It's just about the content?

Last week, Millward Brown produced their top ten hints on maximizing the chance of a brand video going viral - there's info on the report here and here.  It's a pretty remarkable read as I think it captures a lot of what is wrong with research today and, more broadly, a view of how communication works and our ability to create stuff that is 'viral'.

The first warning sign is the assertion that of the 102 ads tested, less than 15% were viral hits (ie watched more than 5,000 times per week on youtube in the US).  Yes, this was a test of what TV ads became 'viral', not content in its broadest sense (and seems a far higher number than the reality of things that really begin to spread and take hold through culture).  'Viral' is defined by times viewed, not its 'spreadability'.  At best, it's a study about amplification of an ad through unpaid media.  A lot of the tips also seem to focus on 'tricks' to buy virality, rather than trying to understand the nature of things that spread.

Then they do their usual trick of saying that often 'viral success' has little to do with what really matters - 'persuasion'.  So, MB are still hanging on to an increasingly discredited view that advertising works only (or at least primarily) through persuasion and that the decoding of commercial messages is critical, not whether the brand is able to play a bigger role in culture (perhaps by creating it).

Finally there is a brilliant moment of have your cake and eat it.  After trotting out their usual ingredients of successful advertising (and if you've done any MB debriefs you will know these by heart) - awareness index, brand impression, use of celebrity - they then have the audacity to talk about "keeping your fingers crossed".  Arguably, the best advice ever given by Millward Brown but not one that I would want after paying tens of thousands of dollars for research predicated and sold on the power of a predictive in market score.

It's research like this that is increasingly to blame for the increasingly drunken way we make decisions (the old Ogilvy adage of people using research like drunkards use lampposts - for support, not illumination) and our attempts to predict the unpredictable.  It may be 'nice' to have a number for the board meeting, but it is often a number that bears little resemblance to reality.

Comments

[Claire Grinton]

So how do we change client expectations and internal practices to move away from using studies like this? How do you suggest we solve the problem of clients expecting the numbers they can report and measure?

[Kenji Summers]

This puffery is an issue that seems ongoing. I'm hoping that some transformation can take place. I think this will only happen when agencies fully allow themselves to be immersed in the cultures they want to engage. I.e., becoming curators for art galleries, community centers, libraries, museums, etc. It will take some agency folk that will allow themselves to become mermaids or mermen. Because it is the subcultures that will inform mainstream culture (s). How does what I've said relate to going viral? MB is just wishing upon a star with the info they presented. Viral mirrors human evolution. (I won't do deeper there. Ha.) Money spent and money lost. Ogilvy was a smart human being.

I concur and even go further, "illumination will lead to the education for the dumb, deaf, and blind".

[Sam]

Boom. This is a textbook deconstruction on the folly of research.

I totally agree that the curation, facilitation and extension of culture is the way forward.

When it comes to being able to 'sell' through to clients, I've found in my oh so brief experience that doing the desk research to show them how these communities exist and interact presently online is the only way to get them on-side. Even if you're pitching the creation of something brand new (sorry Gareth!) the fact that you can find an analagous example online is invaluable.

I'm young and foolish though, so I could be talking nonsense.

Great post regardless.

[Alasdair Lloyd-Jones]

Why do research houses feel more and more like government departments?

What upsets me most about this is that clients do need something to lean against and these research companies are doing nothing to help them lean against something more relevant to today's culture.

Lazy...

[Duncan Southgate]

Interesting criticisms Gareth, though I think they’re all pretty unfounded! It’s clearly not just about the content, as we’ve been at pains to point out. However, our research is the first to quantify the massive effect which creative strength has on viral success. Dealing with some of the other specific issues….

1) Views per week
You’re right that views per week can also be expressed as which ads get amplified through unpaid media. Clearly this in turn is a mix of how widely an ad is seeded, and how quickly it is spread or propagated. We discuss that in detail here: http://www.mb-blog.com/index.php/2010/02/03/big-is-beautiful-when-it-comes-to-viral-media-too/
Since views per week and propagation rate are clearly strongly correlated, I think it makes perfect sense to use the more widely available views per week figure as our measure of viral success.

2) Persuasion
We don’t say that persuasion is “what really matters”, nor that it “works only or at least primarily through persuasion”. Ann Green was simply quoted as saying that persuasion doesn’t have a relationship with viral views per week. We know that ad engagement can sometimes be enough to drive impact and sales, particularly for established brands. However, we do see that persuasion also plays a role for many brands, particularly in the short term and for newer brands.

3) The usual ingredients
The full papers which we have presented at WARC and the ARF make it clear that the recipe has changed. Yes, The Awareness Index is still important. But the measures of Buzz, Celebrity Profile and Distinctiveness are clearly of new importance in driving viral viewing.

4) Support for decisions
I’m sure advertisers will find it illuminating to know in advance of placing an ad whether it has a decent chance of propagating. Before spending significant sums on viral seeding and promotion, surely it’s best to know if the ad is worth investing in? And surely it’s better to do that using validated metrics which learn from previous success?

