You want to make sure that your sample size is large enough so you could run the test again, or it could simply mean that there is no difference. If you are asking from a Marketo functionality how a winner is chosen, that's a different question entirely. The likelihood of it being exactly 50.50 is pretty rare. If you are choosing the winner manually and you get those results it means it doesn't matter and you can go with either one.
I just wrote a blog on this topic. You can find that here. Making the Most out of Your Marketo A/B Test: The Statistics you Need to Know