Fixing a duplicate content issue

[travelin cat]

We have a new client that previously used a company that really messed up their on page seo as well as many other things. The main site, example.com had two other sites called example1.com and example2.com. Apparently the original strategy was to create content on the 1 & 2 domains and point them both to example.com. They also had the main site link to the two other sites. Most of the content was identical and they got smacked hard by Google a while ago.

My main question is this, should we just completely remove websites 1 & 2 with no trace? Or should we point example1.com and example2.com to example.com? Or perhaps something else?

[1script]

If it was literally identical, you cannot do any more harm by 301-in the other 2 locations of that content into one canonical one. Perhaps identifying some really problematic URLs (destinations of bad links mostly) in the process and 410-ing those instead.
When you say "got smacked", all of them? That might be an indication of something other than duplication. If these three could be easily identified as one entity (by same AdSense or Analytics code for example), then they should have been able to pick one and make it canonical. It might not have been the one that YOU wanted but there would still be one site standing if duplication was the only issue.

[levo]

I think the fastest way to remove/nuke pages from Google index is to remove A (or CNAME) record from DNS. I would choose removing over redirecting.

[Planet13]

Hi ya, travelin_cat:

I would tend to remove the pages that were on example1 and example2 having them return 404 / 410 with a custom Not Found page that has links to the most important sections on the original site. It could even have a search function for the original site.

The exception would be pages that have good NATURAL inbound links, which I would 301 to the appropriate page on the original site.

[Robert Charlton]

I wouldn't 301 anything from example1 & 2 to example.com. It's such an obviously spammy setup that you want to completely disengage example.com from that strategy and those sites.

I'd have to give it some more thought about whether I'd go with levo's or Planet13's approach for removal... it should be either total removal, or 410s, IMO... or eventually both.

Just thinking out loud... perhaps having Google see all outbound link pages removed from example1 and example2 before you take those sites down lets Google see a sequence of operations that suggests conscious removal. It's hard to say what they'd make of the A records vanishing or how fast the removal needs to be. Check dates on links and spidering dates if its easy.

Look at other inbounds to the main site as well. Obviously you want to get rid of the outbound links to the two other sites... and maybe check to see whether there might be some link circles involving yet other sites that comes back around.

Check also where all of these sites are hosted. If all on related IPs, that makes the relationships unmistakable.

You've also got a big content rebuilding project to deal with.