1

I'm doing some research on certain demographics on a certain site and basically, because there's no API I can tie into with these sites, I'm forced to do something like this:

https://www.google.com/search?num=100&hl=en&safe=off&&q=site:okcupid.com/profile/ +intitle:sexy

Now Google says there are about 27,100 results which is believable, but it's only returning me results where the word "sexy" is distinguishable, "sexy-momma" and not "sexymomma", whereas the other 'hidden' results probably have the word sexy in it like so: omgshesosexy123

I want to see omgshesosexy123

What am I doing wrong?

Improve this question

2 Answers 2

Reset to default
2

Google indexes and searches on whole words, not regex patterns. They appear to use a dictionary approach for plurals and other variations and it's pretty clear they parse URLs to separate domain names and other segments of the URL into the likely words that have been run together. But that's because URLs are an important match (not as important as matches in anchor text but more important than matches in the title or body of the page) and they are often run together. I don't believe Google is doing the same parsing of individual words in a title. After all, it's obvious why people run words together in a URL but why would they need to do that in a title on the page?

Improve this answer
0

You should note that result counts are not a good basis for academic research. Please see:

http://searchengineland.com/why-google-cant-count-results-properly-53559

for an explanation. I can also confirm that the other answer is correct: search engines attempt to parse run-together words out of domain names, but not text in a page.

Improve this answer