Digital media and traditional media are at a crossroads -- and when it comes to research and measurement, marketers are having difficulty knowing which way to turn. Since the dawn of the TV-era, countless dollars have been spent developing data-capture and analytics tools. But, the TV-centric model of measurement is obsolete. Media channel crossover is becoming more complex: television ads are linked to product Web sites, product Web sites lead to social media sites, and social media sites connect users with other users who are talking to each other about the product. Consumers are engaged with multiple channels simultaneously.
The old marketing and research models that assume media channels are independent of one another are simply no longer working. Without good models and tools, marketers risk making bad allocation decisions potentially sending millions of dollars into less productive channels.
As more digital channels develop, marketers must be able to account for crossover, interaction, mixing and the overall complexity of media in the digital age. Research must also take into account the influence peers have on each other and the marketing ramifications of social influence. This is easier said than done, of course. But let's take a look at how we can bring media mix models and measurement into the digital and social media age.
Traditional media mix models tend to fail when digital is factored in. In fact, at many companies digital still isn't part of their core models. When digital is included, inputs simply don't match outputs. For example, a CPG company spending $50 million on TV and $2 million on digital advertising has a model that says digital is responsible for 40% of sales with TV responsible for only 3%. These results for digital are " too good to be true." Digital advertising alone is clearly not responsible for driving 40% of sales when its budget is barely a blip compared to TV. Conversely, there are models where results are " too bad to be true." So, why the variable? The traditional model treats each medium as independent, which doesn't work when digital crossover blurs the media lines.
The root of the problem is that consumers are receiving messages through numerous media touch-points but research models aren't accounting for this, and are only giving conversion credit to one medium. That is, an ad on TV may drive a person to that product's Web site or to search for that product online. Once he does a search or lands on the site, he is more likely to be targeted with digital display ads for that product or related products because the way he interacts with media determines how advertisers reach him.
This kind of crossover creates a real headache for marketers trying to measure the effectiveness of each media channel. Take the Super Bowl ads for example. Some advertisers look at the increase in number of online searches to show that a TV spot was effective. Here, search behavior is the output of the model. So, which is it? Is digital media one of the inputs (as a paid media channel), one of the outputs (as a response to other media) or both?
To bring old research and measurement models into the digital era, they are going to need a major makeover--but what does this new model look like? How do the various digital channels interact with each other, with traditional media channels, and with consumers?
Moreover, a huge part of digital media is now social, where consumers are taking social influence online, which creates a whole new layer for marketers.
A Model Based on Social Influence
One key element that the old methods of research do not take into account is consumer influence. That is, how consumers are affecting each other -- and through what channels. To fix this issue, it is important that modelers take into account a person's " social graph" as part of the model of measurement in order to be relevant in today's more social society. A social graph represents a person's social network -- those whom they influence, and those who they are influenced by (ranging from personal contacts to product reviewers online).
To bring the models up to speed, marketers must take into consideration that consumers are influencing each other in totally new and different ways -- through digital -- and these interactions are not only causing flaws in media mix measurement, but also flaws in conclusions about consumer behavior. Modelers must develop new models that incorporate all causal relationships of digital media. And after delivery of the model, it should be standard procedure to validate the model.
The Measure of a Plan
In order to better understand consumer behavior, researchers must account for a consumer's social graph and look at both digital and non-digital social networks.
Agencies must continue to innovate to develop more powerful measurement tools, whether that is a tag that can track a single consumer's behavior across channels or a holistic solution that displays all marketing in one place, along with easy-to-understand graphics and actionable insights.
This will allow for a better understanding of where your consumers fit into your marketing programs. The more research that agencies do to better understand how consumers interact with each other and with media, the better they will be at incorporating digital into the media mix model.
This originally appeared in Dave's monthly column in Chief Marketer.