Advertising company JC Decaux's UK operations were becoming increasingly complex and, in an attempt to manage this complexity and the wealth of data being created, the company upgraded its Business Intelligence system to an Infobright data warehouse system from Open Business Associates (OBA).
JC Decaux has an exclusive focus on out-of-home advertising, such as that posted on street furniture, billboards and electronic billboards, and transport.
Following an evaluation of two systems, the company decided to partner with specialist open source management firm Open Business Associates (OBA) in December 2009.
The system chosen was an open source Infobright data warehouse system combined with Talend’s Extraction, Transformation and Loading (ETL) Analytics tool.
The data warehousing system allows the generation of timely reports for use by JC Decaux UK's in-house staff, and provides detailed, up-to-date information to JC Decaux's head office in France.
It has also enabled the generation of a common information bank, aggregating data from three separate databases holding assets and sales data for large format advertising, street furniture and airports.
Why did JC Decaux's choose OBA over the competition?
JC Decaux's financial systems and BI manager Raj Rajkuman said that before it opted for the system, the company had three choices: to stay with the existing approach, go with an established BI firm like Oracle, or investigate the open source market.
JC Decaux UK IT manager Stephan Lavollee said: "The existing approach meant lots of time spent gathering information from scattered data sources, and would have seen the production of lots of reports in different formats using different processes and methods."
"Our users had developed their own complex reporting methods, which were inefficient, time-consuming and didn’t meet our business needs,” added Lavollee.
As an existing Oracle customer, JC Decaux's IT team considered developing a solution on Oracle's platform.
But Rajkuman explained that the company was already using all its Oracle licences and that buying new ones would increase the project’s cost – and so it discounted the idea.
In early 2009, Lavollee's IT team tried developing a system using a MySQL database and Pentaho's Kettle ETL tools. The system offered a 50 per cent saving over the Oracle option and initially performed well.
As more data was added Rajkuman said producing reports went up from 20 minutes to two or three hours.
"We normally produce reports on a Friday afternoon and so having to wait three hours for the data to be loaded into the data warehouse was not suitable. We had to go back to the drawing board," Rajkuman explained.
The tender process
It was following this that JC Decaux's team identified Infobright's BI-focused database as a possible solution.
Infobright introduced their partner, Open Business Associates (OBA) and, after initial discussions, OBA recommended a short pilot project using JC Decaux’s own data.
OBA were given just three days to complete the proof-of-concept, during which time OBA’s team identified and addressed a number of issues concerning JC Decaux's data layout.
OBA then loaded the four million records into Infobright’s database in under two minutes from MySQL tables.
“I was amazed with the results,” said Rajkuman, “queries which had taken more than 15 minutes using MySQL only took a couple of seconds using Infobright. It handled all our data, and we could generate reports quickly and easily."
OBA began rolling out Infobright's system in December 2009, using Talend as its ETL tool.
JC Decaux can now load data in 20 minutes and produce reports in under three minutes.
The project is the first phase of JC Decaux’s three-year plan to roll out a complete BI system for its UK operation.
JC Decaux is aiming to integrate all its historical data into the system and plans to roll out a further three data warehouses in the coming months.
Within two years, all its UK operation’s departments will use its BI system, eventually producing company-wide information through a single reporting tool.
Its UK arm was established in 1983 and it has revenue of almost €2.6bn and over 650 employees.
By eliminating high entry costs for big data analysis, you can convert more raw data into valuable business insight.
A discussion of the "risk perception gap", its implications and how it can be closed