The death march for traditional advertising cut a surprising path through the offices of the Boy Scouts of America.
It was December 24 in Palo Alto, California, and Google's chief financial officer-a Canadian transplant named Patrick Pichette-was driving around his new hometown looking for a Christmas tree. Unsure where to even start his search, he picked up his smartphone and spoke the words: "Christmas tree, Palo Alto." Using algorithmic wizardry developed by engineers just down the road at Google's Mountain View campus, the phone parsed his words, converting them to a search request. Almost instantly, the phone returned a result based on Pichette's location: The Boy Scouts, whose Palo Alto office was just minutes away, was selling Christmas trees; the phone even provided driving directions.
"Eighty bucks later, there you go," Pichette says, in the manner of a child wowed by a brand new toy. "Imagine the hours I would have spent driving around looking for a tree."
Pichette likes telling this anecdote because it suggests just how disruptive Google's technology can be for advertisers. Rather than blanketing an area with billboards or buying time on local radio stations, the Boy Scouts could rely on Google's search results to place its listing squarely in front of a very motivated group of buyers. The Scouts were lucky: Their tree lot came up as the top result. But even if it hadn't, the troop could have simply paid for a sponsored link to be delivered to anyone searching for a Christmas tree in the vicinity. Moreover, Google would then be able to collect meticulous data about anyone who clicked on the link, where they were when they encountered it, and what keywords or images caught their attention. The whole process is ruthlessly analytical, and promises to eliminate much of the guesswork that usually pervades other advertising channels.
Advertising is one of the few industries where it's virtually impossible to know if money spent is money well-spent. Is a Super Bowl commercial really worth $3 million? Is Tiffany & Co. a more alluring brand if its ad appears in The New York Times instead of USA Today? And what makes a good ad, anyway? These and other questions have plagued marketers for decades as they seek to refine the murky art of persuasion. Now, platoons of Google engineers are threatening to make such questions irrelevant by reframing the entire scenario: Let the consumers tell advertisers what they want.
But how, exactly, do you find out what the consumer wants? Why, you ask them. In a way, this is the task that Google was born to do. Every time a user sends a search request through one of the company's servers, they are not only telling Google what they want, but where, when and how they're looking for it. The company tracks and catalogues every aspect of the user's request, which adds up to a lot of information. Every three days or so, Google processes as much as 70 petabytes* of data-more than the entire written works of human history. For advertisers, the key to knowing whether an ad will be successful or not is buried in that data. This digital treasure trove may soon render many traditional methods of measurement-from focus groups to Nielsen ratings-hopelessly obsolete.
Think of the ad as it has existed for decades: Company A has a product it believes will appeal to a certain group of people, so it hires Company B, a creative agency, to help raise awareness of its widget. Armed with as much information as it can collect, the agency produces a campaign it hopes will appeal to the product's ideal consumer. A media buyer arranges to have the campaign run in key markets in media channels-newspapers, magazines or during TV shows-that are known to be popular with the product's target audience.
A number of ratings and measurement agencies have traditionally provided this last piece of the puzzle. Newspaper ad rates, for instance, are based largely on circulation figures-compiled in Canada by the non-profit Print Measurement Bureau-
as well as by demographic information gleaned directly from readers in the form of surveys (for example, the number of people who own second cars, have college degrees, buy organic produce). In the world of television, Nielsen ratings, which track the viewing habits of a selected group of people who have agreed to install one of the company's boxes, have been the industry benchmark for decades.
Both models deliver only a fuzzy outline of a potential consumer. More complete data, including how much time people actually spend looking at ads, where they are when they encounter them and how often they might mention them to friends, is largely lacking. That's where a number of Web analytic companies, of which Google is undeniably the king, have stepped in.
Every time you visit a website, the operators of that website know you're there. They know (roughly) where you are, how long you spent on the site, what content caught your attention and what didn't. Depending on how much information you volunteer, they may even know a little bit about you, such as your gender or age. What Google's tools offer is the ability to measure viewing habits in real time, and provide clients with a detailed analysis of what visitors are doing. The amount of information that gets mined is so vast that some of Google's newest tools are designed to present less data than its software actually collects, in order to avoid overwhelming users.
