Ellerine Holdings has invested in an enterprise data warehouse, powered by a high performance in-memory analytical platform to streamline its merchandising and customer relationship activities and underpin its role as a growth engine for listed parent company, African Bank Investments (Abil).
Ellerine Holdings consists of six furniture brands – Geen & Richards, Beares, Ellerines, Wetherlys, Furniture City and Dial-a-Bed. It has the largest furniture retail footprint in Southern Africa, operating more than 1 000 furniture stores and employing over 12 000 people.
When Abil took over the debtors book of Ellerine Holdings in 2010, the group restructured its operations so as to focus on its core business – furniture retailing. Operational excellence in merchandise planning and supply chain management was identified as key focus areas for improvement in the group’s retail efficiencies.
Ellerine Holdings’ Chief Information Officer Ian Child says an enterprise data warehouse was seen as the foundation for integrating information across the group’s operations and improving data flows in an unsynchronised environment. The Kognitio Analytical Platform was selected as the underlying technology because of its massive parallel processing ability, relatively low cost of implementation and limited support requirements. “It was also possible to get-up and running faster than with traditional technology,” he says.
Ellerines embarked on a proof of concept exercise, simulating some data crunching challenges that had been experienced during the debtors’ ledger conversion to African Bank. “We were very impressed with Kognitio’s ability to grind through masses of data in minutes compared to the many hours and sometimes days it took on high-end relational databases,” says Child.
According to Ellerines IT Executive: Business Intelligence, Stefan Terblanche, the group had been struggling to deliver on-time, trustworthy data to business managers on the legacy infrastructure. “The original system had been designed to aggregate decentralised key performance measures in our 1 000 stores, and batch them overnight into our central information systems. We were running out of night-time and could not land actionable data by start of business the next day.
“This frustrated users and was at odds with our aim of creating operational excellence. The Business Intelligence (BI) Department spent lot of time following up queries and reconciling numbers that balanced at one point but did not match to another point. The focus on improving merchandise and supply chain processes and systems identified weaknesses in the efficiency of legacy systems that we could no longer tolerate.”
Some significant enhancements were made in moving data from the branches to a central data store developed by our enterprise resource planning (ERP) vendors. This environment has since matured and stabilised and become Ellerines’ primary data source for the Kognitio Analytical Platform. “We load the Kognitio staging area through ongoing triggers and take on data into the warehouse data structures once a day,” says Terblanche. “This has removed our dependency on batch jobs that could be carried out only at the end of the branch day.
“We have now arrived at a sweet ‘SPOT’ or ‘single point of truth’, where we can publish timely and accurate consolidated information across operational, tactical and executive communities. The data warehouse is abstracted from the data sources so, in the future, as we evolve our transactional environment to centralised systems, we expect almost seamless integration into the standard industry data models (based on the Association for Retail Standards – ARTS – model ) implemented on the Kognitio Analytical Platform.
According to EnterpriseWorx Chief Technology Officer, Sean Longhurst, Ellerines chose Kognitio chiefly because of its performance. “Kognitio is the pioneer in in-memory massively parallel processing, with many independent units running in parallel. And it is still considered the fastest on the market,” says Longhurst. “This makes it a leader in the big data space. Because Kognitio has an open data model, existing data structures can be ported on to the new platform, without discarding the initial investment.
“Owing to the enhanced data flows and blistering speeds at which Kognitio can consume, process and make data available to users, it streamlines business processes. For example, some business intelligence (BI) models that were available to managers only 24-48 hours after the fact are now on hand by start of business the next morning.”
“Our business is enabled by technology,” says Child. “We understood we were embarking on a substantial data warehousing implementation. This is typically an expensive exercise that can drag on for years with a raft of support people. We’ve done it in 14 months on a thin level of support and with value being released as the system became available. That’s a huge efficiency saving. I commend our BI team and EnterpriseWorx for this exceptional delivery.”
Stefan Terblanche believes Kognitio is a strong foundation for the BI ecosystem across the 14 key communities in the Ellerines business. “We have tackled this project in a holistic manner with strong reference to the DAMA data management framework and our business intelligence competency centre,” he says. “This implementation has allowed us to significantly improve data flows as well as bring extract, transform, load (ETL) and database administration functions in-house. Our execution has improved and we are able to respond quickly to our internal customers.
“A large part of our success stemmed from implementing a single point-of-sale system across almost all our brands in South Africa. We are busy with a similar project across our foreign operations. This will bring greater harmony to our disparate world of multiple systems. System rationalisation has given us