Google Indexed the HTTPS version of an e-commerce site

[brianspatterson]

Hi, I am working with a new e-commerce site.  The way they are setup is that once you add an item to the cart, you'll be put onto secure HTTPS versions of the page as you continue to browse.

Well, somehow this translated to Google indexing the whole site as HTTPS, even the home page.  Couple questions:

1.  I assume that is bad or could hurt rankings, or at a minimum is not the best practice for SEO, right?

2.  Assuming it is something we don't want, how would we go about getting the http versions of pages indexed instead of https?  Do we need rel-canonical on each page to be to the http version?  Anything else that would help?

Thanks!

[OptimizeSmart]

let the people redirect to non https versions. what is the problem here? They won't loose items from their basket when redirected from https to http version. And when they are 'checking out', the protocol will remain secure via SSL as pages which need to be https wont redirect to non https.

[brianspatterson]

Hi Irving, thanks for your reply.  That all makes sense to me except "noindex meta tag them".  They are the same product pages whether they are https or http, so I can't put 'noindex' on exclusively the https page...

Or are you suggesting that I figure out some conditional code that if the https version is called, it inserts a 'noindex'.

Is there a reason nobody is suggestion rel canonical to the http version?

[irvingw]

block the https pages in robots.txt and nofindex meta tag them.

then make sure that all of your links coming off of the https pages are absolute http links.

Your problem is probably relative links on the https pages getting spidered and staying https when they come off the secure pages onto the http pages if that makes sense.

[brianspatterson]

301'ing https versions would not work, because people who belong on the HTTPS versions (because they have something in their cart) would be force-redirected to the non-https version.

I'm thinking that rel-canonical to the http version along with Robots.txt rules as you've suggested may be the way to go.

[OptimizeSmart]

1. it can create duplicate content issues and not a good seo practice.

2. You can 301 redirect all the https versions to the http versions and apply meta robots noindex, follow to the handful of pages that need to be https.