What to do about similar product pages on major retail site

[RyanKelly]

Hi all,

I have a dilemma and I'm hoping the community can guide me in the right direction.  We're working with a major retailer on launching a local deals section of their website (what I'll call the "local site").  The company has 55 million products for one brand, and 37 million for another.

The main site (I'll call it the ".com version") is fairly well SEO'd with flat architecture, clean URLs, microdata, canonical tag, good product descriptions, etc.

If you were looking for a refrigerator, you would use the faceted navigation and go from department > category > sub-category > product detail page.

The local site's purpose is to "localize" all of the store inventory and have weekly offers and pricing specials.  We will use a similar architecture as .com, except it will be under a /local/city-state/... sub-folder.

Ideally, if you're looking for a refrigerator in San Antonio, Texas, then the local page should prove to be more relevant than the .com generic refrigerator pages. (the local pages have the addresses of all local stores in the footer and use the location microdata as well - the difference will be the prices.)

MY QUESTION IS THIS:

If we pull the exact same product pages/descriptions from the .com database for use in the local site, are we creating a duplicate content problem that will hurt the rest of the site?

I don't think I can canonicalize to the .com generic product page - I actually want those local pages to show up at the top.  Obviously, we don't want to copy product descriptions across root domains, but how is it handled across the SAME root domain?

Ideally, it would be great if we had a listing from both the .com and the /local pages in the SERPs.

What do you all think?

Ryan

[JaneCopland]

Hi Ryan,

I guess the first point here is that Google doesn't treat this sort of filtering as "penalisation"; it's just filtering two or more versions of the same content because it believes (sometimes mistakenly) that users don't need to see two versions of the same thing. This gets REALLY tricky in fields like real estate when all the aggregators in the same town have access to pretty much the same feeds or properties.

If Google were perfect, you'd put up the two pieces of identical content for all 55 millions products, and Google would serve the right one given the appropriate query, like the example above ("fridge sale san antonio" brings up the local page; "refrigerator" has your main site rank). And this might happen, because Google is getting better at these sort of query-appropriate results. We still recommend not providing dupe content solely because we can't be sure that Google will get it right.

As an aside, it would be so great if they worked on a tool for localisation in the same way that they have given us the href lang tag for internationalisation. rel="city" or similar would be awesome, especially for big countries.

Your idea about serving the content from a shared source will certainly work (iframe, text hosted on separate URL, JS etc.). The pages serving this text clearly won't be credited with that text's content, which removes its SEO value of course.

[RyanKelly]

Hi Jane, thanks for the response!

I can't understand why Google or any other search engine would penalize a brand for having the same product detail in more than one location on the same root domain.   It's just not feasible to re-write all of the product descriptions for 55 million products.  The only difference is going to be the price, and some localized content on the page in terms of store locations and addresses (perhaps multiple in one area).

What if - kind of like your M&S example - the local product pages pulled product descriptions from another location on the site, but displayed them in a modal window - so a JS event displayed the proper descriptions and details for the user experience, but the HTML is devoid of any "duplicate" product description content?

[JaneCopland]

Hi Ryan,

It's going to be hard to do this without creating duplicates - if they aren't commissioning re-writes of descriptions but just pulling from the database, identical content like this is far from ideal.

One school of thought is that there really isn't any such thing as a "duplicate content penalty" unless you have some huge, gratuitous problem that results in a Panda issue. Google simply chooses the version of the content it favours and drops the other. The local site would still be much more relevant for a query like "fridge sale san antonio".

An example of a big retailer that has a similar(ish) site at the moment is Marks & Spencer Outlet here in the UK (outlet.marksandspencer.com). M&S is probably the most recognisable high street brand in the UK, to give you a perspective on size.

Looking at what they're doing, they're listing pages like this: http://outlet.marksandspencer.com/Limited-Edition-Jacquard-Textured-T69-1604J-S/dp/B00IIP7GY2?field_availability=-1&field_browse=1698309031&id=Limited+Edition+Jacquard+Textured+T69-1604J-S&ie=UTF8&refinementHistory=subjectbin%2Csize_name%2Ccolor_map%2Cbrandtextbin%2Cprice&searchNodeID=1698309031&searchPage=1&searchRank=-product_site_launch_date&searchSize=12

This is the same product as this: http://www.marksandspencer.com/jacquard-textured-coat-with-wool/p/p60056127. I love it that the "outlet" version is more expensive... anyway...

The product details, which are all included in the HTML of the main site, are not included in the Outlet page. The Outlet URL is indexed (what queries it ranks for / could potentially rank for are unknown) - but I would be keen to hypothesise / experiment with the idea that if that product was on a page about it only being available at M&S Moorgate, and looking for coats at M&S Moorgate was as popular a query as [fridge sale location], the Outlet page would rank.

You will never get an SEO to say that you should "copy and paste" descriptions across domains or within them, but essentially the pages have to provide a service / information that makes them worth ranking for relevant queries.