5) Luck
Viral is different to TV. There’s no way we can predict which ads will inspire mash ups, and we can’t easily predict the online spillover effects from TV and other media spend. So it’s honest to say that there will always be an element of luck with viral campaigns. However, we can isolate the role of the consumer response to the creative, and that’s all we’re trying to do here. I guess some folk will always prefer to use their own judgement and estimation, but consumer opinions have always felt important to me.

[Gareth]

Duncan,
Thanks for joining the debate. Well, unsurprisingly, I still take issues with your defense and I don't believe they are unfounded. Rather they are different views on the world. Taking your points in turn:
1. My issue is the focus on video as viral content or even a proxy for viral content (even if the word 'viral' is I think the wrong metaphor). It's bigger than video and fundamentally different.
I know Duncan's research well about big seed theory and I don't think the propagation rate simply equates to the creative power of the ad or even views per week. I guess, fundamentally, I have issues with the methodology and application of your research.
2. Even if you are softening your view (which admittedly seems from your comment considerably softer than the article) I still wonder on the real power of persuasion in this context. A participatory landscape is rather different to the passive 'send/receive' model of TV ads - how does your research allow for this? Even video I believe behaves very differently in this space. Persuasion does work sometimes, for the right marketing objectives, but it works as a weaker force than most of our ad frameworks would suggest (this isn't new or post digital, it's at the heart of most of Andrew Ehrenberg's work on the weak force of advertising)
3. Fair point, although I would argue these factors are all about the content, not the context and perhaps more importantly the people. A lot of emerging research, perhaps popularized most by Mark Earls, points to the fact that it's less what we do and more about what people do to what we do. These are different factors to the content based one you mention.
4. Hmm, I kind of agree with your sentiment but the question is how predictive is the prediction? Isn't a better model one more of agile development, placing lots of small bets and seeing what works in the real world and scaling up behind those successes and quickly killing the ones that don't (the lots of matches to start a forest fire theory). Again, I have some issues with the continual return to ad based models. I'm much more comfortable with a model based on real world learning, particularly as the cost of failing on the web quickly reaches zero, or at least less than the cost of some pre-testing research that really might not be that predictive.
5. People's opinions are important to me too and I take issue with your suggestion that perhaps I don't think this. But their behavior is even more important to me than their opinions. I guess that's where our world view perhaps differ the most.

I don't think luck is unique to 'viral' (actually, another reason why i hate viral as a descriptor as viral campaigns would require no luck) but a reality of a world that is characterized more by unpredictability than predictability.

I guess we just have different views (and we would both argue evidence) of how the world, people and communication works.

Thanks again Duncan for participating in the conversation.

[John Burke]

How the world works is ever changing, daily. There is little reliable prediction outside of yesterdays behavior patterns. This is where analytical data falls short.

Events happen and things are communicated so rapidly that behavior can change. look at Nestle.

So behavior yes but deeper It's about motivation. The why not the how.

[John Burke]

Let me reiterate I have no answers on how to report success back to client in this model other than widgets sold or traffic volume. Easy peasy.

Statistical data to prop up a spend doesn't concern me. Sell the idea on the customer motivation which is not always found in behavioral patterns. Far from it.

What media / medium is usual easy to suss.

My 4cents (spent 2 on the last)

[Duncan Southgate]

Hi Gareth,
Thanks for a very considered response.
Re 2 we account for the importance of sharing (beyond receiving) via our Buzz measure. What we can't account for in a pre test are the media effects (influence of being a featured video, value of promoted video etc.). This is where I'd love to do more research, to see how those investments build on the ad's creative potential.
Re 4, some advertisers may be in a position to place lots of small bets and see which take off, and I absolutely agree that we all (advertisers, agencies, researchers) need to be more agile these days. Even then, I'd suggest that consumer insight into the kind of emotional reactions which are generating that behaviour will help them understand how they may be able to repeat the trick in future.
Agree 5 seems to be the core of our differences. Apologies if I implied you're not interested in consumer opinions. My pov is that the relative importance of behaviour vs. opinion surely depends on your ad objective. Some direct response brands may be exclusively interested in behaviour, but for most brands the views and clicks are just one part of the equation, and the attitudinal response is just as important.
Work we did last year with Google/ Doubleclick for all kinds of online ads showed that there's very little correlation between behavioural response (clicks, interactions etc.) and brand impact. So the combination of brand response surveys and behavioural interactions data seems like the ultimate picture to me. Skilful integration of the two is going to be a major digital research challenge for the coming years.

[Gareth]

Duncan

Thanks for your reply. Clearly we have some worldview differences but I guess that's what keeps life interesting. The one thing I would question though is the point about behavior and opinion depending on the objective. I believe behavior is what matters full stop. All the work by Ehrenberg et al has shown that attitudes follow behavior and not the the other way round. Changing our behavior changes how we think, and not the other way round. This reality is ignored by most ad models and I think is critical to building a better, more effective model of communication

[Mark Earls]

Good stuff, G.

