Google Website Optimizer (Optimiser)

[Runner]

Sorry to raise this again. I have looked around but cannot find a complete answer. I see that people have asked this question before but have either not had an answer or the thread has been sidetracked onto something else.

I can only try... Is it possible and has anybody succesfully used Google Website Optimizer on their Actinic site?

[Runner]

Is there nobody using this on their Actinic site?

[Runner]

I'll try another approach.

Who's not using this and why not? Seems like a useful tool.

[Mike Hughes]

Who's not using this and why not?

In my case Time & Effort.

Too many other things to do.

Mike

[Darren B]

last time i looked it was not something you could intergrate, google checks the pages you list to make sure they exist, so when your comparing two pages and want to add the receipt page as the goal page you cant

[carbontwelve]

I'll try another approach.

Who's not using this and why not? Seems like a useful tool.

Im using it, unfortunatly its quite a complex task for someone who doesn't do behind the scenes technical computer things. However I shall describe the method to you. Actinic by design makes this difficult, many other software platforms make it very easy, just depends on how much hair you have left to pull out really.

1. Upload your actinic website
2. Using another ftp app (I use winSCP under windows or that which is built into osx) bring up the directory structure of your root web folder. You will see several folders and files.
3. Duplicate index.html and rename it as index2.html then open it in a text editor (double click within winSCP or right click and edit under osx) then make the nessessary changes.
4. Log in to the google experiments and give it the url for index.html and index2.html as the two pages for an AB test and set the rest of the settings up.
5. Sit back and "relax" until you get enough data to formulate an appropreate answer.

The AB test is the easiest to do, however it does require an indepth knowledge of html code and template editing from a text editor... although I presume as you know the benifits of SEO experimentation you are able to do the above. If not it will cost you money hiring someone to do so for you (one way in which actinic aids developers to milk their clients, something I am against).

[Runner]

Thanks for the reply Simon. I'll give it a go soon. Have you had any significant results yet?

[carbontwelve]

Indeed I have. The great thing about AB testing is you can fine tune your design to provide the greatest returns. While it is not surprising how much adding a more positive title to a button you wish people to click will improve your conversion, it is not the first thing designers tend to think about.

For example I have been developing online applications and website designs for the better part of 10 years however its only in the past four that I have begun to fine tune my way of working towards SEO and even now I have to ask myself if the design passes the "three year old test" by asking why three times.

A lot of the SEO performance comes from psychological analysis of your audience and developing a design to manipulate them. While there is always a two way interaction between your audience and the design, resulting in the intended meaning being interpreted in a variety of ways by your audience what SEO and conversion improvement is about is making sure that the meaning being interpreted is not too different between person to person.
For example no-body is going to deny that a symbol of a man on a door does not mean "men only". Subliminally we associate similar meaning to the way designs are laid out, an element with a lot of white space surrounding it will be deemed a lot more important by the audience than one stuck in a crowded area, people also get confused and impatient with designs that do not provide contrast to important information. Your buy button should be prominent, it is the most important thing on the page and it is what you want people to click.

Then again there is a fine balance to uphold and thus the AB testing, for example I wondered why over 70% of people where adding things to their basket but dropping out during checkout. After changing things around a bit and making the progression from stage to stage more convenient and obvious that dropped to 15%. A few small changes resulted in a huge performance gain.

The speed of your results is directly related to the amount of traffic you receive as you need a sample of at least 1000 visitors a week to gather data sample worth disseminating, in addition you cant go making too many changes at once, because while one change will give you a performance boost another at the same time may give you a performance reduction, resulting in skewed conclusions. Therefore if you have five ideas of what to change and only get 1000 visitors a week then it will take five weeks to collect all the data to provide you with accurate decision making.
This is where the Multivariate experiment becomes very useful, however I have yet to use it under actinic, as im used to content management systems written in php or asp that I can edit to provide different pages dynamically depending on which url is passed to it.

[kakshaks]

Good thread here on what to do:-
Google Optimizer in Actinic