For my fourth interview in the Word-of-Mouth Wisdom series, I decided to tap two of the smartest people I know in the field of marketing.  Dr. Peter Fader and Dr. David Reibstein both teach marketing at The Wharton School, where I was fortunate enough to earn my MBA.  Both have been friends and advisors ever since graduation, and somehow I convinced them to invest in Bazaarvoice!

Pete is well known on many levels.  He was helping CDnow run analysis back in the pre-boom times.  He has been very outspoken in the age of digital music, advising music companies on how to market in these rapidly changing times.  I remember him best as my Markstrat professor, one of the better MBA classes I had the pleasure of taking.

Dave is also very well known.  He consults for companies all over the world.  He served as the Executive Director of the Marketing Science Institute.  And few know him as the co-founder of BizRate, where he served on their Board of Directors from its inception to when Scripps bought the company for $525 million in cash almost two years ago.

So begins our interview…

1. You two have studied and taught marketing for your entire academic career and now teach at my favorite business school.  What do you think about the recent trend of user-generated content?  Will it change the world of qualitative and quantitative marketing analysis?

Dr. Fader:
First, it’s extremely important to distinguish amount different types of “user-generated content.”  There’s a huge difference between a daily blog entry versus posting some comments on an open-ended discussion board versus a customer filling in a feedback form on a commerce firm’s website.  And this doesn’t even get into “real” content such as a homemade video.  It’s a big mistake to treat all of these sources of customer input as interchangeable with each other.

Second, yeah, there’s a ton of new stuff being created every day, but the big question is whether customer behavior is fundamentally different than before, or are we just getting a new window to help see the behavioral process in a way that we could never see it previously?  I tend to be relatively conservative about this (i.e., I believe more in the latter than the former) and this has important implications for model-building and analysis.  Too many people are saying that it’s a “whole new world” and are throwing out the old models/metrics.  This was an enormous mistake for many firms at the dawn of the Dot-Com era, and I think history is repeating itself again.  As a specific example, most of today’s research on social networks is inferior to work that was initiated back in the 1960’s.  The available data are way better today, but firms are drowning in it and not being very thoughtful about how they use it.  Too many firms are coming up with “fun facts” to impress senior management or investors, and not drawing specific, meaningful insights that really add value to the data-collection process.  It’s important to do some serious thinking before looking at any data.  Come up with some specific hypotheses, or guesses, about the relationships you expect to see, then test them.  The complete opposite of this is “data mining,” which is what most companies tend to do.

Dr. Reibstein:
First, I agree there are different types of “user-generated content,” and they should be treated differently. 

Let me add that there is a reflection that there is significant bottled-up sentiment that many people have to share.  Let me start with a caution about some of this content, however.  As with survey research, and the same holds for user-generated content, beware of non-contributors.  In survey research, we would refer to this as non-response bias.  Nixon (I am dating myself) referred to it as the “silent majority”.  The question is what are the feelings of those that didn’t reply or generate any content.  Were they more satisfied and didn’t feel a need to comment?  Dissatisfied customers, experiences, generate many more comments than satisfied.  Mediocre experiences generate the least.  The assumption with survey research, customer satisfaction surveys, and beyond is the non-respondents are not any different than the respondents.  And, in many cases, that is correct.

All of that being said, there is the common belief that consumers trust other people’s opinions and content more than they trust that of advertisers.  Advertisers’ motivations are to represent their products/services in the best light.  Other consumers are less biased.  Of course, this is a universal statement that should be challenged.  It would be safer to assume that some consumers are more willing to rely on other users’ input.  The question is what proportion.  I am sure this percentage will differ by category.

It is the case that consumers can probe and interact with other consumers in a way they never have previously been able to with companies.  Companies, in general, have not been able to figure out how to deal with the multitude of customers and address each of their questions, yet given a large number of customers, it is likely that customers could ask very specific questions and find user-generated answers.  Some companies have been able to harness this “community of users” to help address such individual questions.  For example, P&G has a user group for detergent where people can ask and get responses about how to launder and so forth and get responses from other consumers.  This has proven very powerful.

