Unlocking SEO testing insights: Leveraging quasi-experimental designs for effective experiments

Abstract

With the development of marketing's digital landscape, experimentation has become crucial for companies' business planning. Testing can now be set and delivered within seconds, making optimisations faster than ever. However, traditional models, such as A/B testing and multi-variate testing, face challenges as marketing professionals seek to expand experimentation to areas where controlled conditions cannot be met. Among these is the case of search engine optimisation (SEO) experimentation, an area that is steadily becoming a crucial element in modern marketing strategies. This paper explores the use of quasi-experimental designs to overcome these challenges, allowing for the collection of insights when traditional experimental setups are unfeasible. Examples are provided to show how quasi-experiments can be effectively applied in multiple SEO scenarios, assessing the performance of the treated pages and using correlation to generate comparable control groups. It advocates for a shift in the mindset of digital marketing professionals, stressing the importance of adaptability in experimental approaches, underscoring the necessity to embrace quasi-experimental designs in modern marketing strategies by highlighting their pivotal role in achieving data-driven insights in an increasingly complex digital world.