Consumer Centric

What question do you suppose would leave a panel of leading experts in media measurement stumped? During an otherwise informative and interesting panel discussion at the opening session at yesterday’s Audience Measurement 4.0 conference (put on by the ARF), such a question was asked that resulted in dead silence.

Irfan Kamal of Ogilvy PR Worldwide was the first to capture on Twitter the noticeable lack of response from the panelists to a question about planning for social media within media mix discussions.

The keynote panel was titled Forecast 2015: Social Media & Their Impact on Audience Measurement.  The discussion was good and the participants were solid:  John Burbank – CEO, Nielsen Online, The Nielsen Company; Chris Cunningham – Founder and CEO, appssavvy, Inc.; Dean DeBiase – CEO, TNS Media; Joe Doran – Founder and CEO, Media6°; and David Smith – CEO, Mediasmith, Inc.

None could answer, however, the audience generated question about “how do I fold in social media into mix decisions?”  Burbank eventually broke the silence by giving a basic “it depends” answer along with “the tools will catch up.”

It’s a tough question.  How do you determine the most effective and impactful spend levels for social media in your overall media plan?  Marketing mix models and econometrics can provide a rear-view mirror analysis to guide budget and mix decisions.  But what if you want to try something you’ve never before done and, therefore, you don’t have any past data, but you still need help anticipating what might happen and how much you should spend on the effort?  These tools alone can’t effectively address such an need.

Enter agent-based simulation techniques.  

Companies like Icosystem, ThinkVine, and Decision Power, among others, are findings ways to create a “virtual test market” from the individual person up, rather than only using the aggregate/total data across the system.  More will come along as they see the benefits of using these tools in tandem with marketing mix and econometric models.

Basically, this technique defines rules of engagement for millions of “agents” (think people as well as media, messages, and other influencers in the real game of life) within a specific environment.  Rules mean how “agents” behave and interact with other agents, etc.  

Once all of the agents are programmed, computers run thousands of simulations to tease out and identify emergent behavior (something that hadn’t been there before and couldn’t be seen without the modeling).  This forward-looking simulation allows you to play what-if scenarios for marketing activities you’ve never before tried.  How does that work?  Because you build the simulation from the bottom-up, you’re modeling individual human interactions and not just looking at the historical relationship between aggregate data like how much you spent and how many people bought your stuff while you spent that.

At our company, PURSUIT, we apply Icosystem’s Concentric platform to our client needs for planning, optimizing, and reporting ROI on all channels, including new and social media. I love it because it is truly people centric in its very design and execution. You have to define how people behave as individuals, then let them interact with each other to see what new behaviors emerge. 

Don’t get me wrong, econometrics and marketing mix models have a solid role to play in analytics and campaign resource allocation.  They’re ideal for incremental improvements in efficiency and effectiveness (how do I get more out of my mix this quarter as compared to last, etc.).  

But, answers to the question posed yesterday will be more fully addressed when marketers and advertisers embrace the application of simulation tools that mimic individual people and how they interact within and across the worlds in which they live.  These tools can directly address the questions of “how do I fold social media into my mix” and/or “what will happen if I try this?”