Monday, December 22nd, 2008
If you sometimes need to search and replace some text throughout your weblog, you should definitely be using Urban Giraffe's Search Regex. With Search Regex you can search and replace text in all the fields shown in the picture below:
Search Regex
If you are replacing some text, first enter the Search pattern and press the Search button. This will show you the results, so you can fine tune the pattern to get only the results you want. You also get a set of links to view and edit the post. Very handy.
Read the rest of this entry »
By Martin
WordPress |
Friday, December 19th, 2008
We've just had to move another client's old site to a new one.
There are lots of inbound links but the page URL structure has completely changed for the better.
The client wants to rank right away.
What do we do?
301 the old site is the traditional answer.
Not so fast says Eric Ward who is one of the masters of link building, having built links by hand for longer than almost anyone else on the internet and for more large corporate clients than any individual I know (there are some SEO companies working fairly stealth with portfolios of almost 100 big names):
I wouldn't 301 it yet. First I'd run a backlink analysis on the old site and then visit each site linking to the old site, and for those that look exceptionally trustworthy and legit, ask them personally for a hand edit to change the link from the old site to the new site.
Painful.
Slow.
Tedious.
Effective.
Frankly for a website with thousands of backlinks, that's just not a realistic option. Well for Walt Disney or some of Eric's other clients perhaps it is. But should the rest of us do?
- Run a detailed backlink analysis (we use SEO Spyglass for this as it's reliable, comprehensive and cross-platform so that I can use the same software on my Mac as the rest of the team uses on their PC's or even Linux machines). That analysis will give you the target pages of all incoming links.
- Make a list of all the pages which have incoming back links and look at the anchor text for those links.
- Find the page on the new website which best corresponds to that anchor text (the new landing page so to speak).
- Write a 301 redirect for each of those old pages to new pages. 301 syntax looks like this:
redirect 301 /olddirectory/oldsubdirectory/oldfilename.htm http://newdomain.com/newdirectory/newpage
- Open up .htaccess and paste in the new 301 redirects (.htaccess is in the root directory of your website and is an invisible file - you need to turn on the option to see invisible files in your FTP client in order to work on .htaccess).
- Paste in the new 301's to your new website.
- Test your 301's by hand (always test everything by hand - a single colon or quotation mark out of place can disable an entire PHP file, html page or .htaccess file!).
- Monitor your server logs for 404's in any case. Any page which 404's often should also be 301'd.
For bonus points:
- Do a site:yourdomainname.com search in Google.
- Find the equivalent new page for each old page (depending on the site a regex redirect will be your friend).
- Write 301's for all the existing pages to the new equivalents (you can group pages of course, i.e. 4 different pages from the old site could get mapped to a single page on the new site).
- Add to .htaccess.
- Test.
If you follow these prescriptions to the letter, you should retain the rankings of the old site. Be prepared to see a three week to seven week dip as Google gets used to your new digs. It's a bit of manual drudgery in comparison to just a single global 301 to the new website.
Still it's a lot less work than trying to dig up the owners of incoming links to the old site and begging them to change the links to the new site.
But Eric has a point. Any incoming links over which you have control or are of particularly high value you should seek to change by hand (although an older link will lose its age value by being changed to the new domain, so what Google gives, Google taketh away). Over time, though, the direct link will be worth more.
By Alec
WordPress |
Tuesday, December 9th, 2008
We have some campaigns out there on Google AdWords for which we now have some very nice organic rankings.
At the end of each month we like to calculate the number of sales of PPC versus organic for a Canadian life insurance client.
Most of our PPC results have a gclid parameter in them so it's clear as day. It's almost certain that a clean result like this one is organic:
http://www.google.ca/search?q=canadian+life+insurance
While this is definitely PPC:
http://lsminsurance.ca/calculators/canada/term-life.php?PPC-ON-ripped&type=search&keyword=life%20insurance&adid=984186361&placement=&gclid=CLmk9Iruk5YCFQhdswodiCp_FA
What about this one with &rlz= in the URL parameter?
http://www.google.ca/search?sourceid=navclient&ie=UTF-8&rlz=1T4ADBS_enCA248CA248&q=greatwest
I couldn't tell. The long strange string smelled like PPC to me.
It out to be nothing of the kind. &rlz= is the string that Google uses for identifying users of Google Chrome:
You may notice a RLZ parameter in the URL when you do a Google search from the Google Chrome address bar. The RLZ parameter contains some encoded information (like when you downloaded Google Chrome and where you got it from). The RLZ parameter does not uniquely identify you nor is it used to target advertising. Google uses this information in aggregate to find out whether groups of people are using Google Chrome actively. Not all users have the same RLZ parameter. The RLZ parameter is based on where Google Chrome was download from, when it was installed, and when certain features were first used, like search.
A RLZ parameter is sent to Google with every search done using the built-in search box. It is also sent separately on days when Google Chrome has been used or when certain significant events occur such as a successful installation of Google Chrome. The RLZ parameter is stored in the registry and may be updated from time to time. The code that makes this work is not included in the open source project (http://www.chromium.org) because it only applies to the version of the browser that Google distributes, Google Chrome.
Great to know. These &rlz= searches are definitely organic then.
That Google would track you based on your browser is disturbing. However if you ever search logged in, Google can track you pretty well via your IP number, putting two and two together.
On the other hand, there are seven of us behind one IP here so the log might be a bit confusing. Still it could be correlated to exact user agent as many of my colleagues use different browsers as their primary web browser (or even a different OS).
I wonder why Google felt a need to go even further in their tracking. Just put away those images of Will Smith fleeing across bridges in Enemy of the State (blu-ray now too) and go back to your typing.
By Alec
SEO |
