Testing Everything

In his novel “Pattern Recognition,” William Gibson creates protagonist Cayce Pollard, a professional “coolhunter” who suffers extreme psychological allergic reactions to well-known brands and logos. In essence, Cayce is capable of seeing the latest trends before they happen but has become so sensitized to the art and science of advertising and branding that she can’t stomach the proliferation of logos (her strongest reaction is to Bibendum, the Michelin Man). In the modern world she has to go out of her way to avoid the marketing we work to place everywhere.

Shepard Fairey, creator of Obey Giant, describes his campaign as an “experiment in phenomenology.” Although you may not be familiar with the campaign, you’ve most likely seen the sticker with the picture of Andre the Giant and the word “obey” somewhere in your travels. Fairey has subverted the ubiquity of advertising and branding, creating a campaign with no product to sell and no cause to promote. Instead, it simply exists to call attention to the act of recognizing the world around us and the messages delivered.

Gibson and Fairey tap into the dark side of our marketing-driven culture and call our attention to the dangers that lurk within. As marketers, we often fall into the trap of “more is better.” If we can show the logo 20 times, why not 100 times? If we can get customers to add our banner to their blogs, why not make the logo on the banner just a bit bigger?

Marketers are trained to believe in the power of repetition, exposure, and constant brand reinforcement. The Web allows us to replicate our message infinitely and to ensure we lead customers through a highly planned and tested brand experience. If on- and offline brands are out of alignment, we see immediate impact on customer behavior. Consider Apple’s on- and offline stores. Why is the physical store so much better at portraying the brand than the online version? Could Apple do a better job, or is this a weakness for one of its competitors to exploit?

Modern online marketing allows us to plan, measure, adjust, and optimize in near real time, 24-7. We can make our own brand nearly ubiquitous if we invest enough time and effort. But do any of the data tell us when we’ve gone too far? Do we know at which point we might be causing an “allergic” reaction to our messages in our target customers?

Here are a few things to consider when thinking about this problem:

  • Has your Web site become a giant logo farm? How many times must your corporate or product logo appear on a page? Should it appear on every page? If you have a globally recognized brand, you likely only need it to appear once. If you have a lesser-known brand, you should focus on creating value behind the brand, not pasting the logo everywhere.
  • Do you run viral and community campaigns that have become brand parade floats? Do you focus on the real content and value of what you deliver when you enter the online community, or are you trying to get your logo to appear on some kid’s MySpace page? Does having it appear on that MySpace page really sell your stuff and drive ROI?
  • Are your off- and online marketing activities aligned sufficiently so the brand experience is consistent from one to the other? Are you in danger of creating brand allergies in potential customers through your overexposure or poor execution? There’s nothing worse than seeing a picture of a hotel room online, then walking into the real room and feeling duped. Ensure every brand touch delivers both value and truth.
  • Have you been testing and optimizing brand messages and visual treatments as well as product content and offers? It’s pretty easy to run a massive multivariate optimization test of discount offers, but what if you simply test the presence or absence (or even size) of your logo when placed next to the offer? Consider the impact you might be having on your customers when you put the big “LOGO” sign in front of them.
  • Online marketing has given us an amazing set of tools and the ability to test everything. The trick for the good marketer is being sure to define “everything” as everything.

    Related reading