Traditionally search has been all about "how do I find the one (or five) documents that answer my question?". However, what about when I want to encourage serendipity and discovery (common in research use cases)? In those cases, I care about increasing the diversity of my search results, while still ensuring the precision (i.e relevance) is decent. This talk will provide a simple formula for measuring your search results diversity (using Apache Zeppelin as a notebook), and offer some concrete tips on surfacing diverse search results.