When I started online, I was a very avid tester of everything... including what I now consider goofy things. Because of that, I know all kinds of goofy things… and a few useful things. I happen to know that a link that says “Don't Click” will get more clicks than a link that says “Click Here”. That's not very useful for marketing because it's not the clicks that count… it's the sales.



I also happen to know that a white background is the most profitable from split tests. Yep; I really did test every other color of background (32 variations of different colors including pink and purple). Now that is useful information. However, doing that test probably cost me $20,000 in lost sales. I've done many, many, many other “goofy tests” that have cost me a ton of money.

Why was that a goofy test? Because it's very easy to do a statistical analysis of other people's sites to get that answer. You don't need my expertise or a large dataset to get that answer. It's extremely obvious. Go to the top 20 most profitable sites on the ‘net. What color of background do they use? Now go lower in the list to get a control. Do you see that a few of the lesser profitable sites on the ‘net are doing what I did… a goofy test. They are testing something very, very basic that has already been testing by literally millions of others. The answer is known.

Think about the process of split testing as “zeroing in” on the most profitable sales page. Do you remember that game “higher and lower” you played as a kid? It went like this. Someone picked a number between 1 and 100 and didn't tell anyone. Then someone guessed a number. The answer would be “higher” or “lower”. Amazingly older kids would guess the number reliably in 6 guesses. Younger kids would take dozens of guesses.

What did the older kids know? They had a strategy. In fact, the most efficient strategy is obvious to the older kids and they would always use it… that is until they realized that it wasn't even a game anymore because the most efficient strategy was known. The older kids would first guess the number 50. If the answer was “lower”, they would guess 25. If the answer was higher, they would guess 75. They would continue to split for up to 6 guesses and they would have the answer… guaranteed!

Split testing is the exact same process. However the number is between 1 and nearly infinity. If we just randomly do split tests, we will be like the younger kids who first guessed 17 and then guessed 77 and even sometimes guessed numbers that had already been ruled out with prior “higher”/”lower” answers.

Here is the best way I have found to optimize a sales letter using the “older kids” strategy.

1. Do a statistical analysis of profitable and unprofitable sites for the general layout, color, items of content (ie: headlines, testimonials, guarantees, etc), font and other such general items. I have already done a lot of these and you can get the answers with the Statistical Copywriting course. Don't test known unprofitable things (like bonuses). That's goofy.

2. Do the same for the words and symbols used on profitable and unprofitable sales pages. I've already done this with Glyphius. Use it. Don't test low scoring stuff. That's goofy.

3. Now that you have a sales letter that contains the profitable items (guarantee, testimonial, headline, etc) and it is arranged like profitable sales letters (white background, a profitable font, a profitable headline color, etc), THEN… AND ONLY THEN… start your split testing to make it more profitable for your particular product.

Don't test any goofy stuff. You really don't have a need to test the wingdings font, a purple background and yellow text. That's just goofy and you'll lose sales if you do it. We already have that data. It's known. It's not profitable. It's goofy to test it again yourself.

Instead start with the Statistical Copywriting copywriting course. Optimize all of the words in your copy with Glyphius. Then start with split tests on the truly unknown stuff that can actually improve your sales.