Checking for content duplication against content on your own site.

[MichealGooden]

We are currently trying to rewrite our product descriptions and I'm afraid some of the salespeople that are writing the descriptions are plagiarizing one-another's writing. Is there a content duplication checker that will allow you to check a piece of writing against a specific site rather than all of the web?

[Cornel_Ilea]

Hi Michael,

Maybe this can help you. http://www.copyscape.com/

Cornel

[MichealGooden]

That site searches the entire web for copies. I'm looking for something to crawl my own site for duplicate content.

[Cornel_Ilea]

The duplicate content from you website is shown in the SEOmoz tools.

Check the Crawl Diagnostics Summary:

Cornel

[MichealGooden]

That's for pages that are already published and crawled. I want to able to search my site for entire sentences and/or paragraphs of text that I have yet to publish so I can make sure it's not being used elsewhere on the site. The crawl diagnostics tell me I have duplicate content after the fact - I'm trying to take a proactive approach rather than reactive.

[Jinx14678]

Just off the top of my head, there are a few low tech ways to do it....

If you have Win 7 the searching has improved greatly - just move all files to a local machine - and search the directory you placed in for the content you are wanting to check - it will give all files that contain the words. (but can become overloading)

If you have dreamweaver or other enterprise level editor - almost all have a site search function to where you can search/profile code/text and have it find one by one which pages contain the searched terms - or globally list them.

Other than that, probably a custom script -or a google search for an HTML profiler might help?

Shane

[MichealGooden]

Those are good answers and would work on a smaller scale site. We currently have over 17,000 product pages so I can't really use either method. It's looking like a google custom search is the best bet even though I can't search an entire paragraph at a time.

[Cornel_Ilea]

Hi Michael,

Having a website that big means that you might have a test or dev environment.

If not create one.

if you have something like  test.yourwebsite.com and submit it  to the SEOmoz tools as a new project you can see a report before your website goes live.

Cornel

[MichealGooden]

I have two dev servers, one of which it is possible to do what you're talking about but that is the absolute least efficient tool to use for this.

The crawl diagnostics are updated about once a week which means I would have to post the new content and hope I got it online in time for the crawl. If I didn't then I would have to wait an additional week to see results.

The crawl diagnostics also limits the amount of pages it will crawl on your site to 10,000. I stated before that I have over 17,000 pages. So even if I did use this method, the chances of that page being crawled is little better than 50/50.

Also, the crawl diagnostics only tell you what pages have duplicate content - not the exact content that was duplicated. That means I'd have to manually find the page I'm targeting, then follow the supposed duplicate content suggestions proposed by the crawler and find the similarities myself.

I think it's very safe to say that the crawl diagnostics, nor any product that SEOmoz provides, is an answer to my issue. If I thought it was, I would have already been using it and would not have posted this question.

[CleverPhD]

I assume that you have an admin section in the CMS where you are editing and entering these articles before they go live.

You need to get a developer to simply write a search algo that when you create a new article and before it goes live, it takes sections of your content and looks for matches/duplicates.  You can set a requirement that it has to match on a minimum of a 4 to 5 word string and other such limitations to make sure you are not matching too many items.  It will take a few tests to find a sweet spot of too many matches vs not enough.

With 17K pages, this is the only way you can really do this in an efficient way, you need some IT support/development. They may have to create a reporting layer as well to help you sift through the results.

Good luck.