Thanks for the shout out but more particularly for bringing Andrew Ehrenberg back into the debate.

As for your well made point about how advertising might work, don't forget the 1984(?) paper by Judie Lannon and (the sadly late) Peter Cooper - "Humanistic Advertising".

For my money, we need to go a step beyond this: it's really the people doing stuff with each other (rather than what they do with our advertising) that really shapes effectiveness.

Clearly, that makes it much harder for MB et al to do their stuff.

[faris]

hahahahahaha

viral.

anyway....not much to add except that, well, kind of have to challenge the notion that function of advertising is to encode and transmit commercial messages.

Because the ones that work best, according to a meta-analysis of the IPA databank, which contains 800+ IPE effectiveness award winning papers, don't have any rational content at all:

http://farisyakob.typepad.com/blog/2009/12/no-rational-content.html

[Graeme]

Well said Gareth, and something that really needed saying

So the bit that MB can't account for is the 'paying for people to watch it' bit, but they can tell you how many fingers you need to cross?

Grrrr. I'd like to be amused by the use of 'viral' but it seems like there are people with large research budgets who will still buy this stuff. I won't pretend to have better explanations than those already in the comments, but anyone at MB still unsure about what they are doing wrong should really read Bud Caddell on the subject http://bit.ly/90Chmh

[John Burke]

Is not behavior and 'attitude' a two way street? Ehrenberg and similar work aside in my experience of the world they are mutually affective.

Though I would agree that changing behavior changes thought without question but beliefs are expressed outwardly. I just do not see a one way street. That ia not my -experience- of the world.

I can think of many life examples to make my case.

This thread is fantastic btw. Thank you

[Mark Earls]

John B. It's not just Ehrenberg - it's what the behaviour change literature tells us, too whatever anecdotes from our own lives or practice we might choose to cite.

Again and again it's clear: our stated opinions and attitudes lie in a much looser causal relationship to our behaviour than we tell ourselves and mostly follow rather than precede behaviour change.

Of course it seems otherwise to agents whose behaviour we seek to change and those observing it case-by-case: that's due to the way our mind confabulates our sense of agency.

[John Burke]

Tx Mark. I was pulling off the work of Talcott Parsons as well as my holistic life experience. Fundamentally I am in complete agreement however.

You make fantastic points. Any additional materials you can suggest along these lines to look into?

[Thomas Wagner]

Great post and debate.

In addition to Lannon/Cooper one might also refer to "The Social Uses of Advertising: An
Ethnographic Study of Adolescent
Advertising Audiences" (http://www.marketingritson.com/documents/socialusesofadvertising.pdf) and "Advertising Uses and Gratifications" by O'Donohoe (http://homepage.ntlworld.com/ashley.payton/p52.pdf) for more on the context, instead of the content part. Both articles are from the pre-"viral" era and well worth a read in my humble student opinion.

[Edward Ludd]

Hey you guys stop picking on poor old Millward Brown - you're behaving like schoolboy bullies picking on the guy who's not into the supposed cool music. My view is that there are creative themes and rules that could be followed (so maybe MB are on to something in there), but agree on the Watts/Herdmeister luck/small bets thinking - maybe MB need to delve deeper into the rules/themes a bit more - but so long as all this sort of stuff it is seen as an aid for creativity and planning not the black box.
On the Ehrenberg front - love his thinking (attitude after behaviour / in fmcg we're not really loyal to most brands and are portfolio shoppers / that role of comms should be to "nudge" (not in the BE sense) a sale) But I just don't see how it refers to most non-fmcg/portfolio brands. Yes get the social/Herd element for lots of brands and categories (get them to do stuff together etc..)- but am a big fan of the Feldwick association/nuances thinking (of course there are social/Herd aspects to this)- I think of Patex Philippe as a premium quality special exclusive heritage Swiss watch brand - I think I've been persuaded that this is the case almost exclusively by image association from advertising and sponsorship (nothing participatory and nothing by my behaviour or any influence from other people I know or have seen wearing one). Guess I'm kind of saying is that there are a lot of different influences out there: Persuasion, behaviour, social and particpatory as well as image association (plus others)- you schoolboy bullies seem to just focus on the ones that suit your work and desire for the new etc ;-)

[Adam Ronich]

It's a case of MB putting together yet another product to sell to clients. Here's my biggest problem with it. How can you have a Top 10 List for Viral Advertising and have "creativity rules: a distinctive and engaging ad is a necessity" as one of your main conclusions.

So in other words, make sure you're original and authentic by doing these things others are doing.

The logic does not compute.

[Rob @ Cynic]

It was worth the wait. Just.

Now can you update this more than every passing Hayley's comet?

PS: It really was very good.