Wednesday, March 20

Measuring Effectiveness: Coke Says Buzz Is Not Enough

Plenty of people have said it before, but Coca-Cola had invested hard dollars to prove it. Online buzz is not enough to have a measurable impact on short-term sales. Online display advertising works better.

While the concept seems to contradict what social media enthusiasts tend to tout, it's one of several studies that not only raise questions about the growing interest in online influence but also refute it. After all, if buzz doesn't drive short-term sales where display advertisements might, then what about influence?

Smart companies don't make decisions based on single studies.

Of course, according to the Adweek article, Coca-Cola isn't ready to toss out the baby with the bath water. Its digital media team points out that the findings were based on one study with one segment of one company that appeals to a particular customer. In this case, one with 61.5 million Facebook fans.

Instead, Coca-Cola will continue to look for ways to measure online buzz and other popular social media counts such as video views and social sharing. The company, one of the early entrants into digital media, wants to find a predictive measure that can pinpoint financial outcomes — at least so marketers may better understand the tradeoffs among media types.

Personally, I'm not sure they have to or that it is even possible. After all, no influence expert to date has considered that obvious problem with attempting to measure online influence. Much like social media, influence does not exist in a vacuum.

The psychology of a decision-making process is bigger than one trigger.

Most people aren't subjected to a singular influence that leads to a single action like influence scoring systems such as Klout suggest. On the contrary, most people are subjected to dozens or even hundreds of complementing and contrary influencers — ideas, people, personal experiences, and other variables — before they complete a multi-decision process that leads to a purchase.

Ergo, if a friend or "influencer" recommends a book, the suggestion is subject to all sorts of variables, ranging from what other friends and influencers have said, the largely random collection of reviews that might be found prior to purchase, any number of customer reviews at the point of sale, the propensity that they have made past book recommendations, whether or not you know the author, how far off it is from your preferred genre, whether or not you are reading a different book, etc., etc.

It seems amazingly silly to credit someone who might not even be the initial influencer with what is an infinitely grand process. But that is partly why all marketing is part art and science, not just science. The human brain doesn't merely string together 0s and 1s like a computer program. It's much more complex.

Keep watching Coca-Cola anyway. It finds out all sorts of interesting things.

While attempting to isolate the impact of social media, especially buzz, might not be possible given the magnitude of the variables, Coca-Cola is discovering some interesting data related to online advertising. According to the article, the company found online display ads could be considered about 90 percent as effective as television while search advertising is only 50 percent as effective as television.

Given the continued changes that Google is making to its search algorithm, this might surprise some SEO proponents because it runs contrary to what people like to think, unless we're talking about mobile. But the truth, I think, is even more surprising. One-off SEO has undermined search as some specialists attempted to divert unrelated searches to gain traffic. People trust display ads more than search because the ad is what it is — the search suggestions might not be what they want.

But it raises another wrinkle in most studies out there. The best marketing strategists — digital or otherwise — know that like all advertising, the operative word is sometimes no matter how hard naysayers try to steer people away from one or all facets of distribution. And that "sometimes" depends not only on the distribution method but also the timing, message, presentation, product, organization, audience, and a half dozen other things that most people don't know to measure.
blog comments powered by Disqus
 

Blog Archive

Google+ Followers

by Rich Becker Copyright © 2010 Designed by Bie Blogger Template