Mereo Uses CoSort to Transform Big Data for Media & Leisure BI
Christophe Monroy, chief system architect at Mereo.
Mereo is a leading provider in the field of revenue optimization and business intelligence. Our clients are major sales houses in the European TV radio and outdoor advertising sector, as well as operators in the Hospitality, Entertainment, and Travel industries. Mereo helps increase its client revenues by implementing revenue optimization techniques.
We have gathered the skills and experience to guide sales houses through any of the stages of revenue optimization, from profit assessment to process improvement, tools integration, implementation, maintenance and support, as well as staff training. Founded in 2000 and based in Paris, France, the Company has assembled an expert consulting and engineering team experienced in yield and revenue management solution implementations in the media and leisure industries.
Mereo clients use a mix of Unix, Linux, Windows-based servers and operating systems to analyze and simulate sales and deal data, run statistical forecasts, monitor and manage inventory, and plan advertising campaigns.
Our applications supply the common ‘Revenue Manager’ module aggregate data that’s generated from complex and sophisticated computations. Those computed results are obtained daily by batch processing performed on large volumes of customer sales and traffic data. Mereo uses the CoSort SortCL program as an ETL engine, integrating and transforming the client data sets rapidly so that we can perform these computations and reach the BI generation stages fast. We deploy basic SortCL operations such as Select, Field Definitions, Join, Sort, Merge, and Aggregate, in combination and through serialization of multiple job scripts in order to rapidly generate output files in the 60GB range.
The CoSort SortCL program is extremely fast, compatible across all our client data processing platforms, and is easy to use – the language is simple and understandable, and extremely powerful for data transformation, remapping, and reporting in its own right. Mereo also feels comfortable with SortCL because its language is procedural.
A recent version of SortCL contained certain syntactical incompatibilities with a prior version amid new feature additions. While understandable, this can result in us having to perform a large number of regression tests because we deploy so many scripts.
CoSort has represented the best choice for Mereo because it meets our criteria for:
- Very fast data sorting and related transformations
- Cross platform compatibility for job design and execution
- Low cost development licenses
- Simple deployment model(s)
We license the CoSort package for internal development and testing, then embed the CoSort SortCL program for distribution within our applications in a transparent way.
IRI and its local office in France have always been responsive to our questions, and attentive to our problems. We’ve been able to easily get help and updates as need: Despite being a 700+ page English manual, the CoSort .pdf documentation is clear, concise, and readily usable as a reference and learning tool.