Data collection has become the chief driver of advertising in the digital age. But how marketers can interpret and use that data is still being explored.
During the last presidential election, Barack Obama's campaign team spent an enormous amount of time sifting through various photos of the Senator-Obama smiling, Obama looking dignified, Obama with his wife and kids, Obama looking decisive-and scratching their heads. The problem: Which of the photos would resonate most with visitors to the campaign website?
Before the era of Web analytics, there would have been a simple way to solve this kind of problem. Avinash Kaushik, Google's in-house analytics guru, calls it the "HIPPO" approach: that is, the Highest-Paid Person's Opinion prevails.
But in this case, the Senator's campaign team chose to let the data decide. Among the many tools the campaign team used on its website was one called Google Website Optimizer. Simply put, Optimizer records everything a user does while visiting a website-what they click on, how long they stay on any given page, where they lose interest and leave-and gives a webmaster feedback on what is working and what isn't.
Using the near-instantaneous feedback, the Obama team tested the various photos and elements on the site. They even added videos to see if multimedia content would encourage visitors to stick around. Their metric for measuring the success of each element? The percentage of people who signed up to receive more information from the campaign team.
The results were swift and backed up by irrefutable data: All the videos were flops, the image of Obama with his family worked best and people responded favourably to a "learn more" button. The staff altered the website accordingly and, shortly thereafter, saw a 40% bump in Web-based sign-ups. "The single best thing you get from the Web is the ability to be proven wrong fast," Kaushik says.
For advertisers, the difference between a digital campaign like Obama's and a traditional one is clear. Run a newspaper or TV ad, and you may not know for a long time that the campaign material's green background turned viewers off. Online, however, you can watch users veer away in real time, tinker with the ad a little, and then check again. The flow of data is constant and conclusive. In this light, it's easy to see why an increasing number of companies are devoting larger portions of their marketing budgets to the digital realm. According to an Econsultancy report released in February, about two-thirds of companies surveyed said they intended to increase budgets for online campaigns; about 40% said they would decrease their print media budgets. The giants of old media have every reason to be concerned.
In his book Googled: The End of the World as We Know It, Ken Auletta, a media critic at The New Yorker, recounts a 2003 meeting between Google's founders and Mel Karmazin, who was then the president of media giant Viacom. When told of Google's plans to take the guesswork out of advertising by giving marketers detailed statistics on which ads work and which don't, Karmazin complained: "You're fucking with the magic."
When Google executives talk about the future of advertising, this is what they imagine: A pizza parlour that wants to run a two-for-one special directed at males, 20 to 25 years old, who happen to be watching a particular game on TV. Google foresees a day when that pizza joint may pay to have its commercial run-on television-for the duration of the game, and only in living rooms within the restaurant's delivery radius.
Such a scenario is actually not all that far removed from reality. Whereas Web, print and television advertising used to live in separate kingdoms, the lines have already begun to blur. Many of the higher-end TVs lining the shelves at electronic stores now come with some sort of Internet connectivity. Pair that with the rise of mobile computing-recently bolstered by the release of Apple's iPad-and there is compelling evidence that digital screens hold the future of magazine, newspaper and television content.
It can only be a matter of time before our viewing and reading habits will be as open to third-party analysis as our daily Web wanderings. If that's the case, the days of the passive, non-interactive ad are numbered. And for an example of how rapid the axe's fall can be, marketers need only look at how quickly classified newspaper ads were replaced by Craigslist postings less than a decade ago.
This time, however, traditional media is less likely to be caught off-guard. Many marketing departments are already committing more resources to the creation of online communities, comments sections, mobile apps and other online tools that facilitate greater interaction between marketers and consumers. It's a smart move, says Kaushik. "Marketing is becoming less about shouting at people and more about participating in conversations and influencing them through that," he says.
For some, this might sound a little subversive. Kaushik's boss, vice-president of product management Susan Wojcicki, has a more delicate way of phrasing it: "We want ads to be useful."