Testing YOUR Way to Success
 

The Numbers Tells It All...

New Easy Way to Test Based Upon Website Design

I just looked at a video by a very successful guru and I think he may have a good idea for easier testing. He plans to launch a new product in January to add to his already large set of offerings. I will definitely check it out and see what it looks like. His discussion was on determining where a problem was developing and could be resolved pretty easily by having the proper layout of your website. He starts with an "opt-in" page, then creates a "sales page", next  comes the "order" page, and last comes the "thank you" page.

YOU Should Know What YOU Expect From Sales

If it's not what YOU expect, what then?

It seems that by knowing just 4 numbers you could determine the exact location of your problem. If not, then look at the number of unique visitors visiting your opt-in page, is there enough to expect desired results? If not, the opt-in page needs work or new methods of getting traffic should be explored. If that number looks ok, then look at the sales page and see if a large enough number of prospects are visiting it. If not, then the sales page needs work. Now, look at the order page conversions, you need enough orders to reach most of your sales goals. If not, sales page needs more work. The thank you page should also be used to promote additional products or upgrades and increase sales. If you don't see enough additional sales from this page it should be reworked as well.

Testing must be done, but some methods are more difficult than others, but they all come down to numbers. Knowing the numbers should point YOU in the direction of the problem. Refining the the areas that are receiving less than expected results should improve your conversion rates and total sales. Testing and knowing your numbers and how to use them is the fastest way to success. Here's a link to video. 

Good and Bad Results Will Happen...
Learn From Them and Make Changes

Test, Test, Test...

Statistically checking to determine what returns the best conversion rate can be one of the smartest moves you can make. Some experts in statistics have said that many of the tests being used in the area of marketing are probably invalid. The tests are not being conducted in a scientific manner with controlled variables. Statistical testing is a discipline of its own, and has to be conducted using specific rules. The way you test can either be good or bad depending on several factors. Usually any testing will help you determine your websites potential, especially on sales pages.

Many of the tests being conducted by most marketers are just that tests; some are better than others. I doubt that many of the tests being conducted are statistically, or scientifically sound. Statistics are usually based upon some truth, or confidence level, and require very specific calculations to determine the relevance of the test.

Should this rule out Testing? Probably not, several marketers have stated that testing has increased their conversion rates. Some testing is probably better than no testing at all. As long as you can control the variables and try to limit the testing to one variable your data may point you in the right direction. Several marketers have stated it worked for them, and the money they make indicates it works. Split testing is probably the best method for testing on a small scale, and appears to be the method used by several marketers.

Split testing is usually accomplished by changing one variable for your test. Many use their lists to perform the split test. After determining the one variable to be tested; the list is split in half, and one half receives the original test, while the other half receives the new item to be tested. It would be best if each half of the list was pulled at random for the testing. The greater the number involved in the test, the more confidence you could have in the results being relative. Many marketers use something called the "rule of 40", I have no idea where it came from so please don't ask. They just say you have to wait for a minimum of 40 to respond to your test; then determine which gives the best results.

The results may, or may not be significant because this isn't enough data to really draw conclusions. Even if the data collected points strongly in one direction the campaign could still return miserable results; and the other could have done better. I am only pointing out; the testing being conducted probably isn't scientifically sound. It doesn't mean that testing shouldn't be done; only that you need to be careful in how you use the data. Don't bet your shirt on it. Many successful marketers are using unscientific tests and doing all right, and some testing is probably better than none at all.

It is probably more important to get busy trying to market; than to wait for everything to be perfect. One thing is certain, if you don't market your product, it won't sell. The only way testing can hurt you; would be if you wager all your money on an inconclusive test, and it fails. If you are getting a decent return on your money, go for it. You will at least be making some return, and that is better than most.

The more you know about your business, the more likely you are to make good decisions. Testing will at least make you collect some data, and that could be significant. Just remember that testing may not be conclusive; and you are responsible for decisions of your business. The method you use in your testing can have a large impact upon the results. Care must be taken so you don't fool yourself, just be cautious and think about what you are trying to learn from the test.

There are some good ebooks available that explain different methods of testing and tracking. I wasn't able to find my copy while I was writing this article, so it only gives some vary basic ideas related to testing. I would suggest you study up on testing and tracking methods before you begin making too many changes. I will try and find my ebook and see if I can offer it for free at a later time.

As some of you know, I worked as a manufacturing engineer for several years. One of the big names in that area was a guy named Taguchi, we used his methods to find ways to improve quality in products scientifically. I knew I had seen a spreadsheet dealing with testing products marketability but couldn't find it while I was writing this article. I have been doing some 'housecleaning' on my computer the last couple days and found the PDF file. Click here to see a scientific method of testing to determine best results. I would be curious how many marketers are using such techniques.

I neglected to tell you that the link for the latest speadheet no longer goes where it once did. It appears someone has turned it into a business and the spreadsheed itself no longer is available. The instructions on how to use the spreadsheet will give you an idea of scientific testing. You might find it interesting, several large companies use the techniques of Dr. Taguchi. I used the method for several years to try and determine the best possible set of alternatives for a given situation. There are some easier methods that exist if I remember correctly there was an ebook called "Scientific Testing".

Testing is probably the only way of actually knowing what is going on; successful marketers are very good at testing and revising their websites and pages to reach the maximum level of productivity. Controlled testing is the key to a better understanding of the data derived from the test. If at all possible only one variable should be changed at a time and it should be compared to your benchmark to determine whether the outcome is positive or negative. Testing is probably the fastest method of accomplishing your objective of making money.

I can assure you that all the very successful marketers are doing some form of testing and using the data to tweak their webpages and promotions. Testing and tracking are the best ways to predict and determine what the outcome of your work will be. You will know much faster what is going on and make determinations about what is, or is not, profitable.