Preparing articles for CS - technique

[comando]

Thread started here


And at this point tomtom rightly suggested to move it to CS section because the thread took a turn: tomtom has been showing us a cool technique to write articles to use in CS.

[tomtom]

Hi

Yeah this is the right place for it.. the even the topic in the last place sort of seemed out of date

You now have an idee of how I work my articles... the keyword you saw in my sample gives you an idee of how good this software can be in an automated blog.. new post every few hours, all year around.. and all good for reading...

tomtom

[tomtom]

I suggest that anyone who is setting up a new set of articles start by using a smaller amount of text to start with.. so it easier to over view it.. some sentences can be reused in coming projects.. and some will fit almost anything. you will soon find your favorite ways of working..

tomtom

[comando]

just to keep some sense of continuity I'll past tomtom last post here:

He man! Thats damm fantastic! Only ONE important fault with it.. it is to well written!
If you produce articles like that, you will struggle to make each sentence REALLY uniqe from a search engine spiders view of things.. You risk only being able to have very few alternative word for each word in a sentence..
I think I work hard with mine, but I am no were close to compleating that task you have there.. but dam it would be great! You surely have the way to work things clear!
The way you build and alter the positions to make the texts unique is 100% correct.. But only your first 5 block contain 415 words!!!! That is huge...

Make a stop here and consider the following:

What is your reason for publishing a website?
a) I think I can really give the visitor some nice knowledge and a good time reading.
b) If I have good content, no one can say it is automated crap, search engines will love me, but any visitor that start reading will soon look for the first best thing to click on to get out!

If a), why do you want CS?
If b) You are on the right way, content should be so good that a manual review will say "OK this guy is maybe from Poland, he needs to practise his english gramatics a bit, and his design is not to good, but its nice of him to try and help people..."

If b), I prefer to have smaller articles, have about 150-300 words in total.. that is enough and makes it viable to complete the set of articles...

I will PM you 2 articles from one of my feeds, so you see what I do.. OK.. you are on the right track here ( I think..) But you have to down size it.. shorter articles wiht a lot more alternatives.. that makes it unique...
Hang in there for a PM and you will see what I settle for..

tomtom

PS. I mean nothing bad about the Polish members of this forum...

hehe, I kind of suspected while I was writing that I was going to have some trouble spinning all those snippets of text once imported in CS. I guess a more basic and rudimentary literary style is needed.

For example, instead of:

Once I figured out what the basic features of [keyword] were and the fact that I was getting a pretty good deal our of it I decided to make the purchase.

should perform better for our purposes written like this:

[keyword] has many good features and it is possible to get a good price so I decided to buy it.

---

Would that be correct?

thanks for the PM   I will keep working on this and banging my head against the wall until I get it right.


Of course my answer to your question is b)  !

[tomtom]

hey

yeah, that is much better, you now have a more workeble starting position..

I try to have alternatives for 7 out of 10 words in a sentence, this can be very hard, but since some single words can be replaced with "word blocks" of say 2-7 words, it comes out pretty good anyway..

The last thing you want to do is have a sentence like this (read from top and down, it is not on one line as I want to be able to have alternatives in next sample I post..):

It
has
always
been
clear
to
me
that
[keyword]
is
a
good (good, great, fantastic, positive, desired, wanted)
thing.

The words with in the ( ) are the alternatives on that line.. this would be REALLY CRAP, AND USELESS TO HAVE...
even if there where alternatives for say 2-3 more words, the "keyword" would always be in position 9, and the sentence would always have a set amount of words.. NO NO NO.. NOt good at all!..

Give me 2 minutes and I will post again wiht a better sample of a short sentence..

tomtom

[tomtom]

A simple, very simple, line of text that can be reused in several diffrent articles or as coments after articles..

We^cap  (we, I, me and some friends, me and other users, alot of people)....
have (have, know about, are familiar with, use, currently use).....
the
keyword
info   (info, content, information, materials, information, relevant info,).....  
you   (you, most people, you and many others, alot of people)....
need  (need, want, desire curently are searching for, are trying to find)...
!


This simple combination would give an output of between 7 and 20 words, keyword would be in posistion 4 to 8.. but Im sure I could fix it better if I pent a few more minutes on it.. something like 7-40 words and keyword position from 4 to 35..

Tomtom

[comando]

Excellent. Thanks for all your help man. I'll keep working on this

[comando]

Just a follow up here, I've been putting into practice this technique and it is looking very cool.

I am able to output articles that differ from each other on a range that usually is close to 95%. I run the output feeds through copyscape.com and get zero results.

And with this technique, articles also differ in size, paragraph structure and keyword density.

Even more, they look somewhat readable. Not well written, but they make some sense.

[comando]

Ok another follow up..

I thought the most time consuming task would be the actual original writing of all the sentences. But it is taking me a lot more time to manually place synonyms for each article inside CS.

Regarding this. A couple of questions:

1 - I have 10 variants for each one of the 10 sentences (i.e. I have 101,102,103...all the way to 110; and same with each one of the others.)

I'm not sure exactly how many variations of articles I can create with such amount of feeds. But I guess 10*10=100, which is more than I need.

So the question is: how many should I create? I know the answer would be: it depends.

