But Google is always trying to ‘improve the search experience’. One way that the Google tries to do this is by detecting the geographical location of the searcher and then providing local results. So if Google thinks you are searching from the UK, then you will get predominantly UK results, even if you choose to use google.com.

So if you are promoting a site in the US for example, how do you see how your site is ranking and who is outranking your site. I have recently come across a Firefox plug in that allows searches to be done as though from different geographical locations. It’s from www.redflymarketing.com but although I had no difficulty in accessing the site and downloading the add-on yesterday, it is unavailable today. For an alternative location, try https://addons.mozilla.org/en-US/firefox/addon/5977 .

Once downloaded, simply do a search in Google.com in the normal way, then right click and choose to run the search for the UK, USA, Canada, Ireland or Australia. Really useful.

A/B and multivariate testing was this was the subject of the WAW (Web analytics Wednesday) meeting in London last Wednesday, 15^th October, and very well supported is was too with about 170 people. This was a particularly educational WAW with three excellent speakers.

First was Mark Simpson from Maxymiser. He was bombarded with questions showing the interest in the room for his overview of the subject.

Ewald Hoppen from Wehkamp in the Netherlands really sold the value of multivariate testing by giving real life examples of how they have tested changes both big and small and the sort of changes in conversion rates that resulted. For this large online retailer, even small percentage changes can make huge differences in term of euros. They also found that they get different results depending on the time of day and the day of the week, so it is not just a question of delivering the most effective pages but delivering the most effective pages for larks, owls and ordinary daytime visitors.

The final speaker was the one and only Eric Petersen, he presented his Top Tips for Testing Successfully. Although The Big Book of KPIs is always on my desk, this is the first time I had seen Eric speak. By 7pm in the evening and for a talk of web analytics, it takes some doing to really engage your audience but he certainly did that.

WAW have been running in London for several years now and the next on is January 21^st 2009, sponsored as always by SCL Analytics. Well worth attending.

See the Atracks article on using Google’s free A/B and multivariate testing tool, Website Optimizer for more information on how to start testing and improving web pages.

]]> http://atracks.wordpress.com/2008/10/17/ab-and-multivariant-testing/feed/ 1 Atracks http://atracks.wordpress.com/2008/09/08/google-insights/ http://atracks.wordpress.com/2008/09/08/google-insights/#comments Mon, 08 Sep 2008 11:35:38 +0000 atracks http://atracks.wordpress.com/?p=24 ]]> Google Insights is really natty new tool from Google.  Choosing which keywords to optimise for is fundamental to promoting your site and this tool is helpful addition to the keyword research armoury.

Choose the UK filter to get results for the UK only and then type in one or more search terms into the search bar.  You won’t get numbers of searches back but where Google Insights really wins is that you can compare different search phrases in terms of search volume and as to whether a particular phrase is on the up or in decline.

For example, Atracks offers products and services related to search engine optimisation and web analytics. Typing each of these terms into the search bar, separated by a comma, shows that although SEO is still the more popular search phrase, it is on the decline whereas web analytics is steadily gaining ground and is now getting something like 60% of the searches that SEO does.

Try it out with two different products you offer and use the information to plan your marketing activity next year!  On and offline.

Using Google Insights to compare ‘plasma tv’s’ and ‘digital radio’ shows how powerful this tool can be.  The seasonal behaviour in tv sales is really evident, whereas it seems that people are looking for digital radios all the year around.

A further useful function of Google Insights is the top searches and rising searches information below the fold.  This offers real gems of suggestions for related terms and again helps you focus on what is in the ascendancy – and keeping ahead of the game is what it is all about.

Part of the purpose of Insights is to give demographic data about searches but as with so many of these tools works much better in the US than the UK.

Google Insights is a development of Google Trends, already a useful tool itself.  There is already more than enough internet related information out in the ether but we think this is going to prove a really benefitical addition.

For example, on the home page, there might be three links to product1, one on the nav bar, one from an image which says ‘Product of the week’ and one in the body copy extolling product1′s virtues.

So how do you track which link gets the most traffic?

One way is simply to add a parameter to the link.

The links from the home page to Product1 page will look something like

<a href="/product1.html">Product 1</a>
 

you then add a tracking parameter to each so they look like

<a href="/product1.html?link=icon">Product 1</a>

< a href="/product1.html?link=navbar">Product 1</a>
< a href="/product1.html?link=body">Product 1</a>
 
 
The only problem with doing this is that Google then sees three different pages but the content is the same, ie duplicate content.
 
You get around this problem by adding the no follow tag to all but one of the links so they become

<a href="/product1.html?link=icon" rel=”no follow”>Product 1</a>

< a href="/product1.html?link=navbar" rel=”no follow”>>Product 1</a>
< a href="/product1.html?link=body">>Product 1</a>
 

Leave one of the links without the no follow tag as you want Google and the other search engine to follow the link.

When deciding which link to leave for Google to follow, remember that text links are more valuable than image links, and links in the body text more valuable than links in the navigation. But of course there are always practical considerations, the nav bar is probably controlled by the template which means you have to leave the nav bar links alone.

Looking at which links capture the most traffic can make a real difference when it comes to optimising your page layout and design.

Simply go to Google’s URL builder at

http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55578

or type Google’s URL builder into Google.

