Why Google went instant | Hacker News Reader

Why Google went instant

technologyreview.com

20 pointsanderzole14 comments

Here is my theory.

Google instant does an even better job of getting the user to use their predetermined search query rather than continuing to type out the users longer tail search query.

By funneling search traffic into these head terms, Google is concentrating traffic around the terms with higher AdWords CPCs. And as more and more users are funneled into these terms it will increase the level of competition bidding on these terms, further driving up the CPC to appear of the first page.

Seems the most plausible to me!

Until you ask, well why do people use long tail search terms? Might it be because the "head terms" do not offer them what they wanted.

Unless, long tail search terms really are something of a waste of time for the user and the head term can offer them just what they wanted anyway, but the user was ignorant of that.

Scenario:

A user began typing what would have been a long-tail query, but realizes the results have satisfied their query without typing the full thing.

I'm not saying long-tail is going to disappear, because it's not, at least not for now. But I'm sure a significant portion of would-be long-tail queries are going to be rerouted into short tail queries.

You're absolutely right. I'm already seeing traffic shift away from (cheaper) long-tail keywords in my campaigns.

My theory is Google is trying to make their search engine as fast as possible and by not clicking the enter button it reduces 1 second and also appear to be faster.

seems that the switch allows them to funnel people into the head terms, away from the long tail, which would improve their cache performance, saving money (which was briefly mentioned in the article, too).

Organic results are cached, but adwords ads are calculated in real time.

There's a certain vacuity to the core point here... "Why did Google do this thing? Why, to improve their bottom line." Well, yeah. Let me tell you, the company I work for also has quite a few initiatives that involve improving the bottom line. Wow.

I just noticed that Google autocompletes 'j' with 'justin bieber'. I realize that letter has a high scrabble value but c'mon.

Google is in a unique position to know the first-order Shannon compression of the letter "j".

for the record, I get 'JetBlue' for "j". "s" gives me "skype", q gives me "quotes".

It's actually kind of fascinating.

Java. Maybe because I'm sitting in the CS dept right now? :)

The results are definitely reflecting the geolocation of my IP (I'm not logged in and have cookies off).

I think I need to move somewhere that autocompletes to java.

Maybe Google is afraid of Twitter's search potential?

Twitter search might be useful to track breaking news, conferences or celebrity gossip, but I wouldn't count on it to find something remotely useful.

I don't even claim that low signal-to-noise ratio is to blame here. I believe it's rather the length limit, it's obvious that the more content you have, the better index you can build for a given document.

PS. Google has its own real-time search, crawling primarily Twitter updates. http://bit.ly/9PUpaf

Reply on news.ycombinator.com