Please point me at sites which allow you to enter words and phrases
and the site then tells you how many times that word has been used in
search engines in the last X days.
Thank you

Hello diamongeezer,

Not surprisingly, the one site I found that allows you to do exactly
what you want is subscription-based, although I did find one which
will partially do what you want for free.

Overture’s (formerly GoTo) free Search Term Suggestion Tool  shows
related searches that include your term/s  and also how many times
that term was searched on last month.  This is accessed from their
signup page https://signup.overture.com/s/dtc/signup/

Word Tracker provides exactly what you are asking:
http://www.wordtracker.com/  A free trial is available, otherwise the
service costs between $6.30 (£4.20) for one day access and $210 (£140)
for a year’s subscription. There are also weekly, monthly and
quarterly subscriptions.

Lencom sells a database of over 280,000 search keywords and
combinations , divided into 256 categories. The popularity number is
an average number of clicks per month for the one average search
engine - from the top seven engines.  Price $29  I could not find any
information about pricing of updates.
http://www.lencom.com/keywords.html

---

Clarification of Answer by tehuti-ga on 27 Jun 2002 16:05 PDT

Thanks to you both, insideinfo and frox, for the extra information. 
For me, it was the lack of info about updates that made Lencom look
suspicious.  A one-off purchase of keywords is not all that useful 
Thank you for looking at it in more depth.  Overture is the only one
that defines clearly where the keywords come from.  Both Word Tracker
and Lencom mention "top" search engines, but provide no information on
which ones (or at least I could not find it).

Excellent info in the answer tehuti. I would like to add as a note on
the numbers you will get on both wordtracker and overture. Wordtracker
info will be searches per day. Overture will be searches per month. So
if you run the number for a keyword in both, the numbers will be
significantly different. Additionally both are the popularity of their
corner of the web and do not include the WHOLE web or all search
engines and directies. I do not know of any such source and use
wordtracker and goto myself in my internet marketing.

Another way you can used to estimate the relative frequency of
searches is with Google's Adwords' program:
https://adwords.google.com/AdWords/Welcome.html

You can "simulate" different adwords campaigns and see how many times
they would show up in Google, i.e. have an indication of how many
times people would be searching with those terms.

It's free (you can simulate adword campaigns without buying them) and
it gives you data on google's users, obviously the most interesting
folks around the globe.

Have fun!
frox

About the Lencom database at
http://www.lencom.com/keywords.html

Where do they get this data?. They say "from the top seven engines".
Ok, does this mean they are buying data from google AND altavista AND
yahoo AND ...
If it's so, why don't they tell?
Also, most of these engines do not appear to be selling these data.
I find it worth of suspicion when somebody sells something whose
sources are unknown.

Also, odd data is contained in their sample records foud at:
http://www.lencom.com/ftp/dtbs/SampleKeywords.txt

Consider these keywords and their frequency:
"internet movie database" (65208)
" internet movie database" (36996) note the trailing space

Note 1: If they cllect data from search engines, a search engine that
does not ignore a trailing space is worth NOTHING.

Note 2: Search engines do incur in variations in writing and spelling.
It's odd, though, that "the top seven engines" they claim to use all
have this single kind of variation (inserting a trailing blank), no
other type of variations (i.e. spelling errors etc.)

Note 3: It's also odd that this variation is SO frequent. More that
half of the "correct" writing. To have a look of what real variations
in search terms look like, see the page
://www.google.com/jobs/britney.html
Here, the first spelling variation in search terms has less than 10%
of occurrences of the "correct" spelling.

Conclusion: I would NOT trust this data!

Just my 2 cents' worth!
Frox