Enterprise Multi Site Migration Questions

[steaprok]

We are currently undertaking a fairly complicated site migration, and are looking for thoughts, ideas, pointers, takeaways, findings from the great minds at WebmasterWorld.

I would love to get a thoughtful conversation going on this topic and see what kind of experience others have had on similar projects. Thanks.

We have 3 websites, each with 1-3K+ pages indexed in Google, at the directive of the executive team, they must be migrated into a 4th existing website with about 2K pages indexed.

Each of these sites have a decent amount of organic traffic, rankings and are leaders within their submarket.

A Little Background

This is a larger company, let’s call it Company A, which went on a buying spree and bought up several of the other smaller market leaders across it’s submarkets. Company A is also a leader in their main market, but they offer 5 different products/services, whereas companies B,.C and D carry one main product that falls within one of the 5 products sold by company A.

These 3 other companies they bought, have some overlap with company A’s existing keywords/services, but there are also some service branches that do not overlap and are completely different.

(Company A) bought Company (B, C and D)

(Company A) after 6-12 months of messaging and cross branding on the sites, is ready to pull the plug on the 3 sites for each (Company B, C and D)

Additional Aggravating Factors:

1) Company A recently changed their domain as well. (3-4 months ago) from exampledomainsystems.com to exampledomain.com. Thankfully, there has been less than 5-10% loss of organic traffic, but they are seeing a lot of traffic from the old domains keywords and losing some links.

2) We are up against a hard deadline, and a hard directive from CEO. Whatever we recommend, whatever we find. We HAVE to migrate all 3 sites into site A, even if it takes just taking them down completely and doing a catch all redirect sending everything to the homepage of the website of Company A.

3) There are a lot of internal politics at play here as well. We are dealing with 4 different teams, each with different priorities, and each pulling to save ALL the traffic on their sites (B, C, D)
Then you have the team from company (A) trying to drill the fact that sites (B, C, D), make up only 10% of the global picture, and are not that important. More important they say, is maintaining the integrity of the main site, site A, which is now going to be home for all sites A, B, C, D.

4) Each site has at least a couple of hundred pages living within subdomains. From support PDF’s, to old content, to blogs.

I hope I haven't lost you yet! If you have read this far, you're probably the type of person I want to hear from. Keep reading ….


Data we Have Pulled

Complete crawl of all 4 sites using Screamingfrog and Deepcrawl.

Identified all existing :

HTML Errors
Crawl Errors
Meta and Semantic Errors (Title, Desc, H1’s, etc)
Existing redirects (301, 302, etc)
Existing 404’s and other errors. (40X, Time outs, etc)
Page Speed

We do not intend on trying to fix ALL errors pre migration, only those related to migration i.e. redirects, 404’s, page speed, etc. Any pages missing meta data or H1’s, etc on sites B. C, D will not be fixed post migration. (thoughts?)

Once migration is complete, and traffic/rankings settle, we will do a post live SEO audit and recrawl and identify all remaining crawl/html/semantic errors.

We also used Majestic SEO to pull link data for each domain

# of links and value per URL
Types of links (image, text, etc)
Anchor text breakdown
Top referring domains and IP’s
etc..

In addition to using this link data for selecting the high priority, highest value pages, we can also use this data for outreach to update important valuable links and anchor text.


Traffic data using GA and GWMT

We pulled GWMT search queries and landing page data Jan-April 2014

We pulled GA organic keyword and landing page data YTD 2014

We looked for overlap across organic ranking keywords, to identify potential targets on the target site that can be used for existing content.

Not all of the sites had goals set up in GA, but the ones that did we are also using that data at the KW and URl level to identify top pages and URLs.

Additional Baseline Data:

We pulled pages indexed data from GWMT and also scraped all of the indexed URLs from Google for each site and subdomain.

We pulled crawl stats for each site in GWMT

We created visual sitemaps of each site using powermapper, which really helps visualize the site structure and layout, when mapping the content across the domains and understanding the overall size and scope of each site.

What we are doing next:

We are currently working on categorizing and grouping all of the URLS and starting to flag all High Priority URLs to identify where they will go on site A.

High Priority URLs = URls w/ decent links, traffic and ranking for some variation of target keywords.

All High Priority URLs must be mapped to pages with same content and page structure. Site A will create subdirectory for Site B exampledomaina.com/siteb, which will serve as a new home page for site B with its own sub navigation and links to important subpages, similiar to structure from old site.

Submitting XML Sitemaps prior to migration on all 4 sites and then post migration, we will submit a new master sitemap for site A, with all of the new subdirectories and pages.


Make recommendations on new site sections for site A, to house new pages for each site(section B, to contain all of the high priority pages from old site B)

Ensure ALL URLs are redirected to the most relevant URL

We also have a team on stand by during the migration: SEO, Devs, PR and Internal decision makers.

This by no means is the complete list of everything we are doing, but I hope it can get the discussion going.

I would love to hear your war/horror/success stories around site migration.

Thanks again.

[rish3]

I would invest a bit of time in some central place to keep all the redirect rules, with a simple way to distribute them out. Otherwise it's difficult to get a grasp on the big picture as things are patched up, changes, etc.

Doesn't have to be an application per se. A mysql table with (host_to_push_to,redirect rule,order) + mysqladmin + a script to push it all out might suffice.

Other things:

• Serve up "410 Gone" instead of "404 Not Found" for urls you want to drop out of the index quickly. Combine with Google's url removal tool in GWT if you need things gone really quickly.
  
