Founded in London, England, by Italian brothers Sergio and Bruno Costa in 1971, Costa Coffee is the United Kingdom’s favorite coffee shop, having been awarded “Best Branded Coffee Shop Chain in the U.K. and Ireland” by Allegra Strategies for five years running. With more than 1,999 coffee shops in the U.K. and more than 1,168 shops in 29 overseas markets, Costa has diversified into both the at-home and gourmet self-serve markets. There are now more than 4,275 Costa Express self-serve machines. Costa is a wholly-owned subsidiary of Whitbread, the U.K.’s largest hotel and coffee shop operator.
Keeping back office operations running smoothly
Coffee is big business: there are 1.7 billion cups of coffee sold each year in the U.K., according to consultancy Allegra Strategies. Costa Coffee is the largest coffee shop chain in the U.K. and the second largest in the world. Not content to rest on its laurels, the company continues to grow at a brisk pace and expects to soon surpass 2,000 shops in the U.K. alone.
Behind the scenes of that impressive growth is the busy back office operations of Costa Coffee. Keeping operations at the head office running smoothly requires high availability and performance of Costa Coffee’s enterprise resource planning (ERP), customer relationship management (CRM), and other critical business systems. Keith Patterson, IT operations manager at Costa Coffee, and his team of seven analysts are responsible for all of Costa Coffee’s head office IT operations, service delivery, change management, and release management. Keith’s group also oversees the operations of the company’s corporate website.
With service level agreements to fulfill, Keith’s team needed deeper, more detailed performance data on the business systems it supports. “We couldn’t measure and track how our infrastructure was working other than whether or not we had any outages or major incidents over the past month,” says Keith. “If an application was using a lot of CPU, our infrastructure team didn’t have the data to understand what was causing it. We couldn’t see spikes in processing times or bottlenecks in the system.”