Google Suggest

http://www.google.com/webhp?complete=1&hl=en

Does anyone know how this works? It seems like if I knew a bit more about this it could be useful for picking keywords to target if there is any specific order to the suggestions that come up.

No idea, it gives me a headache just thinking about it! :)

It seems you type what your searching for, wait a second, and it lists other possible terms in a drop down. Or am I stating the obvious and looking like a fool? :)

Stating the obvious. :)

I was just wondering how it chooses the order of the suggestions. For example, when you type in "free", it suggests "free games", "free ringtones", and "free e cards" next. Information like this could be useful if we knew that the suggested were generated, say, from most popular searches to least. If you have on a certain topic, you could type it in and find related searches that get high traffic. In some cases it might be worth targetting those words too.

Looking at this again now, maybe it wouldn't be much help at all, as it only makes about 10 suggestions.

This may help: http://labs.google.com/suggest/faq.html and perhaps if you are feeling adventurous, this: http://serversideguy.blogspot.com/20...dissected.html

they need to clean up the results. lol. I type in the letter p... first world "Paris Hilton"

You know why Paris Hilton shows up first? Because its one of the most searched keywords. So yes, there is a rough correlation with the number of searches.

i know.. but it is kind of like M$ Word... you can type a word close to a swearword and it wont show it in the autocorrect list, but if you type the word correctly it does not put a little line under it. OpenOffice lists the swear word in the autocorrect and the spell checking dialogue.

It's not like Paris Hilton has to directly correlate with 'unclean' search results. It could just be some little girl (or boy) who idolizes the psycho (I'm sorry, but never wearing the same thing twice and paying thousands of dollars for each new outfit per day does seem to be exactly that to me).

I'd say it'd be good for finding site topics and analyzing the number of websites in those particular niches; but quite frankly, all I can see is program and brand names.