At a press event in San Francisco this week, Google
Google claims the new feature is designed to make Web searches more efficient, by offering interactive feedback for partial queries -- "faster than the speed of type," as the company puts it. If you really were searching for Sausalito or sauvignon blanc or sausage, after all, you'd be done.
But I have my suspicions. I've been testing the new feature and talking to people in the search-engine-optimization business -- folks who spend their whole day advising Web publishers and e-retailers on how people react to search result rankings, and how to elicit free traffic from Google and other search engines. They have some interesting thoughts about the probable repercussions from this latest change in the way Google presents search results.
One of the outcomes -- and I think it's likely -- could be that you'll actually spend more time using Google than before. Another could be a huge increase in competition among users of Google's AdWords search-advertising platform for common keywords -- leading inevitably to more revenue for Google, but not necessarily to more traffic for the majority of advertisers.
As a caveat, I should say that none of these effects is likely to set in right away. For one thing, the Instant feature is being rolled out gradually, so it's not available in all regions yet. It doesn't work for searches from mobile devices. And most importantly, it doesn't work for searches from the browser toolbar, where more and more people enter search terms. In my own case, I never actually go to Google.com, because Google lets me initiate searches directly from the Chrome "Omnibox," the same place where I type in URLs. People who use the search boxes on Firefox or other browsers are in the same boat. Google would have to rebuild Chrome or come up with Instant plug-ins for Firefox and other browsers to make Instant search available from these locations. (Google's Marissa Mayer said at the Google Instant launch event that the feature will be coming to browser toolbars "in the next few months.")
So it's too early to get all worked up about Google Instant. But it is fun to speculate about the forces at work here, and which way they might be pushing. In this business, the stakes are so large -- to wit, Google's 70%-plus share of the search market, and its $23 billion in advertising revenue in 2009 -- that nothing happens by accident, and even seemingly small changes in the way the search engine works can have outsized effects.
1. The Wikipedia-Facebook theory
There's a great edition of the geek webcomic XKCD called "The Problem With Wikipedia." It shows a diagram starting with the query "Tacoma Narrows Bridge." This search leads to articles on "Suspension Bridge" and "Structural Collapse," which, after "Three Hours of Fascinated Clicking," leads -- naturally -- to "William Howard Taft," "Lesbianism in Erotica," and "Batman."
The point being that the Web is a magical garden of enticing distractions. Google has said that when Google Instant is switched on, each query results in the delivery of five to seven times as many result pages. (On their way from "S" to "sausage," for example, San Franciscans will see at least three pages of preliminary results, for Skype, Safeway, and Sausalito.) That means there are many, many more opportunities to stumble across something you weren't really looking for but that might interest you anyway.
"They are making the bet that users are going to be interested in having their question change on them in mid-typing," says Ray Grieselhuber, the founder of GinzaMetrics, a Mountain View, Calif.-based startup developing a real-time SEO platform. "It may speed up each individual search, but it may cause you to spend more time searching in general. Which is great for Google. ... When you look at it in that light, it makes a lot of sense."
Why would Google want you to distract you with stuff you didn't know you wanted to know? Lots of reasons. The company might, for example, be looking enviously at Facebook's stunningly high time-on-site statistics -- 32 minutes per visit, according to Alexa, compared with just 12.8 minutes for Google. (Update: comScore reports that Web users spent 41.1 million minutes on Facebook in August, which was more than the 39.8 million minutes they spent on all of Google's sites combined.) It wouldn't be surprising if the company is working on various ways to make Google.com more interesting as a destination site -- a place you might want to spend a little time exploring before you click off to the next page.
