This process may lend itself to PPC a bit more than SEO. When split testing you will need to be aware of duplicate content, and considering that your ultimate goal is to figure out which landing pages are more effective, you will end up removing some of the pages anyway. On a large scale this isn't going to be as effective.

I would consider running a PPC account to test these pages and not have them indexed. Then, once you have a landing page that performs well, create it on the site and promote it with SEO.

If your customers would find it useful I can't see it being too much of a problem. I like a small selection of links in the footer. Often customers will look in the footer for contact links and information.

If it's just for SEO and you are putting in some keyword links I would probably say avoid it.

Hi Walid,

I would personally eliminate A and B purely based on the fact that they aren't targeted enough. I would put the "Free dliver on $50 purchase" in the meta description to encourage the click through, but C seems the most concise and targeted out of all of them.

Plus I'm a big fan of the " | " symbol

In that case I would suggest changing a couple and reviewing any changes in rankings as a result. This is one to approach slowly.

Not sure how you have access (are they paid?), but by changing a large amount of links at the same time it may alert Google to some suspicious activity. Large changes in anchor text can be a sign of changes to bought links.

Personally, I wouldn't change them all either. Even though the year is wrong, there's still going to be benefits. e.g. if you are trying to rank for "holiday in new york 2013", your rankings will still benefit from keywords saying "holiday in new york 2012".

Take it steady and keep reviewing the results, but bear in mind you may be better focusing your efforts on building new links with new anchor text.

Hope that helps,

Ben