Split Test­ing Ex­plained - What Is Split Testing?

Diagram: What is split testing?You’re prob­a­bly aware that split test­ing is used to make your web site more successful.

Ba­si­cally you’re test­ing 2 ver­sions of your web page to find out which one per­forms better:

  • Which of 2 split tested squeeze pages gets the most con­firmed subscribers?
  • Which of 2 split tested sales let­ters gets the most sales?
  • Which of 2 split tested PPC Ads gets the most clicks?

Most of the work is done by split test­ing soft­ware run­ning on your web site. The only thing you have to do is to setup the 2 ver­sions of your web page. Then the split test­ing soft­ware will con­trol the test, record all data, and show you the results.

In the di­a­gram the split test­ing soft­ware is rep­re­sented by the gears

  1. Each time a vis­i­tor ar­rives at the page, the split test­ing soft­ware sends him to one of the two ver­sions of your test­ing page.
  2. Each time a con­ver­sion hap­pens (for ex­am­ple a sale), the split test­ing soft­ware will col­lect the in­for­ma­tion which web page that vis­i­tor has seen.
  3. In the end you sim­ply com­pare which ver­sion cre­ated the bet­ter re­sults. You pick the win­ner and put it on­line, you dump the loser.

The re­sult of your split test may look like this.

Sales Let­ter A Sales Let­ter B
Vis­i­tors to that page 1024 1008
Sales from that page 7 14
Con­ver­sion rate 0,68% 1,38%

In this split test sales let­ter B clearly sells bet­ter than sales let­ter A. So you’d end this split test, you’d drop sales let­ter A, and con­tinue run­ning the win­ner, sales let­ter B.

If you’re smart, you’ll im­me­di­ately start cre­at­ing a new sales let­ter C and run­ning a new split test C against the for­mer win­ner B.

It’s im­por­tant that re­peat vis­i­tors see the same page each time. Oth­er­wise you could not find out which ver­sion is re­spon­si­ble for a conversion.

  1. When vis­i­tors ar­rive at the web­site, the split test­ing soft­ware checks to see if it’s a re­peat visit or not.
  2. If it is, the vis­i­tor is sent to the same sales let­ter as last time (to en­sure the same of­fer is shown).
  3. If it’s a first-time vis­i­tor, a sales let­ter is cho­sen at ran­dom, and the vis­i­tor is sent there.
  4. The en­gine records which sales let­ter the new vis­i­tor is sent to (ei­ther by set­ting a cookie on the visitor’s com­puter or sav­ing the visitor’s IP ad­dress), so re­turn vis­its can be routed back to the same sales letter.

As you see, it’s not so hard to un­der­stand split test­ing. If you’re just start­ing out, keep it sim­ple. Don’t hop on su­per com­pli­cated se­tups for your first test, maybe do some­thing very easy, like just us­ing 2 dif­fer­ent headlines.