Posted: September 16th, 2008 | Author: Mihai Bojin | Filed under: Google, SEO, Search engines | Tags: domain, fast inclusion, Google, htaccess, redirect, SEO, SERP | 1 Comment »
For those who don’t know what redirection is, there are a lot of helpful resources out there; try Wikipedia for example.
Google published a redirection best practices tutorial in April about moving pages around. Let me sum it up in a short list:
- Use 301 redirects; start small with a couple of pages in the begining and after you check that it works go further step by step until you redirect your old domain to the new one.
- Don’t do both a redesign and a full domain redirection at the same time
- Check on site and off site links, changing them accordingly to point to the new domain
- Keep control of your old domain for at least 180 days.
The trick here is to redirect traffic from an older domain to a new one without losing standings on the SERPs and traffic coming to your domain. It is harder than it seems, but not really that hard.
I tried this technique with one of my domains and now I’m sharing the results !
Everybody fears and avoids redirecting domains because of the most probable loss of traffic.
Let me tell you how I did it:
- I had a site with around 2500 unique visitors per week
- I wanted to change it’s domain name so I did a 1 on 1 redirect to the new domain; I did not proceed with caution 1 page at a time, one directory at a time; that’s because I knew what I was doing, the domain name was coded in my configuration script and once I changed it all of the site’s internal links were updated to the new ones.
- I did not move the site to a different server (in my opinion this is demonstrating to Google that you own both domains)
- I even left the site in the same directory on the server and I implemented the 1-1 redirect within the .htaccess file; it’s important to keep the old site registered with Google Analytics and also keep it’s sitemap right where it were; because I had the same directory for both old and new domains I had to create some custom htaccess rules:
- I didn’t contact the sites that had outbound links to the old domain; I just left them as is, mainly because they were not high quality links and partly because I was lazy
# Old files that need be kept on the domain
#Google webmaster tools verification file
RewriteCond %{REQUEST_URI} ^/googlea000011112222.html$ [OR]
#Sitemap file
RewriteCond %{REQUEST_URI} ^/smap.xml.gz$ [OR]
#Sitemap index file
RewriteCond %{REQUEST_URI} ^/smap_complete_idx.xml$
#Any other files that you want kept accesible on the old domain (without redirecting them) you will have to place
#above in a RewriteCond line ending with an [OR]
#Serve the requested URL on the old domain
RewriteRule (.*) $1 [L]
#www to non-www redirects and old to new domain redirect
#I wanted to use the non-www form for the new domain so I redirected all other forms to example-new.com
RewriteCond %{HTTP_HOST} ^www.example.com [OR]
RewriteCond %{HTTP_HOST} ^example.com [OR]
RewriteCond %{HTTP_HOST} ^www.example-new.com
RewriteRule (.*) http://example-new.com/$1 [R=301,L]
- The old domain was indexed in Google for about 9 months; the new one was bough one day prior to the redirect, so it was virgin: no links to it, no Google history etc
Okay, now come the facts:
- The old domain had around 2500 uniques per week (measured one week before the change) with 10,000 page views
- I pulled the redirection hard, not doing baby steps but Titanic sailing in Googly waters
- After around 4 hours of the initial redirection and listing of the new domain sitemap in Google Webmaster Tools, the site appears in Google results with around 500pages indexed. I was shocked ! It took 4 hours to add this new virgin domain to Google
- 6 weeks later, the new domain pulls in 2000-2100 unique visitors with 6000 page views per week
- In the meantime I lost control of the old domain, so that’s another thing on Google’s list I haven’t respected; right now there is only the new domain online
- In these 6 weeks, GoogleBot visited the new domain through redirects and added pages from the new site at an incredible rate; after this small timespam the new domain has around 40k pages indexed, while the old domain even if it doesn’t work anymore still retains 30k pages (it had around 70k before the redirection and Google was dropping around 2500pages per week but now it stopped dropping)
- I also have to note that after the first three weeks Google had in it’s index around 30k pages indexed for each domain
So what can we learn from this ?
- A full domain to domain redirection has costed me around 20% of the traffic which is neither little or big
- If you can be sure that the internal links are updated when you redirect, you can start redirecting all at once
- If you want to get a new domain indexed in Google, get an older one and redirect all traffic from it to valid pages on the new domain! It took 4 hours to get my new domain listed in Google!
- I guess that if you have enough traffic on a group of pages, you can pull this fast Google inclusion trick by only redirecting a part of an existing domain (maybe even for a small time). I wasn’t able to test this because I lost control of the old domain, but will test it with the next opportunity
- Also this seems a method of splitting the traffic from one domain into 2 domains. Because Google doesn’t delete a page from the old domain immediately after indexing it on the new domain, you could get away with more summed up total traffic on both domains; I will test this whenever I’ll get a chance
Finishing up, I am really curios of the new domains PageRank at the next TBPR update. I will post the findings. Be sure to check back.
If you have any information that you’re willing to share on this subject, please comment on this post !
Posted: September 12th, 2008 | Author: Mihai Bojin | Filed under: Google, SEO, Search engines | Tags: content, direct traffic, domain, facts, Google, incoming links, links, PageRank, PR, refering sites, search engine traffic, SEO, TBPR | No Comments »
I have been working like crazy in the past few days and I wasn’t able to write anything… until now.
Alright, so I’ve got two domains that were launched almost at the same time 1 month apart. Fresh domains. No links pointing to them. PageRank of 0(zero).
