Twitter, the social network evolving faster and regular usage by millions of people. Spam is defined as unwanted content that appears on online social networks (OSN) sites that are driven by several goals such as to spread advertisements, generate sales, viruses, phishing, or simply just to compromise system reputation. Twitter has gained a lot of popularity throughout the world, it continues to be a very interesting target for spammers and malicious users for spreading spam messages. Several machine learning algorithms exist to detect the spam messages in the OSN but still it is a challenging issue. Not only the significant classifiers used are essential but also the selection of relevant features is also very crucial. Binary and Continuous Feature Engineering (BCFE) analysis is proposed in this paper for better feature selection. The features are evaluated using significance strategies and the best features are selected for effective spam classification. © 2020, Springer Nature Singapore Pte Ltd.