Last week, Google rolled out a major change to its search engine’s interface. American users, and some users in other countries, began to see the results for their Google searches appear onscreen while they were still typing. For example, if you type “b” into Google’s search box, you might see the results of a Google search for “bank of america” appear below the box, along with a sponsored link from Bank of America. Keep typing until you have “bat” in the search box, and the results could suddenly change to match a search for “batman.”
Google Instant, as it’s called, is primarily about making searching faster and more effective, according to Google. The company says the service saves two to five seconds per search, and helps users refine their queries by providing instant feedback.
Google doesn’t deny that maintaining its search ad sales was a primary focus in testing the service. The company only gets paid for an advertisement, or sponsored link, when a user clicks on the ad, and ads are targeted to specific searches. By displaying a search’s ads onscreen a couple of seconds sooner, the frequency of users clicking on those ads could only go up.
Google’s search ads are a $20 billion business, thanks to the popularity of its lightweight, minimally intrusive text ads sold against specific searches. So it’s unthinkable that Google would risk its cash cow simply because employees thought instant results were cool.
But are users finding Google Instant useful? Some Internet pundits have been skeptical. One Web developer conducted an eye-tracking user study of seven people chosen to represent a wide age range of heavy Internet users. His test found that none of them were jumping down to the early-arriving results. Instead, they continued typing until they had finished the search query.
Google insists that its own heavy testing says otherwise. Company spokesman Jake Hubert says, “We actually found that people do look at the predictions and the results, but they don’t necessarily notice that’s what they’re doing.” Google is known for extensive testing of even the slightest change to its moneymaking search engine. Other recent products, such as the Google Buzz social network and the Google Wave everything-plus-the-kitchen-sink collaboration tool, didn’t catch on as hoped. But Google Instant isn’t a new product. It’s a presumed enhancement to the most successful product on the Internet.
Google Instant is also an impressive engineering feat. How can Google deliver a nonstop stream of instant results to hundreds of millions of users without melting the Internet? It turns out that the instant results are powered not by high bandwidth, but by advances in Google’s back-end data centers and inside your browser.
Othar Hansson, tech lead for the feature, says that loading search results as you type only uses a few thousand bytes of data for each new set of 10 results and any text-based ads targeted to the search. “The people that manage our traffic barely even blinked when we said we’re going to have this effect on search traffic,” he said in a phone interview, “because we have this other thing called YouTube.” The five to 50 kilobytes of data in a new set of search results, even if they include image thumbnails, are tiny compared to even one second of online video.