This takes you to a very easy to use tool which builds the code for you.

With the first client, it was clear something was wrong. Google Analytics should pull in the cost of campaigns very accurately – and for all my other clients this has been the case – but here GA was showing campaign costs approximately half what the Adwords campaigns were actually costing. Also there was a huge discrepancy between clicks and visits, with visits outstripping clicks by a factor of about three. I would certainly not expect clicks and visits to correlate but a factor of 10% is more usual than 300%!!

We looked at lots of different possible explanations but in the end it turns out that GA cannot always recognise the tracking codes built by the Google URL builder. Now this does seem extremely strange, but it seems that the automatic (hidden) tracking that GA uses to pick up Adwords data is more robust than the manual system used by the URL builder.

Removing the manual URL builder code has removed the errors with the cost data – the two correlate exactly now as they should. The PPC campaign has now paused temporarily and we are still getting a lot of visits (about 100 last week) from no clicks – either this is returning visitors from the previous PPC traffic or there is still some tracking error.

This is the first time I have had any issues like this with Adwords and GA not communicating perfectly, but it highlights the dangers of analytics – information is only as good as the quality of the data generating it and it is not always easy to pick up data collection errors!

Atracks web analytics and nso training

I am just about to move my website from one host to a new one and am worried I might lose my rankings in Google which are currently excellent. Is there any need for me to be concerned and is there is anything I should be doing to protect my good rankings?

In theory there is no reason why moving host should affect rankings at all – though there might be a slight dip if Google happens to come in during the short time that the site is actually being moved.
However, there are some problems that can arise if you are not careful.

The main one is that moving host often coincides with a site rebuild – perhaps the site needs to install a CMS (content management system) and sometimes hosting has to be moved to accommodate its requirements. Its when the site is rebuilt that the danger often arises. Rankings are for specific pages so if you have a No 1 ranking for an internal page, call for example, http://www.mydomain.co.uk/product.html then if this page no longer exists, the ranking will be lost. The home page address may also change, from .html to .asp for example.

There are then two options. Either the new site can be configured so that the page names are left the same – this might require the use of a mod rewrite (or similar). This is the best option if it is possible.

If this is not possible, then the new pages can be uploaded and a custom error page with a link to the new home page used so that any visitor, or bot, who lands on a ‘missing’ page will find a link to the new home page. Visitors and bots will then be able to find the new pages.

When page names are changed, use Google’s Webmaster Tools to submit a new XML sitemap to let Google know of the new pages as quickly as possible. Also, use the URL removal tool to remove the old pages from Google’s index. Then monitor the web crawl diagnostics on Webmaster Tools to check that everything is going well and no broken links have been detected.

Bear in mind though that the value of any incoming links to the old pages will be lost, which of course will affect rankings. You should be alright with links to the site that do not specify a page, ie of the type http://www.mydomain.co.uk

The worst case of loss of rankings that I have seen associated with a change of host was when a lot of images failed to be transferred correctly resulting in lots of broken links. That really did scupper rankings until it was sorted out and if it hadn’t been picked up fairly quickly using Webmaster Tools (it was not obvious from a quick visit to the site), it could have had a very long term effect.

Both were half day courses with most students choosing to attend both sessions.

Introduction to PPC advertising looked at the various options that Google provides when setting up a text based ad. Most importantly we concentrated on how to use this powerful medium to deliver cost effective results. By careful focusing of keywords, creative message, geographical distribution as well as targeting the right kind of online distribution, Adwords can be a very effective advertising medium. The danger is that without understanding how PPC works, it is very easy to end up spending a lot of money generating the wrong sort of traffic.

Introduction to Web Analytics in half a day is an ambitious undertaking! The enormous power of the little bit of tracking code on each page though is hard to under-estimate. The focus of this course was on taking away the problem of information overload – the biggest obstacle to successful analytics use – and looking at ways of using analytics information so that it can be done quickly, targeting what really matters in any particular situation and ensuring that Action results. Without Action, web analytics is a waste of time but with the intelligent and discerning interpretation of web data, the webmaster is able to optimise his site and out-perform the competition.

By running the two sessions on the same day, we were able to cover the analysis of PPC data in detail and in context. Running Google Analytics in tandem with Adwords enables the return on investment for different campaigns to be determined. But what is even more exciting is drilling down and looking at the ROI of different keywords and different creatives. This is the information that allows really effective, well targeted campaigns to be run which deliver to the bottom line.

Student comments included: “Great knowledge provided”, “Excellent day, will recommend”, “Excellent”, “This helps my business planning, I shall be outsourcing these functions but now when I outsource I will be able to give clear instructions and know what I sort of service I should be able to expect”

I have been intending to set up a blog on the Atracks website ever since it was launched last year (previously I had been part of Search Engine Workshops UK which had a very successful blog) but there is always more work than time and it hadn’t happen (the story of the cobbler’s shoes spring to mind but that is a different story).

Last Thursday, I listened to a webinar presented by Aweber, a company I use to send out email newsletters on behalf of clients. Much of what was being said was reinforcing what I already knew so as I was listening I set up a blog quickly on WordPress – and very easy it was too!

As I say I know blogs are powerful, but I admit to being a bit surprised when I received a Google Alert less than 72 hours later showing that Google had picked up the blog.

And the moral of this little story – blogs work.