A tracking and analysis mechanism used in advertising to compare the performance metrics of different ads in a campaign to determine which one(s) perform better than others.
In theory, ad campaigns benefit from initial testing on a small scale because most ads will fail or under-perform. Split testing can therefore be helpful in identifying those ads that are successful so they can be scaled up once a strategy has proven itself.
Also known as A/B testing although this usage is less frequent in online marketing.