The document discusses how companies often approach big data wrong by having data scientists manually collect and clean data instead of using big data as a service solutions. It provides the example of a data scientist named Jackie at a publishing company who needed pricing data but found manual and scraper methods too slow, expensive, and difficult to scale. While good for small volumes, these approaches did not meet the needs of her large company. The document advocates for using big data as a service which can provide specialist skills, scalable data sources, predictable costs, and reclaim over 50% of data scientists' time spent on data preparation instead of analysis.