But I'd just like to have some sort of reference. For example: it depends, if you want to create content for 10 domains with 500 pages each, then you need to work with 10 articles. (assuming I'll be able to produce an output of 50 output articles per seed article). Ok, Im guessing this is VERY conservative as I could create a lot more content with 10 seed articles.

2 - The next question is strictly related to the first one, so it's really a continuation of the first question: How do I know how many articles I am able to output for each seed article making sure all the outputted articles look different from each other?

I should expand further here. Let's say my article has 110 tags and 600 values for those tags. This means in theory I could output 66,000 different articles.

Correct? (110*600=66,000)

Now, if I'm not mistaken, that many articles could be exported but that doesn't mean that the variation from article to article would be any more significant than the replacement of just one word. Correct? For example, article number 65,489 would differ in just one word from article number 65,490.

Correct?

Now.. assuming that I'm correct there, what would be the optimal number of articles to be exported, in relation to the amount of tags and values?

Would it be something like: (tags*values)*10%     ??

Or 5% ? Or 30%?  Or what....? Im just guessing.

My guess anything  below 30% should be ok..

My take on 1 and 2, here is where I answer my own questions... but I'm just guessing: (speculation)I think that if I have 3 seed articles I can probably export at least something in the range of 15,000 to 20,000 original articles of decent quality per article, so about 60,000 in total.

That would allow me to feed more than 100 domains with amount of pages for each domain in the range of the 500.(/speculation)

[comando]

Something I've just thought about: the rewriting process should get faster as I start reusing some of the lines of text, but I'm not sure CS makes this possible.

To make my point.

Let's say in Article 1 I use:

102+202
406
605

and in Article 2 I use:

102+204
507
709

--
If you look at those numbers, you'll see I am reusing the snippet of text 102. Since I already took the trouble to rewrite this snippet in my first article, I would like to be able to reuse this piece of work in Article 2.

I don't think you can do this in CS.. or maybe I'm missing something?

[tomtom]

Hi

You are right about one thing and that is the amount of variation in one article, BUT your math for how many possible aticles you can produce is wrong..

110 X 600= 66 000.. yes but this is not correct math in this place..

if you have 110 words that you can replace with an average of 5 each (that would be the 600 words..) the math to this would be:

5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5x5= pretty dam much..

I cant even calculate this.. so lets break it down to more handelble size:

One sentence contains 7 words. From these 7 words 5 can be replaced with 5 words each:

word1 word5 word5 word5 word1 word5 word5.

This would equal: 1X5X5X5X1X5X5= 3125 alternative sentences from this seed of 7 words.

So ok, lets take a paragraph of 3 sentences. Each sentence contains the amount of words and alternatives as above. This would give us the following nummber of output alternatives:

3125 X 3125 X 3125 = 3 051 175 780 ALTERNATIVES.. read that again.. that is more then 3 000 million alternatives on that paragraph alone!

BUT, and this is a big BUT.. this is just like you describe it, article 152456789 would only differ ONE word from article 125456788..
Please remember that this is only on a 21 word paragraph.... BUT any way.. this is why I use different sentences for the different positions in a paragraph.. add to this the length and nummer of words and the key word possition in the articles.. ..  to put it easy:

This 3 sentence paragraph that contains 21 words can be made in to 3000million+ sentences, but if I only use it 500 times it will be pretty damm unique..

Still not convinced? Here was a 21 word paragraph from the same feed sentences, I have shown you earlier how to make alternative start and sentences, length of paragraphs and how the amount of words can be altered in a sentence.. how many unique articles I create? It is unlimited....

Hope this helps you

tomtom

PS. I will read your post again as I only answered one thing here and forgott what the others where about...

[comando]

So ok, lets take a paragraph of 3 sentences. Each sentence contains the amount of words and alternatives as above. This would give us the following nummber of output alternatives:

3125 X 3125 X 3125 = 3 051 175 780 ALTERNATIVES.. read that again.. that is more then 3 000 million alternatives on that paragraph alone!

Ok so...let's say of these 3,051,175,780 possible outputs, I only use 1% of them, that s still over 30 million (30,511,757)

Given that these 30 M represent 1% of the total possible outputs, would I be correct thinking that they would differ from each other in 99% of their composition?

Also, when I create the feed in CS I have this option: Export [] number of each selected article. What's that about? Does it mean if I leave it at 1 it will only output 1 version of each article? If I enter 1,000,000 it will create 1,000,000 versions?

As always, thanks tomtom for your help

[tomtom]

Yep.. that would be right, BUT (there is always one...) you would then have about 30 000 000 articles that where really uniqe when read, but the amount of used words in the articles would only be 81... not good to create 30 000 000 articles using just 81 words.. so again, dont let the math behind it all fool you.. stay realistic...
The export nummers of articles I dont know.. I still use CS1 mostly for 2 reasons,
1. there was no way I could move my articles from cs1 to cs2.. and as you understand, I did spend a lot of time making articles..
2. I havent recived my free upgrade yet.. ( I forgot to update my contact info..)..

tomtom

[majordude]

..the keyword you saw in my sample gives you an idee of how good this software can be in an automated blog.. new post every few hours, all year around.. and all good for reading...

n00b question: How do you get content (articles, not rss feeds) from CS2 to WP?

[crabman]

Thanks to Comando and especially Tomtom for the info in this thread!

I read this some time ago and will make a start today trying to implement it.
Hope I still have CS1 on my server.