As for the impact on research, my suspicion is no.  The type of “user-generated” data does not fit traditional modeling and it is coming in such mixed forms, people have yet to fully figure out a systematic way to analyze it.  Let the fun begin, but as Pete says, don’t throw out all the good we have learned to this date about how to analyze problems.

2. What does this consumer-as-producer trend remind you of most from your background?  In other words, is there a previous marketing framework or trend that you relate this to?  [I am a believer that most trends are cyclical – and that is part of the reason I chose the word “bazaar” as the first part of our company name].

Dr. Reibstein:
I believe the “consumer-as-producer” trend is a recycling of all the old research from the 60’s of “word-of-mouth” communications.  We saw it resurface recently in such things as “buzz marketing,” “guerilla marketing,” “coolhunting” (the article), or even Gladwell’s The Tipping Point.  None of these were looking at the electronic world, but all were looking at how some consumers were influencing others.  We can even take this back to Veblen’s The Conspicuous Consumer from the 1890’s to see how far this really dates back.  No, I do not date back that far!  All the trend setting from the fashion world has been based on this understanding.

3. What do you think about the current trend of user-generated advertisements?  Do you think this is a fad, or a permanent change in the way ads will be constructed?  Will we see more of less of this in the future, and how will it evolve?  Some recent examples: Doritos, Oscar Meyer, GM Tahoe, and Yahoo!.

Dr. Reibstein:
In some sense user-generated advertisements are not as unique as they might seem.  Companies have previously tried to harness these consumer-initiated responses by showing ads which were interviews of “customers on the street”.  In these cases, it was sponsored by the companies, but they were selecting real people and getting authentic reactions.

Today, there are major differences—first of all, in the company-sponsored ads, they were able to be very selective and only show the ones that were positive for their company.  Today, it is out of their control.  Secondly, the technology is so much better, accessible, and less expensive, that people can develop their own ads.  The examples you gave are all very energizing and giving the power to the consumer, yet, once again, we have the companies selectively editing the ones to be shown.  So, this is no different.  Nonetheless, the approach has now got some customers energized to create very clever ads.  The agencies should be worried.  (I might note, this is consistent with the notion of reality TV.)

Another comparison is companies putting their name on apparel for customers to wear around.  This is a little different, but to a very large degree, it is having individuals electing to be walking billboards for their products.  It is one thing for Polo to put their brand on shirts and everyone can see who is wearing them.  It is another step forward for Budweiser or Harley to have their name on t-shirts and have people wearing them around.  This has been a long-time strategy of having users influencing other users and a form of pre-Internet social network marketing.

If this approach grows much, we can anticipate the novelty to wear off pretty quickly.

3b. I remember how often companies used video testimonials in ads; that is a good historical point.  The difference now is that consumers are creating the ads themselves.  And the main incentive for consumers to do so is social recognition.  That feels like a major shift to me and another unforeseen benefit of the Internet as a low-cost, easily-accessible, globally-connected form of communication.  Isn’t this the start of a permanent shift for agencies?  How do you think they will adapt?

Dr. Reibstein:
I agree that self-created is a shift, but the ones we cited are not truly self-motivated.  In each case, the action was initiated by the firm.  To the degree there are ones that are self-developed, like when people tattoo a brand on themselves, as has often happened with Harley, that is a real shift, but didn’t just start now.  It is a true sign, though, of consumer commitment.

Dr. Fader:
As Dave suggested, right now it’s a pure novelty, kind of like e-mail marketing back in 1999.  I can assure you that there will be no fuss about user-generated ads in the 2012 Super Bowl.  That doesn’t mean that such ads won’t exist, but no one will care.  This is my main point here: customers do not care about the origins of ads (how many consumers can name even one ad agency?).  They only care if the ad moves, entertains, or persuades them in a meaningful way.  If consumers can do a better job of this than advertising professionals (which, in general, I doubt they can do), then that’s great.

