Energy supplierE.on Retailhas gone live with a financial transformation software
project that will give it an accurate view of the profits generated
by each customer and better forcasting of future
demand.
The system is the fruit of a three-year system development
project that cost more than £10m. E.on said it almost had to revert
tointensive "bare knuckle" programmingto reconcile its billing
accounts with its accounts settlement system to establish its gross
profit margins.
The company has accounts for more than eight million customers.
Project leader Lawrence Edwards told Computer Weekly that it must
reconcile its customers with the accounts it runs with power
generatorsto establish its gross profit margins. "It is a massive
problem because of the data volumes and complexity," said Edwards.
"Others have tried and failed, but we persevered."
Edwards said, "Operations are based on 150Gbyte tables
generating 1.8 billion rows of new data over approximately 14
hours." The system consists of in excess of 6Tbytes of customer
data and is growing rapidly.It has to integrateseamlesslythousands
of elements containing data from different sources into one system.
For example,a house may be listed in a table for gas supply and in
another for electricitysupply and in a third for account holder.
The system combines the data from all these tables to produce a
single, comprehensive picture of the account holder.
The database alsohas to cope with high churn in its customer
base, which requires intensive data cleansing to produce accurate
bills.
"The system generates billions of rows of data daily," Edwards
said. "We couldnot do it with our [SAP] financial system, or indeed
any other off-the-shelf package."
E.on turned to the Oracle database management systemrunning on a
dedicated 48-processor Sun Sparc unit. The key application
development tool, said Edwards, was
Oracle's version of Olap (Online Analytical Processing).
This is a specialist tool that uses a database with a
multidimensional data model that combines aspects
of
navigational databases and
hierarchical databases to make it faster than standard
relational databases. This allows it to quickly produceanswers to
complex analytical and ad-hoc queries against huge data
volumes.
With better information, E.on should be able to make better
decisions, Edwards said. However, pricing is a hot topic in the
industry. He was reluctant to be drawn on how the system will
affect E.on's pricing, but he said it will in the longer term make
E.on better at sending customers bills for the right amount. This
should mean it gets paid quicker, and better forecasting means it
should be able to buy cheaper. Combined, they should make E.on more
competitive and hence more attractive to customers.
E.on's financial transformation project – what it has
achieved
• Cut the time to create long-term demand forecasts from 12
weeks to five days
• Cut end-of-period reports from weeks to hours
• Allowed users to build detailed "what if" scenarios based on
customers, products, weather changes, and rebates
• Established a single, common data source and full audit
trail
• Made it possible to accurately predict gross margin and future
financial position through comparing forecast revenue with forecast
costs
• Gained real-time updates on the popularity and financial
position of each of the company's products and services
• Ensured faster settlement with partners, suppliers, and
clients.