Hello i will try to explain the topic more clearly , is it possible to determine what users likes in terms of statistics just from collecting data from software download sites like sf.net (open source ) or download.com (shareware) in simple words .. is it possible to make statistics on users from free resources on the net ?
For web stats:
For top software downloads, you may want to look at download.com & sourceforge and such and see what downloads the most.
Sourceforge top downloads:
see this - http://www.quantcast.com/
Be wary of all of these "analysis" sites.
I've found massive (as in, orders of magnitude) discrepancies between these sites measuring a single site, and between my own numbers for my sites and what these say.
Furthermore I've asked other people about it -- both experts and Twitter queries -- and many people agree with this.
Use these sites as a very general guide but don't go crazy about it.
Competitive Intelligence is an integral part of internet marketing. It's not all obvious why however, but I can best explain it with an analogy.
Relying solely your standard web analytics package (such as Google Analytics or Omniture) is a lot like driving a hot sports car while never looking out the windows. You may know you're doing 60 MPH, but you have no idea if the other cars (your competitors) are doing 80 and passing you by.
CI tells you this. Furthermore, a good service will tell you exactly what they're doing as quickly as possible (within 24 hours).
I started out as an affiliate marketer and created such a technology in 2004 to figure out what the established guys in my space were doing to drive traffic. Within the course of a few months I became the #1 reseller in the category and was collecting $100,000 checks with less than four hours of work a week (It's harder now, to be sure... but I still know people who are doing very well). You can read my story in this book: http://www.amazon.com/Ultimate-Guide-Pay-Click-Advertising/dp/1599183633
I ended up commercializing the software and anyone can use it today (http://www.adgooroo.com/). It isn't cheap, but it is by far the best software available for this purpose. It's also the most widely used CI solution for Fortune 500 and search agencies. If you use it right, you WILL have what you need to dominate any industry.
Anyway, back to my reason for quoting you. CI tools do NOT look at your site and will NEVER match exactly with your analytics data. They are either spying on actual users (panel-based) or are measuring reach/frequency of advertisers (crawler-based). Both are useful, but the panel-based approaches are generally limited to sites that generate more than 5,000,000 uniques per month. Crawler-based approaches (such as ours) are useful for everyone, but you have to make sure that they gather enough data to be statistically reliable. The cheap services do not do this (how could they at that price point???)... and this is why the data never matches up with reality.
A few questions to ask of a vendor:
A) How often do they go out and collect your data? It should be at least 10x/day. B) How often does the data get refreshed? You want it within 24 hours. C) Do they have permission from the search engines to gather this data? D) How do they gather the data? Do they have servers hosted in the actual countries? Or are they using URL hacks to trick the engines into thinking they are crawling from there? This is really important because the URL tagging tricks don't work for natural data (and this is why most of them can't collect SEO data).
Hope this helps.
Richard Stokes CEO, Founder AdGooroo