Process for splitting a site

[hazmat]

Hi all,

Background:
I am in the process of splitting off a large site. Large sections of the site have grown out of hand with content, and diluting the more productive areas of the site.
There is a lot of potential for these other areas, and splitting these sections to separate domains seems to make the most sense.

Questions:
What would the best general plan for splitting this be? We will most likely need three or four different additional sites, and development time and follow up content creation will be fairly huge.

I am considering noindexing or using robots.txt for these existing sections in the meantime, is there any harm in that. Likely to be flagged by G in any way?

Thanks.

[Planet13]

I think that most people would block the NEW sites from bots until they are ready first, and when the new sites are ready, then "throw the switch," meaning:

- Un-block the bots from accessing the new sites

- 301 the old pages to the new pages

since it sounds like there is a large development time, you should probably work on one new domain at a time?

If you really need to do something now, maybe you could just sort of reduce the internal linking to those sections which you are going to delete? Maybe don't link to them from the home page or the other major sections of your site?

Hopefully, others will chime in with better ideas than mine.

[aakk9999]

I am doing the same at the moment, removing one section of the site into a new self-standing site. Planet13 got it right on the high level.

- We blocked the new domain that will be a new home for the sawn-off section from everyone by IP filtering.

- Then we developed the new site. The site structure is not exactly the same as in the process we have merged some pages and split the others and have decided not to move some (to let them die instead). We have also taken this opportunity to make URL structure better, i.e. we did not use the same URL patterns from root down as it is on the "mother" site. This has slight dissadvantage (and more work) with regards to creating 301 redirects, but we took a view that new URL structure will be better long term and is worth a hastle

- Transfer (copy over) the content (but still leave the content on the "mother" site). At this point we stopped creating new pages in the section to be transferred on the mother site.

- On the new site (blocked from the world), implement friendly URLs and test the site thoroughly whilst still blocked

- On the new site - fix the links in transferred content, that is, make sure all internal links from transferred content are correct - the links that link within the section to point to new URLs / new pages (if URL structure or page structure(s) have changed) as well as change the links that on the "mother" site are going to other sections and which in the new site will now be linking out to "mother site"

- Prepare the list of URLs that will redirect on the page to page basis, make sure redirects go to correct places where the pages were merged/split

- Prepare the list of all internal links on the "mother" site that go to the section that will be hosted as a new site with the corresponding link which will be replaced once new section goes live

- Prepare the list of all URLs from the section on the "mother" site that currently redirect - as we will want to change these redirects to go directly to the new site URLs to avoid chain redirects

- Prepare the list of all URLs from the to be sawn-off section that will not be transferred to the new site (e.g. thin content URLs) as they will be set to return 404/410

- Prepare the list of inbound links that link to pages in the section to be sawn-off - these sites will be contacted later on in attempt to have them to change URL to point directly to the new site

- Prepare new sitemaps for both sites, the "mother" site without sawn-off section, and for the new site

- Prepare and upload robots.txt of the new site

WHEN YOU ARE READY TO SWITCH:

- Unblock the new site (remove IP filtering)

- On the "mother" site, switch off the new section from the navigation menus

- Upload the previously prepared list of 301 redirects to the "mother" site (so that all URLs from this section redirect to equivalent URL on the new site

- Upload/create 404/410 response codes for pages that are not being transferred from the "mother site" to the new site

- Comb the pages on the "mother" site and change all internal links to point directly to the new site URL (the previously prepared list should speed this process up). Note that there may be a small window whilst there is an internal link that redirects to the new site instead of linking straight through, but as soon as all internal links are changed, the links will go directly to the new site.

- Upload sitemaps to both sites

- Add the new site to Google Webmaster Tools

- Using a list prepared previously, contact sites that are linking to the "mother" site section that is sawn-off and ask them if they would change their links to link directly to the new site page

It would be more manageable if you do it section by section and you could also watch the impact on ranking of the "mother" site once the section has been removed, in case this section was supporting other parts of the website.

[Planet13]

This bears repeating:

- On the new site (blocked from the world), implement friendly URLs and test the site thoroughly whilst still blocked

Testing is a great thing. Test, test again, and test some more.

[hazmat]

Hi, thanks very much to both of you, that would have taken some time to write all that up. Really useful info there.

301ing on such a large scale will hopefully be ok, Probably about 100k pages, all of which is unique content/products. Funny I feel the need to specify that! I will probably test on one of my smaller sections 20k first, if possible.

[Planet13]

If the pages you will be moving have lots of incoming links, then of course, your main site (from which you will be removing the pages) will lose some page rank.

On the other hand, you can probably revisit the structure of your site so as to make the site more focused.

Hope this helps.