Million Dollar Question

[huile]

Ok before i actually ask the question, i must say that this is indeed the million dollar question, that i am sure many before me have asked.

THe question is. WHat would be the best way to preform mass niche identification and corosponding keyword classification>?

I have reviewed a few different programs, the one that i will ential here is nichetracker.

Niche tracker is a good but i am more intersted in producing something in house. and i am looking for mainly opinions on this matter. So in actuality there is no wrong answer here.

Would scraping different affiliate marketing website be a viable solution in designation of different niche types and keywords? Or is there a better more viable solution to this?

With one thing that i have leanred is you have to be very organized in SEO.

So that brings me to my next question.

After all the niches are identified in your opinion what would be a good way to break down the sub niches and corosponding keywords in an automated matter?

A good example of this would be the niche 'porn'

within that niche there is possibly hundreds of different sub catagories such as.

Voyer
Teen
Tits
etc....

Would be creating something to identify how words relate togehter be a viable solution to something as this and use the base niche as the governing factor here?

After identification would a ranking system then be best placed. and have the ranking system be bassed on search querries, ammounts, traffic etc...

So that is my newbi question. I have set the stage with a few base ideas, and i am interested to hear if that is something that is even possible. and yes i am very well aware that this does not even come close to solving the other part of the equation and that would be content.... But that is not the platform for this current question.

 

[ImagesAndWords]

Good questions, and ones that can be answered in many different ways - depending on where you're coming from. As for me, I have an analytical mind and I enjoy working with numbers and statistics. Therefore, I like to approach niche research with
1) Gathering of data / stats
2) Analysis of the data using a combination of miscellaneous tools and a little helping of "gut feeling" and common sense.

Alright, so I sound vague but - what I do for niche research is to use one of the many free online tools that give you search-volumes for a lot of terms. These could be in general by month (like the one over at SEOBook), or day-by-day (like the "In the News" type of words).

For mass research, I like to gather lists from these tools of all kinds of words and then put them into an Excel spreadsheet. Excel is nice for sorting/grouping/filtering out rows that meet certain criteria. I use my own custom spreadsheets that take into consideration things like search volume, competition, cpc, ppc, etc.
I also try to group related words where possible.

Although I can do hundreds of niches at a time like this, it is still a very manual process of course. And what you're probably looking for is a more automated (or at least scriptable) way of doing all this.

I'm interested to hear other people's input on this.

Another way that saves a lot of time is to get a hold of one the huge niche databases like *cough*Dowser*cough*

 

[huile]

That is some good input, using a spread sheet is viable. Data storing in my opinion is one of the easiest parts of this entire question. ANd i want to thank you for taking the time to address this. Because without data storage and manipulation in an effective manner the system will crumble.

I am interested in this part specifically though

Excel is nice for sorting/grouping/filtering out rows

My exact question to help nail this down for my own personal understanding would be the specific word use of 'grouping'

could you please define this a little better for me.

Are you making a raw classification bassed on prefix and suffix?

or

some other form of relationship


Or are you talking more of a raw aspect in respects just grouping similar terms bassed on the different values of the equation. which would consist of Querries, CPC, CPM etc etc etc.... bassing your filters, and rules on soley those base numbers, and not taking part in any text indetification/manipulation

And if you are referring to the above mentioned making a classification on prefix and suffix, how would something like this situation be handled and also words that mean 2 different things, but are the exact same words.

For example..

Lets take the word 'Blue' into account.

RIght off the top of my head 2 very different uses for the word come to mind.

I am feeling blue

and

The sky is blue.

Both statments say entire different things but the word blue relates to the word 'Sky' and 'Feeling' differently even though this is the same exact word spelling and all.

Would a good approach just be filter out all the potential noise words, and build a statistical relationship between the meat of the sentance?

Any opinions on any of this?