Every guru marketer out there knows that they need to test their sales page, however those gurus also have a tremendous amount of information at their disposal – namely what’s worked for them over the last 10-15 years in the online world.
You might thing that I can simply “borrow” what they’ve done and it will work for me.
What may work for one marketer, may not work for another. You also need to remember that when a purchase decision is being made the reputation of the individual will play some sub-conscious role in deciding whether to make the purchase, or not. As well the history you have with that individual will also play a role. It is for that reason may marketers that teach others stress that you must build your reputation and win your readers/visitors heart before you will get access to their mind and be able to sell them anything.
Suppose you’ve done your homework you have a product that you know will sell and open up your sales page and product to the public. But nothing happens. The sales are just not there. What could be wrong?
Wrong colors? Wrong font type? Wrong graphics? Wrong words used in the sales page? Wrong headlines? Wrong sub-headlines?
Now multiply that with 2, 3 or even 4 values for each variable. Now what?
The key is to test but how can you test 8 variables, with a potential of 4 values each? That actually results in having over 65,000 pages of combinations. You’re not going to be doing that manually.
That’s where multivariate testing comes to play.
Multivariate testing allows you to test multiple variables on a web page with a virtually endless number of values.
For example, on a sign-up form – should you use an image or not? If you use an image, should it be male or female? If it’s male, should they be dressed casually, or more business attire? Should the person be older or younger?
Already from one simple item on a web page we’ve found so many different combinations to test.
Multivariate testing will allow you to test these minor differences easily and best of all – automatically.
How Does Multivariate Testing Work On Web Pages
At a high level, this is how it would work but of course each vendor of the software will have their own method of implementation.
You would set-up the various variable you want to test: headings, heading colors, heading fonts, heading sizes, images, no image, etc., As visitors come to your site they may or may not choose to purchase – the software will record their responses (or no response) and self adjust. Here is a quick example of how this would work.
Taking the sign-up form that I spoke about above:
Bold headline with a male image dressed casually and in their 40′s – might result in 10 clicks to subscribe to your newsletter
Bold headline, in red with a male image dressed casually and in their 20′s – might result in 40 clicks to subscribe to your newsletter
Bold headline, in blue with male image dressed casually and in their 20′s – might result in 30 clicks to subscribe to your newsletter
Bold headline, in black with male image dressed in business attire and in their 30′s – might result in 70 clicks to subscribe to your newsletter
Each element is tested and retested and with the results you can then determine which is the best combination to use that generates the most responses.
There is an excellent overview and discussion on multivariate testing over at Marketing Experiments that I would suggest you read. It will give you a lot of insite into multivariate testing.
Now the one thing that I have not done is given you links to software – the goal of this post isn’t to sell you something but to give you information about multivariate testing. There are several types of software out there that will help you with multivariate testing, ranging in price from $300.00 to thousands. The $300.00 software is available from James Brausch (I haven’t mentioned his name on the blog in some time — not that he cares, I’m sure) and I did buy it some time ago however never put it through its paces. The name of the software is called MUVAR and you can find it on http://www.muvar.com.
James may not support this product any longer, because from his main blog he offers another program called TestiVar for $150.00 > http://www.businesshqr.com/prd/1004/?sku=1004. Again I don’t know how well it works, as I’ve not used it.
Other products & sites that I’ve read about:
Alot of thes products are very expensive.
In all honesty James’ product has had a lot of recommendations and it is priced the lowest — which doesn’t mean it doesn’t work. From many of his blog posts he says that he uses his own product on his own sites.
This post is not a paid advertisement for James or his product, nor do I receive any commission from the sale of the product if you purchase from the links here.