According to an announcement by Microsoft, cookie-based experiments are currently available on Microsoft Advertising. If you select this option, you will only be shown ads from either your experiment or original campaign when they search.
With experiments, you can create a duplicate of a campaign and test it on a segment of its traffic. This will let you see if this works better for your business or not.
Microsoft recommends setting your experiment split at 50% to obtain enough volume more quickly and make comparisons more easily.
While this may yield more accurate data for some experiments, it may also take longer to accumulate statistically significant comparison data as opposed to if you did a search-based experiment.