And don't forget that Google is run by computer scientists who get a kick out of anticipating people's information needs. In a now-famous (and oft-misinterpreted) remark to The Wall Street Journal, Google CEO Eric Schmidt said, "I actually think most people don't want Google to answer their questions. They want Google to tell them what they should be doing next." I don't think Schmidt meant to come off as Big Brotherish; he was just saying that Google ought to supply information tailored to whatever users are doing, at the instant they're doing it. Matt Cutts, head of the team at Google that's responsible for detecting spam sites, made the point this way in a blog post about Google Instant on Wednesday: "Peoples' information need[s] often change over the course of a search session. Google Instant makes that process even easier: People can dig into a topic and find out new areas to explore with very little work." The more new areas they find, of course, the longer they'll spend on Google -- and the more likely they'll eventually click on an ad, earning Google a pay-per-click fee.
Let me be clear: I'm not suggesting that Google Instant is simply designed to draw out the search process. For certain kinds of highly goal-driven searches, such as product searches, it may be faster than the old method of typing out your whole query, and then waiting to see what comes back. But if you're open to suggestions -- well, you're about to get a lot more of them.
"Search intent is a big part of the equation in whether this is really helping efficiency, or increasing the time that you're searching," says Jim Yu, CEO of BrightEdge, a San Mateo, Calif.-based SEO startup I profiled in June. "If I'm looking for size 12 Air Jordans, this gets me much closer to buying something faster. If I'm in research mode and just trying to find information about a topic, it's going to get me a broader set of topics."
2. The AdWords theory
Even before Google's press conference was over, the blogosphere was buzzing over the question "Is SEO dead?" In its white-hat version, anyway, SEO is the process of adjusting the content of your website so that people can easily find it through search engines, using the keywords that would come to their minds most naturally. The answer most experts are giving to the question is that SEO is definitely not dead, because Google Instant doesn't change the way results are ranked -- it just lets users flip through more pages of them. So it's as important as ever to make sure that your website is ranked highly for common keywords and combinations of keywords.
But closely related to SEO is search-engine marketing, or SEM, the art of winning traffic through the keyword-based text ads that appear alongside the unpaid or "organic" search results. In the Google universe, this means knowing how to employ Google's AdWords platform, which hands the top spot in a stack of text ads to the highest bidder in an electronic auction for the given keyword.
Google Instant won't kill SEM any more than it will kill SEO. But it could throw search-engine marketing into considerable chaos. Here's why: People using Google Instant are likely to follow one of Google's automated suggestions before they finish typing their full query, even if Google's guesses don't fit their original question exactly. As Yu puts it, "Google Instant is going to drive more volume to the keywords that are being auto-suggested." (Studies suggest, in fact, that 60 to 70 percent of the time, searchers settle for one of the suggested keywords.) Which means, in turn, that there's going to be a lot of competition for those keywords -- and in particular for the common "short head" (as opposed to long tail) keywords and keyword combinations.
To use a simplified example: If the first page of results that pops up for the partial query "c-a-r" is about "car insurance," then car-insurance companies will have a bigger incentive to bid for that keyword combination than before. Which means there will be more bids. Which means it will cost more to win. Which means Google will make even more money. (It's a little more complicated than this in reality, of course, since results shift depending on the user's location and history.)
"Google is targeting to optimize its revenues," says Horst Joepen, CEO of Searchmetrics, an SEO consulting firm based in Berlin. "AdWords users will either have to pay higher prices to participate in a short head race, or live in a cheaper but less frequently visited long tail."
So, longer time-on-site and higher AdWords revenues -- two pretty good reasons Google would have assigned a team of more than 50 people to work on Google Instant.
Google said Wednesday that "using Google Instant can save two to five seconds per search." I'm sure some of the participants in Google's usability studies did find what they were looking for faster with Google Instant. But in the real world, my guess is that the multitude of new options will send people down unexpected alleys, adding to their overall search time. Which couldn't possibly be a bad thing, from Google's point of view.
But it might take a while for such effects to trickle down and have an effect on Google's bottom line, or on AdWords users' monthly bills. "Google is good at releasing incremental changes that may not result in huge impacts immediately but over the long term [may] certainly affect the way people search and do business on the Web," says GinzaMetrics' Grieselhuber. "That's one of the things I really like about them."
More from Xconomy.com: