media world is complex, continuing to splinter, and driven by creative and ever more demanding consumers. This requires business leaders to be insightful and create new frameworks to embrace the multitude of access channels to those consumers if businesses wish to remain relevant in their consumers lives.

-C 2010, The Nielsen Company

Blackshaw Media Mix Modeling 2.0 13

About the Author

Pete Blackshaw, EVP of Digital Strategic Services, Nielsen Pete Blackshaw, whose professional background encompasses public policy, interactive marketing and brand management, is Executive Vice President of Digital Strategic Services for Nielsen. Petes strategy group works with many of the worlds top brands and corporations to develop cohesive, consumer-centered digital programs and strategies. A 2010 grand prize recipient of the ARFs Great Mind award, he is the author of a recent book by Doubleday entitled Satisfied Customers Tell Three Friends, Angry Customers Tell 3,000: Running a Business in Todays Consumer Driven World, and writes a bi-weekly column in Advertising Age centered around the books themes. A former award-winning interactive marketing leader at P&G and founder of consumer feedback portal PlanetFeedback.com, Pete co-founded the Word-of-Mouth Marketing Association (WOMMA). He is also the Chairman of the Board of the National Council of Better Business Bureau and in that capacity also sits on the National Advertising Review Council. Hes a recipient of industry achievement recognition by both Ad-Tech. He advises a host of non-profit organizations on digital strategy including the United Way of Greater Cincinnati, National Underground Railroad Freedom Center and the Cincinnati Youth Collaborative. Pete, his wife Erika, and their three children live in Cincinnati, Ohio. Pete is a graduate of Harvard Business School and the University of California, Santa Cruz.

Other Contributors
Nina Stratt, Senior Analyst, Measurement Science, Nielsen Alka Gupta, SVP Research, IAG, Nielsen Kim Cox, Senior Analyst, Buzzmetrics, Nielsen

For more information on Nielsen Buzzmetrics or Nielsen in general, please contact Josh Hammond at (859) 905-4973 or Joshua.hammond@nielsen.com

2010, The Nielsen Company