The New E-Commerce World, Seen Through A Google Lense
Written by Evan SchumanThe typical American business executive likes incremental changes. An 18 percent speed increase, a little more efficiency over here, some additional functionality over there.
But what scares them is anything huge. Despite the fondness for slides saying paradigm shift, it’s the last one thing most executives honestly want. Anything that monumental shifts power. To the typical executive with a large company, any monumental shift raises the frightening possibility that they’ll end up a lot smaller than they are today.
That’s why it’s typically the startups and small niche players that embrace big picture change. If only they could get anyone to listen to them. The Internet was one such change as was its subsidiary, the World Wide Web.
But what happens when one of those startups riding that major change gets big? Hence, the challenge for the $11 billion search giant Google. At a recent media day at their New York City offices, Google executives and managers discussed the enormity of the E-Commerce conundrum today and how they are struggling with finding the best way to deal with it.
Consider: In the early days of E-Commerce, a typical retailer and manufacturer might have between eight and 12 products that it is actually marketing, especially with search engine links. Today, that same company would typically be doing the same kind of Web marketing with some 12,000 products, said Tim Armstrong, who serves Google as its president of advertising and commerce for North America.
This new environment, Armstrong and other Google people said, requires a very different approach. Google’s answer is something called an Asset Map and it’s a way to visually lay out every single one of a retailer’s assets, including all products and services. This theoretically allows a company to see not only which of their products are not covered, but to try and project some kind of return-on-investment analysis.
Armstrong argues that this is morphing ad budgets into operational budgets. Is a Google ad akin to a traditional piece of advertising?something that an ad budget should fund?or closer to the cost of a car dealer building a new showroom and dealership?
Google is making the argument that their auction-driven pricing model is more than a marketing cost. They argue that it’s actually a tool to help match inventory and purchase patterns with inventory. That’s because, they argue, the variable pricing allows budgets to fluctuate with consumer interest.
In theory, that should allow better realtime information about demand, in a much more predictive way than simply examining purchases. In other words, if the number of times consumers look at ads for SUVs fluctuates in the same way during different months (or different weather patterns), that can help influence core business purchase decisions.
Another core change for E-Commerce is the explosion of social networking and video sites?primarily launched for a younger audience?including Facebook, Myspace and Youtube (now owned by Google). Those sites create the potential for customized focused campaigns in a way that simply didn’t exist a decade ago.
Armstrong said the social networking sites caught him offguard?”the traffic is really incredible”?because he didn’t initially expect search to be a factor. He saw videos as something people would merely browse through. The high demand for video searches was “a very nice surprise.”
Then there’s the mobile movement, which places limits on ads (minimal screensize, less RAM and much slower bandwidth) but also opens up possibilities by being with a consumer at all times and including very precise location information.
The mobile potential: “Every phone gets an individual ad, built on the fly,” Armstrong said.
This all combines at a time when consumers and businesspeople are becoming much more comfortable with Web searches. “People are getting much more sophisticated about how they search,” Armstrong said, pointing, for example, to significantly longer average search queries.
The simple process of guessing what search terms people will use is also getting much less simple. The term “ipod,” for example, is searched for using any of 7,000 words.
The biggest challenge in the near term for Google, though, are privacy concerns. Google’s current policy, for example, forbids the use of third-party cookies, said John McAteer, Google’s head of retail. That’s an unusual stance to take for a company that is typically seen as pushing the privacy envelope.
In search engine ad placement, there are two traditional ways to sell ads. The first is contextual, where ads look at the page where it’s to appear, and behavioral, where it looks at who the site visitor is and potentially that surfer’s recent traffic.
Google’s self-imposed ban on third-party cookies would presumably take it out of the behavioral game, but non-cookie techniques are being explored. For that matter, all policies can be reviewed and amended.
In theory, an advertiser should find behavioral ad placement more effective. After all, the reason the contextual ad is placed on a certain page is because of the kind of people likely to read such a page. A behavioral search could be more precise on delivering to those prospects, regardless of what pages they felt like visiting.
But even behavioral cookies are not perfect, as people can accidentally hit irrelevant pages or perhaps are looking to buy a onetime gift for someone with very different tastes. The only ideal approach is to use both techniques simultaneously.
Although the mobile phone offers the most specific geographic targeting potential, Google is experimenting with IP addresses and search terms to try and deliver very geographically localized ads. Using IP addresses to identify location is about 85 percent accurate today, said Brett Goffin, a Google industry sales manager in retail.
Many things account for the 15 percent that is inaccurate, such as broadband ISPs whose locations appear instead of the address of the person surfing. They’re often close, but not precise. Certain domains?such as AOL.com?are notorious for giving little hint as to a member’s location. Virtual Private Networks (VPNs) can also wreck havoc with IP address location accuracy.
There are still some sites that strongly suggest location, such TV station web sites, weather sites and the wording of the searches themselves (“furnishings in Austin”).
The amount of time people spend on the Web today is roughly equivalent to how long they watch television, McAteer said. Mixed in with all of that Web traffic is some eight billion searches a day, he said.
There is little doubt that the biggest potential impact on Web advertising will be merged channel retailing. McAteer argues that retailers have only begun to see the evidence that online ads can impact brick-and-mortar sales.
The explicit example are sites that allow coupons to be printed and then brought into the stores. The retailer can track how many times those ads are printed (showing presumed interest) and how many times they are redeemed.