This document outlines the six steps Plumbee took to build their analytics capabilities from scratch. They started with third-party analytics to get data fast but found it lacked flexibility. They then collected everything, processing vast amounts of raw data with Elastic MapReduce and Hive. Over time, they automated processing and built a relational data mart for faster queries. While making progress, they have more work ahead, such as replacing custom tools with open-source solutions like Flume and Cascading to improve scalability and enable data science capabilities.