When a visitor lands on a tested page, Nelio A/B Testing generates a random number in JavaScript and stores this value in a cookie named nabAlternative. This value is then used to determine the variant a visitor is supposed to see on any given running test. For example, in a regular A/B test with only two variants, an even value corresponds to variant A and an odd value corresponds to variant B.
Since these values are randomly generated in your visitors’ browsers, it’s possible that, when a test starts and there are only a few page views, one variant has more page views than the other. This is completely normal in a random process and should automatically result in an even distribution as the number of page views increases.