I took the same steps of adding these sites to Google and launching them (more on that on another post).
In a future article I will write about Google Webmaster Tools, Sitemap submission, Preffered domain form and other basic SEO stuff.
One domain, let’s call it A, is an online portfolio website with a few about us pages, some contact information, privacy policy and a couple of portfolio items listed on one page.
The other one, let’s call this one B, is also some kind of a portfolio site, but for drawings. It gets daily updates with one new drawing per day, the only SE recognizable text on each page beeing the image’s title and date of posting.
Let’s sum up some facts about the 2 domains:
| Domain A |
Domain B
|
| Launched on June 18 2008 |
Launched on May 13 2008 |
| Almost the same domain age |
Almost the same domain age (12 day older) |
| Exact same steps towards submitting in Google |
Exact same steps towards submitting in Google |
| Added to CSS galleries prior to any page appearing in Google’s index |
Added to CSS galleries prior to any page appearing in Google’s index |
| Time it took the site to be indexed in Google: around 5 days |
Time it took the site to be indexed in Google: around 5 days |
|
Statistics after the first month:
- 82,132 page views / 28,438 visits = 2.89 pages/visit
- Bounce rate: 43.88%
- Avg time on site: 2:07 minutes
- New visits: 83.99%
- Traffic breakdown: 81.27% refering / 18.23% direct / 0.51% search engines (only Google)
|
Statistics after the first month:
- 8,869 page views / 4,617 visits = 1.92 pages/visit
- Bounce rate: 66.58%
- Avg time on site: 1:34 minutes
- New visits: 84.75%
- Traffic breakdown:86.20% refering / 13.17% direct / 0.63% search engines (mostly Google: ~99.99%)
|
|
At the time of writing:
- Domain A has 4 links pointing to it (that pass PageRank)
- Domain A has 12 pages indexed in Google
|
At the time of writing:
- Domain B has 43 links pointing to it (that pass PageRank)
- Domain B has 19 pages indexed in Google
|
| Uses the non-www form |
Uses the non-www form |
| dot.com domain |
dot.com domain |
| Site has 27.86% code to text ratio (higher is better) |
Site has 23.78% code to text ratio |
| Site is XHTML Transitional 1.1/ CSS 2.1 Valid |
Site is XHTML Transitional 1.1 Valid |
| Basic SEO facts were taken into consideration and implemented (page title, metas, alts, title attributes, keywords on page, h1,h2,h3 tags etc) |
Basic SEO facts were taken into consideration and implemented, but because the lack of text on the pages, all that is written seems like keywords to the SEs. |
| Site content did not / does not change at all (at least until the TBPR in July) |
Site was updated daily from May 13 until July 2 (but the content of a post on the site is 1 image / 1 image title / date of posting). |
| Site content did not change since launch (all pages remained the same, no SEO tweaking, no new links) |
Site content did not change since July 02 2008. Like with the A domain, there was no SEO involved during this period. |
Along came the July 24 TBPR (ToolBar PageRank) update, which was mostly fixing penalties given on the January TBPR update to directories and link farms.
So guess what:
Domain A gets a PageRank of 1 (before the TBPR it had zero);
Domain B gets a PageRank of 4 (before the TBPR it had 0) !! This is the site with no-text in it; it seemed strange at first.
At the time of the writing (that’s in September 2008), the domains keep their above PageRanks (no TBPR update since July), but:
Domain A (PR=1) gets 2576 visits in September(days 1-12), of which 41 came from Google searches;
Domain B(PR=4) gets 294 visits in September(days 1-12), of which 2 came from Google searches;
So let’s do some simple math:
Domain A: 0.015% visits came from Google (PR 1 site)
Domain B: 0.006% visits came from Google (PR 4 site)
Which in my mind means that PR is not directly related to the traffic on the site, nor that the traffic a site gets from Google is directly related to it’s direct traffic, a metric that could be measured with the Google Toolbar.
I can draw some conclusions at this point:
- PageRank is not directly related to the traffic on your site (bigger PR does not mean more traffic hence more visits); this is somewhat logical because the traffic on your site depends on all 3 primary sources(direct/search engines/refering sites) and also on the topic of your site;
- PageRank doesn’t seem to depend on any of the most important SEO guidelines that you can implement in your sites and that you probably know of from all the blogs/sites/books etc you’ve read them in;
- PageRank does not seem to be directly influenced by the number of incoming links (this is more of a hunch because I didn’t have an astounding difference of incoming links between the 2 sites);
- PageRank DOES seem to be influenced by the quantity (and most certainly quality) of new, fresh and unique content on the site (even if the content consists of IMAGES). That means that you can get a higher PR from adding any kind of content as long as it’s original.
That’s about it ! Of course I have worked with a relatively small set of data in a short time span, but maybe the two most important, logical and well-known facts are:
- CONTENT IS KING ! Add more content, be original, have periodicity (I’ve got to learn this one
); add as much content as you can, but not too much. If you have 10 articles today and only 2 tomorrow, don’t add 10 today ! Add 6 or 7 and keep the other ones for tomorrow; this of course does not apply on news.
- LINKS ! Get as much incoming links as you can, but like with content, not all at once; some of them will count towards your pages’ PR, some of them won’t. Incoming links from pages/sites with related topics to your own count MORE ! DON’T SPAM general topic forums; it’s for your own good.
If you have knowledge or information on this subject, please share it with the rest of us ! Also I can provide more detailed analytics data to anyone that is interested.
Care to comment?!