The Wharton Future of Advertising group hosted a two day “Empirical Generalizations on What Works in the New Age of Advertising and Marketing” conference, where academics and various private sector companies presented their findings. 

On the first day, the session kicked-off at 8:30am sharp under the theme "What is a Touch-point Worth." First up were some former colleagues of mine from Yahoo!'s research labs (now Microsoft research), examining how well we can forecast individual user behavior from their social network. In short, despite the “birds of a feather” expression, the conclusion was that past the top 10% of people demographics or our past behavior is the best predictor of future intent. Darn!

During a different panel, Keller Fay and Market Share Partners attempted to quantify the role of social voice - both off- and online in marketing. Some key highlights were:

• Social Voice can be shown to drive nearly as much as organic search activity as marketing
• Social Voice acts as an amplifier of marketing spending, causing an average increase of 10% online search query volume
• They also demonstrated how online/offline brand and product mentions can serve as a lead indicator to sales

Not surprising results, but it got me thinking about how to better include social data in market mix modeling. Keller Fay's approach is survey-based and while monitoring actual behavior is probably better, the lack of historical data often prevents us from doing so. 

I was very excited to hear the next speaker talk about “High Value Facebook Activities – Value of a Fan or a Like,” as this is a question that comes up frequently. As marketing professionals, we often hear questions like:  

• Does liking a brand on Facebook make the consumer more valuable to the brand?
• Are there conditions under which there is value?

Looking at 63 brands rolled up into three categories, Beauty, Food and Retail, they found that those who “Like” a brand on Facebook are nearly eight times more predisposed to it (sounds like a chicken and the egg scenario). Those who “Like” a brand increased brand interactions by 85% vs. baseline 30 days after “Liking.” However, that was only 1/100^th of a percentage point of total brand sessions in a given month. Nearly all increase comes from a very small number of new fans who revisit the brand page after “Liking.”

What they also found was that Newsfeed impressions are of little value to brand performance, but time with the brand on the page does have value. As a result, marketers should do whatever they can to get users to revisit the fan pages and to go to owned media outlets after such visits.  

Bruce McColl, CMO of Mars Foods, gave an incredibly entertaining talk titled, "Inviting Disaster Lessons from the Edge of Technology" during the second session of the day. He used a bridge analogy (Google Tacoma Narrows Bridge) to talk about how while the technology used in construction has changed dramatically in the last five thousand years, the laws of physics have not applied it to today's CMO and called to the audience to help break marketers of the mental models which hold them back.  A few of them are:

• Share my passion: Marketers love brands and think others should to. In Bruce’s mind, this is a mental model that should be challenged. How passionate can consumers really get about candy bars? He cautioned the need to remember the classic principals of reach and not to focus too much on fans or engagement.
• Propel me forward: AKA desire for change: “The re-formula” Bruce talked about how CMO's start their position wanting to change everything and often that ends with updating their resumes. Anyone who has ever had to go into a review when a new CMO comes in can appreciate that.
• Technology seduction: New isn’t always better. Bruce talked about how CMO's tend to focus on the new shiny objects (no kidding). He reminded us that we need to refocus them not on technology for technology's sake, but on how it will drive their business. 
• Social media: Bruce reminded us that mass marketing is critical. He needs 7 Billion consumers to focus on remembering his brand and several million fans won’t move the needle. Hmmmm… have I been drinking too much new media Kool-Aid?

To tackle this and other issues, Mars has created a marketing institute wrestled with applying science to copy development. He didn't share any findings, but outlines these three core areas of focus:

• In copy development, are there elements of execution that repeatedly associate with effectiveness?
• In media, is there a currency exchange for reach across different mediums?
• Are there patterns of amplification and interaction between mediums?

By the way, the “Betty White” spot more than tripled sales!  Wow!

Next up was Media 6 degrees and NYU Stern with, “Evaluating and Optimizing Online Advertising- Forget the Click.” This was a very interesting presentation. They found ad clickers do not resemble future converters and therefore Click Trained models do not predict conversion well. In fact, Click-based models do worse than random targeting. Though not surprising given how low CTR’s are, there are a lot of vendors selling click trained models out there. 

Interestingly they also found that visits to the brand’s website are an excellent proxy for future converters as long as you have enough volume to the site.

Coca-Cola was up next talking about “Moving from Insights to Provocations.” There call was for “big fat fertile briefs” which require data to become the new soil within which our ideas grow.  Once they have these initial provocations, they then lead to a more liquid brief.  They also spoke about using social media not just to listen to consumers but to converse with them. Sounds a lot like Mel Varley’s thought piece, “Are you in Control Enough to Let Go.” 

Following the day, we had great dinner at the University Club. Now, in our world dinner means a time to relax and maybe have a cocktail or two.  At Wharton, you actually get homework and we broke into small groups to summarize the day. 

By 8:30pm, my brain was tired and I called it a day.

Day two began at 8am sharp and, I felt as though I was back in college. We were seated in a lecture hall at the Annenberg center and the theme was “Synergy Effects.” I would characterize it as highly academic. Most of the findings were on small samples and lacked any real-world application that I could see. However, I took copious notes so if you would like to learn more about "statistical insurance against the excesses of variable selection," Give me a ring.

Overall, it was an interesting two days and I came away with a few thoughts to summarize:

• Creative is still king, regardless of the channel
• We must resist the urge to jump on the bandwagon with the new shiny thing
• Mass media is far from dead and I am glad I didn't buy stock in Facebook

If marketing is to be the new finance (Varian 2007), it needs simple and transportable performance metrics, akin to finance’s. 

Lastly, I was reminded how glad I am that I chose not to go into academic research. Working on real-world brands is much more challenging and, the dinners are also more fun.

Theresa LaMontagne
Managing Director, Analytics and Insight, NA