Weighting by URL length

[b2bwebmaster]

Hi,

Need my "new" catalog pdf files to show up before the "older_catalogs" pdf files in my search. I changed the "Weightings" to "URL length" = "strong adjustment" (so the shorter URLs would show up first) but they aren't.

Here are my results for 31750:

1. http://69.161.159.122/datajabsco.htm
Your source for industrial and marine pumps in centrifugal, gear, flexible impeller, lobe, diaphragm, and submersible for metering, chemicals, coo...
... Series.pdf 720K 31670-Series.pdf 349K 31700-Series, 31705-Series 2004.pdf 456K 31750-Series, 31755-Series 2004.PDF 286K 31765-0092, 31765- ...
Score: 14
2. http://69.161.159.122/catalogs/older_catalogs/marine/catalog109/12.pdf
... GPM Volts Amps Inlet/ Outlet Dimensions (H x W x L) JA 31750-0000 3.7 12- 24 1- 10/" Barb and Qest ...
Score: 14
3. http://69.161.159.122/catalogs/older_catalogs/marine/catalog108/13.pdf
... Amps Inlet/ Outlet Dimensions (H x W x L) 1 2 JA 31750-0000 3.7 12- 24 1- 10/" Barb and Qest ...
Score: 14
4. http://69.161.159.122/catalogs/older_catalogs/marine/catalog107/13.pdf
... Amps Inlet/ Outlet Dimensions (H x W x L) ING GUIDE JA 31750-0000 3.7 12- 24 1- 10 1/2" Barb ...
Score: 14
5. http://69.161.159.122/catalogs/new/marine/catalog110/12.pdf
... Volts Amps Inlet/ Outlet Dimensions (H x W x L) HOME JA 31750-0000 3.7 12- 24 1- 10/" Barb and Qest ...
Score: 14

----

Here are all of the "weightings" adjustments:

page title = normal
description = normal
heading = normal
file name = +1 boost
keywords = normal
link/alt text = normal
body content = +5 boost

word positioning = no adjustment
content density = no adjustment
URL length = strong adjustment

-----

Do you have any suggestions to make this right?

Thanks,

b2bwebmaster

[David]

All the results seem to have scored the same (14). So for this search word the documents must look about the same.

The URL length weight was meant to favor very short URLs compared to long ones. Thus favoring pages like the site's home page, that aren't many levels deep. In your case the URLs lengths for your PDFs are more or less the same. There needs to be a significant different before this has much of an impact.

What we do on our site with our old documentation is to 'deboost' the PDF files. So that they come last almost all the time.

To do this make a .desc text file (see the User's Guide for details) and include some content something like this in the .desc file.

<title>Marine Catalog 109 - 12 (old 2009 edition) [PDF]</title>
<meta name="description" content="PDF file with the 2009 marine Catalog 109 - 12 containing details about blah blah blah</">
<meta name="ZOOMPAGEBOOST" content="-5">

You'll need a .desc file per PDF you want to adjust.

As an added bonus, the same values will be automatically picked up by a sitemap and fed into Google. Meaning when you generate a site map file (which can be done without extra overhead while indexing using Zoom Search), then the Priority Tag in the Google site map file will be automatically populated. The consequences of which is that other pages on your site are given priority by the Google engine as well as Zoom Search.

[b2bwebmaster]

Thank you for the quick response and your suggestions.

I have a few more questions:

Do you think it would help if I put the current catalogs in a higher directory? We have quite a few archive catalogs (several hundred pages) that our company wants to have available. They all have titles but not all have subjects (descriptions).

Is the "<title> tag" you specified above, the Title line of the Document Properties in the PDF? For instance, in the Document Properties, the Title for page 12 in catalog 109 is "Archive Marine Catalog 109 - Page 12". I am thinking this needs to match the code in the .desc file exactly?

Also, in the PDF, the Subject is "Jabsco Variable Flow Water Systems". Is the "<meta name = "description" content="Jabsco Variable Flow Water Systems"</"> required in every .desc file?

Your assistance is greatly appreciated!

"b2bwebmaster"

ps - I love your program!!

[David]

The <title> and meta description tag are optional. If you don't include them, then meta data details from within the PDF file will be used instead. I only included them above because it is rare that people include good meta data in PDF files and if you are making a .desc file it is a good chance to include good meta data.

So a one line .desc file would in fact be enough for the ranking issue.
<meta name="ZOOMPAGEBOOST" content="-5">

[b2bwebmaster]

Thanks for this information, it's working better but still bringing up some of the "older catalogs" before the newer ones... so now I'm creating the: <meta name="ZOOMPAGEBOOST" content="+5"> .desc files for the current catalog pages to see if this helps? Do you need to reindex after you have created and uploaded these ".desc"?

[David]

Yes, you need to re-index your site.