• If you're going to use Apache rewrite rules, remember that they don't, by default, pay attention to url parameters. You'll need RewriteCond %{QUERY_STRING} ... to handle that.
  
• Skip, or entirely remove, any existing www->non-www (or vice-versa) rewrites on domains that will end up pointing entirely at the new site. Then you avoid two layers of redirects
  
• If you have access to curl, it's handy to see if the redirects are happening correctly:

  curl -sIL http://somedomain.tld

  (shows just the headers, so you can see the HTTP code (e.g: 301/302, etc) and any Location: headers.)
  

[steaprok]

awesome thanks for the replay, useful info .

[steaprok]

Anyone else have feedback thoughts?

[netmeg]

(You'll probably get more answers during the week; things tend to be somewhat slower on the weekends)

[Robert Charlton]

steaprok... welcome to active posting here after a long while, and thanks for your very clear overview. Yes, it is a weekend, but here's a rushed question that perhaps could be clarified....

We are dealing with 4 different teams, each with different priorities, and each pulling to save ALL the traffic on their sites (B, C, D)...

Then you have the team from company (A) trying to drill the fact that sites (B, C, D), make up only 10% of the global picture, and are not that important....

This can get into thorny issues. I've seen B, C, D sites get decimated in such a merger. What's important, I think, is to preserve a navigation structure that allows the unique service of B, C, D to remain visible within the overall structure of the site.

For now, let me ask whether B, C, D are known brands. In situations I've seen, they were. How are you going to be dealing with maintaining these identities... and how much of the original site teams and branding are being retained?

Also, how far off is that deadline you mention? I should add that the deadline is unfortunate, because the best way to do this is to do it gradually.

[steaprok]

@robert

thanks for the reply on the weekend.

To your first point, yes that is our main concern, it seems the team for A, is of the attitude that the other sites are not important, and only make 10% of their revenue. They have made it clear there would be no tears shed by them if B, C and D were decimated. Problem is site B has some really good rankings, and even out ranks site A on some really good keywords. And they have had alot of upward momentum in the last months, going from 150 top 10 keywords to 237.

With that said, and to your other question. Yes they are known brands within their market, all of them. With A being the largest, most well known. But B not to far behind in its sub market. C and D are both well known as well, just to a lesser extent.


The teams for B, C and D are being retained, and slowly folded into the main company A team throughout the next year. Also, to add some context, Just to add another level of difficulty, the team members from site A have final say! They can override the requests of the members of B, C and D, which are still trying to look out for their baby (aka the site and company they sold to site A. That is why I said they are all pulling in their own direction. We have to navigate this minefield and make the sale to team A as to why they should keep certain URLs, etc.

We are also working with a PR agency and company A abd B internal team on customer messaging continuity issues, as well as Branding concerns. They have been using email newsletters and social for messaging about the new migration

Since the purchases, sites B, C and D have had their existing sites slowly converted over to site A branding. So it will not be a complete shock, but still I am sure we can use all the help we get, as there will be some confusion.

The PR team is monitoring social for any feedback regarding the brands and UX or site issues for the following 3-6 months post migration.

And an additional post launch team to monitor the metrics we pulled for baseline I mentioned above.

To your last question, this is another problem. Deadline is looming. CEO gave 6/15 as hard deadline to either implement this migration plan or just bulk redirect everything to site A home (aka what we call the nuclear option) - NOT Recommended

Agreed on the gradual recommendation. We made the same on when we were discussing the project during the discovery process. It was shot down before it even got discussed.
6/15 is latest we could push it.

I am well aware of the risks here, and have expressed them clearly to the client, and am planning on covering the risks in depth during our migration plan presentation again to keep it fresh in their minds, I just want to make sure I cover all bases, and do everything we can to mitigate the inevitable drop in organic traffic for B, C and D

Thanks for your questions and input.



PS hey @netmeg long time no see.

[RedBar]

I'm not being rude however does anyone have a clue what is supposedly going on?

I know precisely what I would be telling everyone however, reading between the lines, you would need my sized...reality:-)

[Robert Charlton]

steaprok - It does sound like there is a dysfunctional internal battle going on. It's surprising that they've kept B, C, D personnel on at all.

Regarding the SEO part of it, which is all we can deal with here... I'd restructure the site to follow up whatever product divisions are agreed upon, and to allocate link juice flow in the new site structure so it's more or less proportional to anticipated combined market share. This both with regard to redirects and to the new structure of site A.

In some acquisitions I've been close to, the acquired brands retained their identities... and in others their branding was completely lost. I'm remembering a situation in which there was keyword overlap at the broad keyword level. I was consulting for the new parent company, which wanted to make good use of the acquisition; and I suggested that a differentiation for the acquired product line could be accomplished by adding modifiers to make the acquisition a new sub-niche, slightly deeper in the site, with its own branded landing page.

What was important was to provide enough exposure for the major sub-niche products (or product categories) from that landing page, and also to integrate them elsewhere into the site structure. These were very high-ticket items, and this plan was consistent with why the line had been acquired, and the differentiation was meaningful. It worked out very well... enhancing the existing targeting, and providing more revenue for the new line and for the new parent company.

I've been in other situations where sites I'd built (and the companies my employers had built) were absorbed thoughtlessly into a large company that obviously didn't have a clue about what they had gotten. Similarly, I've seen companies change sites with changes of administration, even where the old site is what put them on the map. In general, you're not going to win those battles.

As much as possible, I think you need to clarify the situation and to optimize within it, and not try to change it at this late date.