Forum Moderators: open

Message Too Old, No Replies

SPAM and Page Rank

What about a paradox solution?

         

RMorg

6:51 pm on Jul 27, 2004 (gmt 0)

10+ Year Member



We displayed about 34,000 pages for antique radios and we got many indexed by Google – naturally also Forum pages etc. and arrived at PR6. We then added 38,000 radio model pages for the US collectors but yet have only very few data because to fill the data for each model is a very big job, done by us 2000 members.

We dropped to PR5 and the daily indexing dropped in numbers – an alarm signal. We first were told that it possibly could be the cloaking but as I learned in /forum24/622.htm it is not.
But our content – very valuable for collectors – and often unique – is too little compared to the line headings and the design elements which are ALWAYS the same.

First step: Line headings only where there is content on the line = realized.
Second step: We have to bring the meta tags to a minimum length.
Third step – and here comes the question:
As one can see on model pages we now have added about 50 lines (30 to 70 – always different number) below the shown model. These show country code, name, year, first valve and some notes for randomly taken radio models – a line per model – some sort of table.

This will only be of interest to a new collector if the line (the model name of it) will be a link and the user can see other models directly by this – now it is more or less SPAM to not get a penalty for SPAMMING … The process to calculate takes 7 days and we are in the 5th day and see some progress on indexing already. With - rmxorg random - one googles such pages as examples. The combination of rmxorg and year brings also the ones not yet in “the new style”.

Is there a danger?

The even more difficult question:
We calculated it in a way that the 45 random models can be changed to a link and the linked model will link back. Until we get many of these pages indexed I don’t dare to open this as links – but think it would be most fortunate if we would. But: Now these links would all link out to models not yet indexed – only 1 % would be linking back!
What would happen to such a model page? Can somebody predict?
Do you know a better approach for 72,000 pages alike to distinguish them with as such correct content?