<!DOCTYPE html>
<html><head><title>Why does my split test start over when I change a URL?</title></head><body>
<b>Why does my split test start over when I change a URL?</b><br />
<br />
Changing the pages involved in a live split test before it’s finished essentially renders the stats useless …<br />
<br />
… so any time you modify your Primary URL or add new variations to a split test, the system will automatically reset your test and start it over.<br />
<br />
This is the same reason you can’t change a split test URL.<br />
<br />
Instead, you should finish your current split test and then start a new one to test any additional variations.<br />
<br />
This ensures that your split test stats are as accurate as possible, and it also makes it easy to start a new test any time you’re ready by simply removing the losers and adding one or more new variations to test.<br />
<br />
<br />
<b>Does the split test start over even when I add a new variation to test?</b><br />
<br />
Absolutely!<br />
<br />
It’s hard to explain without getting into the nitty-gritty of statistics and the ridiculously advanced math involved, but the short answer is that adding a new variation in the middle of a test without restarting it is a really bad idea if you want to accurately determine a winner.<br />
<br />
Here’s a silly analogy that paints a perfect picture: Suppose you have two people who are blowing up balloons to see who can cause theirs to pop the fastest… Then mid-contest, a third person asks to join in on the game. Obviously, you’d have to ask the other two people to start over for the third person to have an equal shot at winning. If you didn’t have everyone start over, the results would be meaningless.<br />
<br />
Similarly, to determine a true split test winner, every URL has to “start over” whenever the contestants change.
</body></html>