Split Testing

[pinkequine]

I couldnt seem to find much on this subject, so posting to see whats possible.
As at first thought it seems like its not really possible to run effectively in Actinic.

What I'd really like to do is not to test just landing pages, but to test product pages as well. I could create a hidden page, but once they navigated off that page they wouldn't be able to get back to it (unless they pressed the back button which is unlikely if they take a look around the site first)

I could run two parallel sites, but this is a bit messy. Anyone got anything working with any kind of success?

[leehack]

Create a top level section for each split you want and then have inside that TLS, the parts you want that particular split to see. Point them to their TLS and they will move around just that part**

**Please note this advice only counts if I have grasped WTH you are on about.

[gabrielcrowe]

to what degree is the difference across the pages?

minor?
how about:

default

page.htm

url1

page.htm?v=variation1

url2

page.htm?v=variation2

in your page:

<script>
fullURL = parent.document.URL 
myquery = fullURL.substring(fullURL.indexOf('?')+3, fullURL.length)

if (myquery=='variation1'){
document.write ('this appears on page 1');
}

if (myquery=='variation2'){
document.write ('this appears on page 2');
}
</script>

please note, that i just made this up, and it probably does not work, but as long as you get the idea, then thats fine

[leehack]

Please note, that i just made this up, and it probably does not work, but as long as you get the idea, then thats fine

LOL G, that was very funny.

[johngwms]

No posts on this subject for a while.

Does anyone have more recent experience of Split testing with Actinic and its integration with Google Analytics?

I am planning on starting some split testing of our home page as part of a major upgrade. It would be nice if different variations of the home page could be shown in sequence and use GA to plot the performance of each.

John

[Duncan Rounding]

I'd also be very interested to hear of any success in this area.

[fergusw]

John at Stinkyink, and his team, wrote an interesting blog article on this very subject:
http://www.stinkyinkshop.co.uk/blog/...ite-optimizer/

It is getting on 2 years old, so some of the finer detail and instructions may not be the same, however it shows one way to do it!

[Duncan Rounding]

Useful thanks - I'm not sure how successful this was mind, Stinkyink has moved on since then.

[leehack]

Split testing on adwords seems more appropriate to me. As the web matures and people become more and more web savvy, i think the effectiveness of page split testing goes down relatively to the experience of users going up.

I'd liken this to have secure payment images with credit card images, if you don't take credit card payments nowadays, then what do you take, pounds of flesh? Showing an american express card logo or a paypal one has merit for sure. For me though, showing credit card logos is much the same as saying 'we are an online shop' i.e. darn bloody obvious nowadays.

[johngwms]

The problem is that everyone is different and a page designed to steer one user towards a sale may have a totally different affect upon another. It is even more difficult for a web designer to place themselves in the eyes and mind of a person browsing the web no matter how good they think they are.

Making relative judgements about the effectiveness of a page design through split testing is the only way of getting to understand the mindset of searchers in your site's market. Without such testing you are relying on instinct and guesswork.

John

[leehack]

I appreciate your point John, it was well put, but I do think you're a little off the mark. You seem to think you will able to cater for everyone, that's impossible, pleasing the masses is the very best you can do with a website. A good designer knows how to please the masses, if he doesn't know that, then he's not yet a good designer. Screen sizes, surfers age, browsers, web experience - all things different for every person.

At best a split test can give you an indication of how you can cater for the masses, a good designer can also do the same, he does it from experience though, not split testing, his last 7-8 years in the industry having worked on hundreds of sites has been his split testing.

[johngwms]

You misunderstand me - I agree that it impossible to cater for the masses.

That is the whole point of SEO for example - to optimise your site for your target audience. Yes general principles of design and layout gained through years of experience are helpful and often essential.

What was going through my mind was not specifically design but content. The value each user places on the words that are used varies considerably from one market segment to another. Keyword research can identify most of the best phrases to use but the style of text and its meaning are often only really understood by a person exposed to the same market that the web site is operating in.

I have been working in the electronics design business for the last 33 years both as a designer and now in marketing. That does not make me an expert in either and frequently discover where the holes in my knowledge show.

I have used three SEO companies to attempt to optimise my sites over the years and all have failed because they simply do not understand the market I work in.

It is for these reasons that I believe that split testing is important. To test both design and content.

John

[leehack]

That is the whole point of SEO for example - to optimise your site for your target audience. Yes general principles of design and layout gained through years of experience are helpful and often essential.

Poor design and great SEO = no sales
Poor SEO and great design = sales possible

SEO gets them there, design keeps them there.

What was going through my mind was not specifically design but content. The value each user places on the words that are used varies considerably from one market segment to another. Keyword research can identify most of the best phrases to use but the style of text and its meaning are often only really understood by a person exposed to the same market that the web site is operating in..

One of the biggest diifculties on a site is getting people to read at all, the language or style of writing is irrelevant when the majority buy on 4 basic principles - price, image, in stock, delivery - none of those involve any great reading.

I have been working in the electronics design business for the last 33 years both as a designer and now in marketing. That does not make me an expert in either and frequently discover where the holes in my knowledge show.

It's all too easy to over analyse the web, it's really not that complicated. I've been lucky enough to work on quite a few very successful sites now and i can tell you one thing, they are all almost identical in how they work. There is no special way of writing text that counts, SEO can in fact be quite poor, usability is always excellent, product images excellent (real biggie this one) and info is complete. Price is keen, delivery is fair and more important than anything, product and company service is great. Monitoring style of text in relation to how a user reads it, would be more akin to checking the effectiveness of a blog, your site would have at least 50 things more important than that to change. IMO you're not barking up the wrong tree with that, you're actually in the wrong forest.

I have used three SEO companies to attempt to optimise my sites over the years and all have failed because they simply do not understand the market I work in.

SEO companies are a minefield for sure, doesn't mean your 4th would be a wrong one though. Getting into the top 3 or on the first page of Google is well overrated though IMO with regards to the success achievable, when i look at the top sites i have ever seen, top spots on Google are not pre-requisites for success, in fact great SEO results can often be non-existent on these sites. Again i can only speak as i see, SEO is sometimes an overrated requirement forcing you into the shark tank, it often serves to detract you from the quick, instant, more profitable and time efficient success possible.

The product and the price at which you buy it, have been and will always be the two most important factors on the success of a site IMO.