If you have been following me for a while then you will know that I am an avid split-tester. What this means is that whenever I create something that tries to get a prospect to take an action, I put up at least two variants to see which will outperform the other/s.

Conclusions:

I know that conclusions are supposed to go at the end of an article, but I wanted to get a few out of the way first:

1. Testing DOES most definitely work

In many occasions I have managed to double or even triple sales and optins. In a one step sales process this means that I have brought in double or triple the amount of sales. In a two step process, this means that I have been able to quadruple or multiply ‘sales’ by 9. Therefore if you are not already, then you must start testing.

2. It is important WHAT you test

I have tested certain things such as font size or text color. Although this may have increased conversions in certain cases, I have found that there are variables that when changed give a far greater increase. Examples of this are price, headlines, order-buttons, time-delaying an order button and a totally different salesletter or video concept. Therefore, if you are going to test, then it is definitely important WHAT you test.

Some Unmentioned Results

There are certain things that I have found now that I am doing serious testing that I haven’t really heard spoken about before but are incredibly important:

1. Generally, when testing you need a large amount of data to make sense of it. If you are testing 4 variations of a salesletter and receive 50 sales, then depending on the data, in many cases the data will be almost meaningless. For many people, receiving just 50 sales is a lot of sales and therefore it will be very difficult to make any conclusions. For example, if I release a new product to my list then making even 100 sales will not give me an conclusive scientific data. It then comes down to acting on hunches until you get accurate data, which does kind of go against the whole point of testing.

One of the problems of testing is that although you are aiming to increase your sales conversion, while you are doing that, you are also losing sales as you are showing variants to prospects which are not resulting in the optimum level of ‘sales’. Therefore, it is important that you lose the losers and keep the winners as quickly as possible. One way of doing this is skewing the ‘pages’ that you show according to what is winning at that point. Google Optimizer does take care of this automatically for you if you wish.

2. Prospects View Your ‘Page’ Multiple Times

(I have never seen anyone else write about this before, so please do drop me a comment below if you have)

In a typical internet marketing launch, this may not matter so much. However, on salesvideos from other niches, I have found something very interesting happening.

I had one ‘page’ which was performing for a few months at (let’s say) 1% conversion – therefore 1 in every 100 customers was buying. I thought that I may be able to increase that, so created a second video and split-tested the two videos. I found that the 2nd video outperformed the first video so started showing that one twice for every time the first one was shown. After a few months, I found that the conversion on the two videos was almost equal. So, I added a third video and found that that was now doubling the current conversion rates of the first two. The first three videos steadily converted equally, so I added a 4th and the same thing happened.

So, what I was finding was, that each new variant that I added performed really well and then suddenly reduced in conversions to approximately the same as the previous videos. My personal conclusion was that visitors were visiting the site many times and would only watch the video if it didn’t seem familiar to them. Therefore, each time I put a new video on the site, it performed really well and made ‘sales’ but over time, once visitors had seen it and not bought they would not watch it again.

The solution of course would be to add a new video every month or so or to rotate the videos. However, I did wonder if you had noticed this phenomenon and if you had any other solutions.

I found two solutions to this ‘problem’:

1. Create a new variant of the video every month or so which means that it will always appear fresh to visitors.

2. Rotate a number of videos, cookieing each visitor to make sure that they will never see the same (fairly different) video in a row

Would love to hear your comments on this or anything else in the article, particularly if you are testing yourself and any interesting results that you have had. Please DO leave a comment below.
Keeping it real in an unreal world,

Ben Shaffer