Articles by Brian de Luna
Experimentation & Measurement for Search Engine OptimizationThis article describes how Airbnb measures the effectiveness of its search engine optimization efforts to ensure that its landing pages appear at the top of search engine results. To improve its rankings, Airbnb optimizes the content of its pages for relevance, performance and authority. They used a new landing page called “Magic Carpet” that is designed to increase relevance and decrease page load times, among other improvements. Since exact rankings cannot be measured, Airbnb relies on traffic as a proxy for an increase in rankings. However, since A/B testing is not suitable to measure increases in traffic caused by changes in external search engine rankings, Airbnb uses a market-level approach. This approach utilizes canonical URLs, which are unique to different cities, towns and regions, and are surfaced on search engines. Airbnb sets the unit of randomization of its experiment to a specific canonical URL, meaning that when a search engine bot scrapes the site, it consistently sees the same treatment for each page, and the ranking for that page updates accordingly.