BBC 面临着一些挑战，例如遗留系统降低了生产力和内容交付速度，以及需要大规模的可扩展性和更快的响应时间。

HEROLD's initial challenge was to quickly monetize enormous amounts of addresses, household and corporate data that it acquired from purchasing a directory business. The company was seeking a better way to connect its customer-the businesses that advertise with it -with user of its digital properties, therefore HEROLD needed to create new digital products that could summarize and structure HEROLD's data assets so it could be productized and resold. Additional challenges included: Integrating and Managing Complex and High Volumes of Data. Integrating, accessing and searching the massive amounts of acquired personalized and address data could not be accomplished with HEROLD's existing relational database technology. Much of HEROLD's company data resided in as many as 15 separate systems and included records on more than 360,000 businesses, comprising of unstructured pieces such as product information, user reviews and recommendations, ratings, photos and text descriptions, which came from product feeds and manual entries. Aggressive Timeline to Demonstrate Return on Investment. HEROLD needed to integrate the newly acquired structured and unstructured data into its systems to quickly create and introduce new products that would help them stay ahead of the competition while demonstrating return on investment to management.

The AIRBUS flight test program is a key component of aircraft development. Every part in every possible configuration must be tested in as many scenarios as possible. This creates multiple types of data, all in different silos — plane configuration criteria, flight data criteria, flight event behaviors, crew reports, and other unstructured content criteria. Historically, each test flight required manual research into previous tests to create a set of specific requests, and the effort to leverage data to optimize those new flight tests was considerable. While a lot of data was generated during testing, AIRBUS’ process to search within the data in order to simply retrieve a specific condition and its result was a slow, manual one. And, flight test environments are constantly changing, with a series of maneuvers done at specific locations, which meant that traditional data approaches couldn’t keep up with this complex data challenge.