The advertising world is ever-evolving. No longer are advertising campaigns hatched from creative meetings or three martini lunches, as depicted on Mad Men.  Now more than ever, agencies are looking to data geeks to uncover the facts that generate the most strategic and effective campaigns. At the core of insightful recommendations are the correlations those geeks uncover.

Correlations define intrinsic relationships between media tactics and business successes, have emerged as the central focus for impactful decision-making.

So what are correlations and why are they critical to the advertising world?

Correlation is a statistical term that defines the relationship between two variables. The most common correlations that analysts conduct in the advertising sphere are multivariate correlations, which focus on comparing one variable against multiple variables. One of the most successful advertising campaigns for AT&T was the “It’s Not Complicated” campaign, which used data-driven creative to generate an additional $50 million dollars in sales. By analyzing over 40 variables and testing over 370 ads, AT&T was able to quantify and define everything from the characters to interactions to humor. Through defining variables that resonated most strongly with AT&T’s audience, “It’s Not Complicated” was able to drive higher audience reach and engagement.

Why correlation modeling?

Correlation modeling enables us to gain insights that shape more strategic creative and help us make impactful decisions on recommended advertising tactics. There are three core reasons why multivariate correlation modeling is so important.

  1. It is the driving-force in advertising decisions. When regressing a multivariate model on a consistent basis, analysts are able to provide the media strategist and clients with actionable insights regarding which tactics work and which do not. Brands have no prescription for the most effective advertising campaigns—every product and geographical market is unique in its target audience and consumption behavior. By running correlations on a consistent basis, ad agencies are able to optimize tactics to create more conversions for the client.
  2. Multivariate correlations are extremely cost effective and end up saving clients money. By making key optimizations on specific tactics that are derived from regressions, clients are maximizing their brand in the market, creating a higher likelihood of attracting new business. When a client invests in a media plan, correlations provide insights that enable the agency to omit tactics that are not driving awareness, engagement, or conversions. This will eliminate any advertising waste that does not benefit the client.
  3. ​Multivariate regressions also enable much more effective reach and engagement. For example, in the “It’s Not Complicated” campaign, AT&T was able to define key variables that resonated most strongly with consumers in their messaging. This drives wider consumer engagement and more effective marketing for the defined target audience.

 What tools are available to conduct correlation modeling?

There is an abundant amount of statistical tools available to conduct multivariate regression modeling. Most recently, MMI Agency partnered with KNIME, a free, open source data analytics, reporting and integration platform. Within KNIME, analysts are able to either connect to other data sources or pull in Excel spreadsheets in order to use its Linear Regression Learner and Regression Predictor. KNIME’s core benefit is not just its ability to easily conduct regression modeling, it can also visualize the data in a digestible way to marketers. This strong visualization component enables advertising and marketing agencies to effectively communicate strategic insights driven from myriad data points.

MMI uses Knime for a variety of client engagements. One example is Amazon optimization, where we pull data for clients, such as Boogie Wipes, as well as their competitors to gain valuable insights that influence our Amazon strategy.

So the next time you sit down with a creative team, engage the quantitative researchers and data geeks. Remember that the most memorable campaigns are based on critical thinking and data-driven creative, which start with correlation-driven insights.

 Kate Nazareth Kate Nazareth
Quantitative Research Analyst



Leave a reply

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>