As search engines become more sophisticated, their role with respect to branding and advertising is changing from unbiased summarizer to informed influencer. Put another way, Google isn't merely searching; it's coming alive. Smart marketers will take advantage of this evolution by thinking of the web as a channel and a distinct target customer segment with its own flavor of brand awareness.

Mind of its own
Scientists recognize that when a system becomes very complex, it adopts a kind of mind of its own. For example, a flock of birds appears choreographed even though the group's sinuous motion is governed by simple rules that individual birds follow. In a similar way, the web exhibits behavior that transcends the content and activities on individual sites. On a micro scale, the web is a piece of engineering; on a macro scale, it displays emergent properties that affect brands.

In addition to providing quick access to individual sites, search engines let us watch the "flock." And they can distinguish among brands. For example, Google has a new service called Google Sets (labs.google.com/sets) that returns lists of terms with things in common. If you type in "red," Google Sets returns "red, blue, green, black, yellow and white." (Note the difference between Sets and Search, which returns links to websites). Google Sets is still in beta, but it can be used as a kind of aided awareness test to illustrate how Google "thinks" about brands.

For example, at the time of this writing, here's what Google associated with Subaru: Toyota, Volkswagen, Saab, Suzuki, Volvo, Nissan, Rover, Triumph, Rolls, Porsche, Mitsubishi, Mazda, Pontiac, Honda.

Here's what Google associated with Toyota: Subaru, Volkswagen, Nissan, Volvo, Saab, WILLYS, Suzuki, Honda, Mitsubishi, Mazda.

In this experiment, Subaru is associated with a wide range of 14 brands, while Toyota is associated with a narrower range of 10 brands. These results suggest Google "thinks" of Subaru differently than it does Toyota. The difference might be explained by the fact Toyota has a monolithic brand architecture, while Subaru embraces endorsed and product brand architectures. Whatever the reason, the fact that Google can tell brands apart may influence search results, which could, in turn, affect sales.

Here's another simple prompted-recall test. When you search for "Hondai" on MSN, the search engine responds with, "Were you looking for Honda?" Based on this, it can be said, Honda has stronger "brand recognition" than Hyundai. The order is switched on Yahoo, which responds with, "Do you mean Hyundai?" The Hyundai brand seems to be strongest on Google, which doesn't question the spelling and takes the user to a page filled with Hyundai links.

'Smarter' search engines
While search engines may have rudimentary minds of their own, it's well-known that what they "think" is influenced by sponsorship and other forms of search-engine optimization (in the example above, both Google and Yahoo had Hyundai-sponsored links, while MSN did not). Monetary biases notwithstanding, as search engines get "smarter," they will be harder to influence using conventional approaches. For example, changing a site's content and code (on-page factors) is less effective than being linked to by heavily trafficked sites (off-page factors). In fact, it's becoming increasingly difficult to distinguish between "off-page search-engine optimization" and traditional PR.

Looking ahead, it will be important for agencies and marketers to think beyond conventional search-engine optimization and devise ways to "shape computer behavior" instead. For example, in another Google Sets experiment, I typed in "mountain biking." The set it produced included what I would have expected -- hiking, BMX, cycling -- but it also gave me "horseback riding." What would this mean to a marketer? If lifestyle-brand managers were aspiring to shape how computers think about their brands with respect to mountain bikers, they might consider endorsing equestrian product lines. It seems counterintuitive, but Google associates mountain biking with riding horses.

Artificial brand intelligence is important because it has the potential to affect brands, consumer behavior and sales. Today, computer programs subtly guide customers; in the future they will be customers. As computers learn more about brands, they will make some purchasing decisions with less human participation. Furthermore, as the web reaches beyond personal computers, the impact on consumer behavior will take surprising turns.

Informed influencer
For example, computer programs might someday plan travel routes based partly on brand preferences. Today GPS manufacturers such as Garmin are experimenting with real-time advertising. In the future, software running on internet-connected GPS navigation systems such as the Nokia N800 might take this a step further and automatically plan routes based on the promise of retail brands along the way.

Here's an example: I'm traveling to a friend's barbecue and need to pick up chips and dip along the way, so I turn to my GPS navigation system to guide the way. I can't be sure which market it will take me to; I'm leaving that decision up the device. Eventually, artificial brand preferences might join factors such as traffic and road conditions in calculating travel directions (see figure).

The point is this: From search engines to car navigation systems, there is evidence that services and devices connected to the web have, or will soon have, rudimentary brand preferences. Many computer programs underpinning the web are developing an awareness of brand images and experiences, and their purpose is evolving from unbiased summarizer to informed influencer. The marketers and agencies that account for this newfound artificial brand intelligence have it right: They will be recognizing a growing, increasingly influential market segment, and they'll have a jump-start on developing sophisticated new marketing strategies and opening up profitable new opportunities.

(Source: Advertising age)