Furniture retailer Supreme furnishers, part of the JSE-listed Profurn Group, has launched a continent-wide data retrieval and management system in an effort to centralise and simplify its database management processes.
The Profurn group, which includes branded chains Morkels, Protea furnishers and HiFi Corporation, is South Africa`s largest furniture retailer. Supreme, with its head office in Braamfontein, has a network of nearly 100 branches across Africa, from Lesotho, Swaziland and Namibia to Botswana, Kenya, Ghana and the Ivory Coast.
"Because we have so many branches spread across such a large area, our data sources are both disparate and unsynchronised, which makes centralised data analysis costly and time-consuming," says Supreme IT director Mark Christensen. "Combined with the notoriously poor telecommunications infrastructure on the continent, we`re sitting with a real problem."
Supreme`s initiative incorporates a satellite-linked network from its remote branches to its head office in Johannesburg, which feeds through localised data to its Universe data warehouse.
"Central to this process is the extraction, translation and loading (ETL) layer, for which we`ve specified Ardent`s industry-standard DataStage software," adds Christensen. "Our biggest headache comes from consolidating information from multiple data sources, including multiple instances of the same record. With DataStage in place, this process can be automated, significantly decreasing the time spent on data extraction and increasing the time available for data analysis."
Christensen uses the example of a salesman transacting through multiple tiers of the sales chain to illustrate his objective.
"When a transaction goes through one of our branches in Ghana, for example, the salesman`s details and all the details of the transaction are captured to the local Universe system," he says. "This data then gets fed through other systems, such as the payroll system, at which time it usually changes in some form or captured in a different way to the original data."
"Data from all our branches and systems is continuously fed back to head office for processing, and a single record - in this case the salesman - can have different instances in the data across different systems," he continues. "It is therefore impossible to analyse our consolidated data accurately without ETL software that integrates across our multiple systems and natively understands the multiple data streams."
The filtering process to extract information for a monthly performance review currently takes a week - sometimes longer - depending on the complexity of the data. Christensen says that once the system is operational and DataStage has been implemented to filter and feed the data warehouse, performance results can be processed and released in minutes.
"With our satellite network system and the efficiency of processing we`ve designed, we hope to set a precedent in the group that our sister companies will adopt in time," he concludes.
Share