I would like to ask if it is necessary to allocate traffic on baseline (original page). Could I only use variation 1 and variation 2 to see which convert better. Will be my result statistical significant then?
I'm not sure if you meant you want to send 100% of traffic to the variations and 0% to your original, or if you want to alter the baseline as viewed on the results page. Both are possible, so I'll briefly outline what you should do for each scenario:
1. Allocate 100% traffic to the variations
On the traffic allocation page, set the traffic for the original to 0% and the remainder to your variations. Hit the apply button and then save your experiment. Please be aware that visitors who were previously bucketed into the original will continue to see the original unless they clear their cookies.
2. Change your baseline on the results page
On the results page, if you right-click on a variation, you can select an option to 'Compare other variations to this one'. This will set the variation in question as your new baseline. You can change it back or switch baselines at any time.
Hope this helps!
Actually if visitors were already bucketed into a variation, they will continue to see that variation even if you set the traffic allocation for the variation to 0. I forgot to mention previously, but if you want to completely exclude visitors from a variation, the best thing to do would be to pause the experiment, duplicate it, set the traffic allocation as desired and then start the duplicate. This way none of the new visitors will have cookies and you'll be starting with a clean slate.
Hope this helps!