Big problem with my new crawl report

[ankali]

The urls are: /product/search&filter_tag=%D0%B1%D0%B8%D0%B6%D1%83%D1%82%D0%B0

after = there are a lot of combinations. Is it correct to put this in robots.txt

Disallow: /*?route=product/search&filter_tag=/

[AlanMosley]

When you no index a page, any links pointing to those pages pour away link juice from you indexed pages. you should never no-index pages IMO

I assume you are using a CMS or some sort of plug in, this is a common cost when you do so. CMS create very untidy code, not good for SEO

[ankali]

I am using opencart. I dont know what to do. Before I had 50 errors, now they are more than 500 after this plug in. The plug in removed the previous errors, but now there are many different errors. I have 2 options:

1. Remove the plug in

2. Do something with new errors - the new errors are only because of search, I have dublicate page content because when you type PDODUCT NAME in search box, there is same content as www.mydomain.com/category1/PRODUCT NAME

Maybe this plug in removed the canonical urls in search or I dont know what.

In robots.txt there is row:

Disallow: /*?route=product/search

The duplicate content is mydomain.com/product/search&filter_tag=XXXXXX

Instead of XXXXX there are many paths.

I decided to add another row in robots.txt:

Disallow: /*?route=product/search&filter_tag=/

Do you thing it is correct or to remove the plug in?

I hope you understand what is the problem.

[ankali]

Please I need help

[Everett]

Hello Anastas,

I agree that you should block the search folder from being indexed. I'm going to assume that nobody is linking to your search pages and that you have other paths (e.g. SEO-friendly navigation, sitemaps...) for search engines to use to access your products).

I don't understand why you have formatted the disallow statement that way, however. Unless I'm missing something (and could be since I don't know what your site is) you only need to do this:

Disallow: /product/search*

And of course after doing this you should test it in GWT to make sure that A: You are blocking the pages you want to block, such as search pages with lots of parameters, and B: You are NOT blocking other pages you don't want to block, such as product pages. Here is more info on where to find the testing tool in GWT if you don't know: http://productforums.google.com/forum/#!topic/webmasters/tbikAxJiIZ4

Let us know how it goes. Good luck.

[AlanMosley]

as long as no one is linking to the search pages including internal links.

[Everett]

I have tried different methods to fix this. First-hand experience tells me that oftentimes it is better to just block the paths (assuming there is better navigation on the site) from being crawled or indexed using robots.txt than to use a noindex,follow tag in order to save the pagerank you're sending via internal links. It is very easy for Google to get bogged down crawling around in the internal search results area.

Unless there are lots of links to search pages from top pages on the site, or a big list of search page links from every page (sitewide footer, for example) I really don't think the waste of internal pagerank is noticeable in the rankings, or worth salvaging if it risks sending spiders into a maze or a trap.

Yes, best practice is not to link to pages that you are blocking. In the real world though, search pages can be very useful to visitors, and to merchandisers who don't have the ability to create more targeted sub-sub-sub categories will often use them, and link to them on the site, as landing pages for promotional purposes (emails, PPC, sales...).

Everyone has their own strategies, and all we can do is make recommendations based on our own experience and knowledge. Thanks for helping out with this question Alan. Feel free to elaborate so Anastas has more input to help guide his decision.

[AlanMosley]

The problem with PR leaks is that they are scalable, If you are losing 10%, then you get some quality links, 10% of them will be wasted, every effort you do in the future will be discounted by 10%.

There are ways to fix all these problems, for example I would make a search to be POST and not GET so that links to search pages can not be made and therefor search pages will not get indexed.

We work so hard to get good links, why waste them when you do?

[Everett]

Alan,

Thank you for the great advice. If one has enough control over the eCommerce system, or the internal site search product, to change from GET to POST so these pages act more like real dynamically generated "search pages" than an infinite amount of "landing pages" I think that is a fantastic solution. It would keep merchandisers and others from linking to those pages - because we all know that they will continue to do it even if the SEO pleads on hands and knees for them to stop.

However, I have found it to be the case that most eCommerce businesses (from small mom-n-pop shops to fortune 500 companies) do not have the ability to do this because the internal site search functionality they use is out of their hands. Site search vendors like Endeca and Celebros serving enterprise eCommerce businesses don't typically hand over the keys to the client.

If you know any site search vendors or solutions that allow one to do this it would make a great contribution to this thread if you could share a few of them. I'd definitely look into recommending them in the future!

Thanks again!

[AlanMosley]

I use Bing search API,

By the way, you want to change from GET to POST, not the other way around.

[Everett]

Thank you again Alan.

Typo fixed.