The Variant WON!
There was a change.
This test has statistical significance.
Statistical significance evaluates whether or not a result was achieved due to the actual change that was made or due to a sampling error. In order for a result to be considered statistically significant a test should reach 95% statistical significance, or 90% at a minimum.
The optimal test length generally falls anywhere between 2 weeks and 8 weeks. After 8 weeks any given test should be stopped. This is because users cookies are often deleted and they can end up being shown numbers variations, which would disrupt the user experience.
The length of time to reach statistical significance depends on how much traffic a page is getting and the pages conversion rate. The greater the traffic and the higher the conversion rate difference, the faster statistical significance can be reached. However, if that level of significance cannot be obtained within eight weeks, another test with either greater traffic or a greater difference should be performed.