There is a huge difference between user-generated ads and user-generated comments.  Yes, consumers often want to see opinions from other regular people.  In that case, the source of the content really does matter.  But this is entirely different from advertising, and when the two come together, bad things happen…

3c. Pete, will you elaborate on what you mean by “bad things happen…”

Dr. Fader:
I’m sure you’re familiar with the growing list of stories that involve a firm posing as a regular user and “seeding” comments in a user forum.  And surely, for every firm that’s been caught with their hand in the cookie jar like this, there are a hundred others who have gotten away with it.  Not only do these episodes harm the credibility of the firm but they also harm the credibility of the forum itself (even if it’s not necessarily their fault for letting the fraudulent comments get through).  Anyway, this is the red zone between user-generated advertising and user-generated comments, and firms will continue to find ways to toe the line…

4. How do you think brands will figure out how to advertise in social networks?  Obviously the “eyeballs” there are enormous as MySpace is now one of the top-ten sites of the Internet.  And eMarketer predicts that social network advertising will top $1 billion this year (5% of what eMarketer projects will be spent on online advertising overall).  But are these mediums the right place for a brand to reach a customer?  MySpace users seem to fight “the intrusion” often.

Dr. Reibstein:
If you have people going to MySpace or YouTube or other social network sites and searching for keywords, they become primary venues for firms to do their advertising.  The risk is alienating potential customers by invading their space.  I suspect customers will be ready to adapt and not be too offended as ads start to appear more and more often.  For some customers, it will even have appeal as they are interested in certain topics and finding links to products/services might be of interest or ease for them.

An interesting phenomenon I am aware of is some companies, particularly start-ups, are producing their own video’s and rather than buying airtime or webspace are putting up these ads on YouTube and encouraging some individuals to go there and to “forward to a friend”.  So, these social networks are being used directly as an advertising medium.

[Note from Brett: see Wayne's post on Blendtec for an example.]

Dr. Fader:
RetailWire ran a discussion about this recently, and here’s what I had to say:
“I frequently use Facebook to snoop around on my students and I see many of these commercial-related groups there.  People gladly align themselves with brands that they favor, but this by itself doesn't mean anything.  No one really does anything there, and they can't compare to more socially oriented user-driven groups such as "So-and-so was incredibly drunk last weekend and I have the photos"!  So it might seem like a "competitive necessity" for brands to stake out some turf in these new worlds, but I don't really see why.  Many of these networks will collapse under their own weight due to excessive meaningless links (I think MySpace is in great danger of that right now), and serious investments that some brands may make in this arena will be largely wasted (albeit in a different way than the students they are trying to appeal to)”.  The full discussion is here.

And then there’s the use of these networks for “viral” advertising.  See my comment about e-mail marketing above…  much of it will diminish as the novelty wears off, but there will always be room for really good ads to be accessed and shared on YouTube, etc.  But the metrics that firms use to gauge the success of these ads needs to change.  I’m a firm believer that the Subservient Chicken campaign, for instance, was useless, regardless of how many people (including me!) wasted time on it.

4b. Dave, I hear what you are saying about YouTube and the viral effect of some video ads online.  Philips took a risk with the “Shave Anywhere” campaign and it paid off.  But it seems to me like it is going to be hard for YouTube to monetize their traffic unless they add a pre-roll or post-roll ad to each video.  And won’t that hurt the pass-along effect?  I can see how a company can produce a “made for YouTube” ad as a standalone “unit”.  But a pre-roll or post-roll seems much more difficult to me (yes, Yahoo! has these for their videos, but they don’t have near the video traffic that YouTube has).

Dr. Reibstein:
I think we will easily adapt to pre-roll and post-roll ads and learn to skip them.  Look at CD’s, videos, and DVD’s that have all added pre- and post-promotions.  We have learned to accept this, and in some cases, because of the capability of the targeting, have even appreciated the ads because they are relevant to us.  Certainly, when we go to movie theatres we now see the bevy of ads before the feature, not just trailers, but also for other products.  We have accepted these, and as I suggest, even enjoy them in some cases.

4c. Pete, doesn’t your response contradict one of the goals of branding?  If it is about mind-share, doesn’t it matter that PINK Victoria’s Secret has 210,620 fans?

Dr. Fader:
Sure, any PR is good PR and all that.  But the value of these connections is minuscule.  I suspect there is no discernable difference in the collective lifetime value for these V-S customers with or without these connections.

Likewise, I’m not saying that Subservient Chicken, et al, are bad ideas – I’m just saying that they don’t generate nearly as much in the way of sales as these firms would like to believe.

So it’s great for firms to jam a stake in the ground and say “we’re here!”, but they shouldn’t expect much economic reward from it.  The same is true of most branding efforts, even in traditional media.

5. Dave, you served on the Board of BizRate (now Shopzilla) since its founding.  Obviously that business was based on consumer-generated ratings of retail businesses, which powers their shopping search results.  Now new shopping search engines like and are popping up frequently.  How do you think customer ratings and reviews will help evolve shopping search?  What opportunities does this pose for marketers?

Dr. Reibstein:
As a co-founder and board member of BizRate/Shopzilla, I, from the beginning, have believed that consumers want to hear from other consumers about their product and merchant experiences.  It is an unbiased (or less biased) view that has some greater degree of trust.  The first opportunity this provides to the marketer is the feedback from the marketplace about how consumers perceive the offerings.  In some sense, the marketing research is being done for them.  The opportunity is to listen to it, not just to cringe from the bad ratings.  It will also be the opportunity to observe what seems to matter.  Is something poorly rated on certain dimension, yet still captures sales?  There is information in these ratings.

6. We started this interview with how the world of marketing analysis is changing.  I would like to revisit that in two ways.  First, what are academics doing to evolve that, given that both of you agree that companies are drowning in data and the methodologies being used are not advanced enough?  And, second, what would you recommend to companies that are planning to embark on a user-generated content future in their marketing so that they ground themselves properly? 

Dr. Fader:
Unfortunately, the gap between academics and practitioners in marketing is widening further than ever.  We teach the same old stuff as we’ve taught forever (albeit with fancier technology), and much of the content and methods we teach are far removed from the needs of managers.  This is particularly true in data intensive businesses such as e-commerce.  So we deserve most of the blame, but managers bear some responsibility as well.  The 24/7 crisis mentality of today’s executives makes it impossible to be thoughtful and deliberate in approaching data.  Quick and dirty solutions are not only tolerated but they are rewarded, while doing things the right way is not a popular path to follow.  So overall, it’s a pretty grim situation, which is very disappointing considering the coolness of all the new data that are coming out of companies like Bazaarvoice and its clients.  And there’s no indication that things will get substantially better in the near future – both sides will continue to drift further apart…
Dr. Reibstein:
I believe there have been several academics that have been making some good strides in understanding the wealth of data we now face.  One of them is my colleague and co-responder in this blog interview who has been too modest, Pete Fader.  He and some of his former doctoral students such as Bruce Hardie (London Business School) and Wendy Moe (University of Maryland) have been leading the front in how to analyze the onslaught of data.  Others such as Alan Montgomery at Carnegie and Florian Zettelmeyer at Berkeley have also been making major strides.  The outlook is not nearly as dismal as Pete paints it.
There are some companies as MarketingNPV and others that have been groundbreaking in helping firms organize their data and connect it to marketing outcomes.

Dr. Fader:
Too many people in the “user-generated content” area (as well as its cousin, social networking) think that everything is totally new and there’s nothing to learn from the past.  They would be amazed at some of the work that was done in sociology and other related social science disciplines years ago.  Likewise, they would be amazed at the power of the statistical methods that were developed back then (and unfortunately long since forgotten by marketing types).

I can suggest a few readings – a couple of classics are listed below.  But note that both are out of print, which speaks volumes about the state of affairs in this area.  Both require a bit more than the usual MBA statistical training.  It is essential for an advanced manager to have such skills.  In fact, I go further to suggest that any CMO (or CMO wannabe) should be able to read and appreciate this kind of work.
Greene, Jerome D. (1982), Consumer Behavior Models for Non-Statisticians, New York:Praeger.
Massy, William F., David B. Montgomery, and Donald G. Morrison (1970), Stochastic Models of Buying Behavior, Cambridge. MA: MIT Press.

My hope is that Bazaarvoice, with its deep grounding in fact-based management, will be able to escape the morass of today’s custom analytics, and I am eager to help them on this important mission.

Thanks again to both Pete and Dave for doing this interview.  They encourage you, the readers, to ask further questions or make comments via the comments section below.

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>