Business Intelligence Innovation

Madan Sheina investigates how business intelligence software continues to bring value to enterprise adopters.

Business intelligence consistently tops CIO spending priority lists as it continues to spread its wings into more corporate departments. Forrester Research estimates that the BI reporting, analysis tools and applications market will top $7.3bn by next year.

Driving this growth is a confluence of technological trends that promise to make BI more pervasive, accessible and affordable than ever before. Here, CBR investigates how some of those trends have encouraged the use of BI in the real world.

Appliance of science
For example, Virgin Media, the entertainment and communications division of the Virgin company, has adopted Netezza's NPS data warehouse appliance for complex BI querying of customer call data sourced from its various merged and re-branded NTL, Telewest and Virgin Mobile companies.

Virgin had previously outsourced part of its telephony analysis due to the sheer complexity and volume of the data involved. Performance issues with the old data warehouse system meant that Virgin was restricted to querying specific data on international, mobile and premium rate calls for high usage analysis.

But the appliance model has opened up query and analysis to all call data types and allows managers to run historical reports without fear of runaway queries grinding the system to a halt. It has also allowed the company to bring this back in-house affordably, with a rapid return on hardware investment, and with a sufficiently high-level of performance, with queries running on average over 250 times faster than with the previous Oracle-based data warehousing system.

Bringing the analysis of NTL telephony data in-house has been a boon for Virgin, allowing its credit services and revenue assurance teams to access and query data in near real-time. The appliance also protects Virgin's existing BI and data warehousing investment in Informatica and Business Objects technologies, which sits and works comfortably alongside NPS.

Specialty chemicals firm Johnson Matthey, meanwhile, has tapped into the dashboard concept to transform its exclusive financial planning tool into an integrated operational reporting and analysis system that can be accessed by more of its business managers. The company is using interactive OLAP analysis and reporting dashboard technology from Applix called Executive Viewer (EV) to roll out analytic capabilities to its employees. Initially deployed as a financial query tool running on a single server, EV has now been extended across the enterprise and now serves over 150 staff working in multiple departments of the company.

Prior to installing EV, Johnson Matthey had grappled with a spaghetti tangle of incompatible spreadsheets sourced from different divisions. The effort of reconciling inconsistencies between these spreadsheets meant the company could only analyse data on a fortnightly or monthly basis.

With EV, analysis has now become a daily activity. For example, daily analysis of precious metal commodity information is critical to identify variations and trends in meta accounts. Johnson Matthey also uses EV as a tracking system to monitor sales and turnover, fundamental in helping the company to make informed business decisions.

The simplicity and richness of EV's visual interface has been instrumental in adoption at Johnson Matthey. Any non-technical user in the company can make sense of complex data, allowing them to develop their own reports and drill into over 20,000 customer and product details, rather than seeking assistance from the IT department.

Predictive treatment
Insurance claims fraud is a big problem for the NHS, costing it millions of pounds every year. The NHS' Counter Fraud Service (CFS) is using predictive analysis technology from SAS Institute to crack down on fraudsters. CFS is using the predictive data analysis techniques in SAS' Intelligence BI platform to anticipate where fraud might occur, and in doing so prevent it from happening by tightening up appropriate policies and procedures.

SAS' Intelligence platform works as a black box system that analyses NHS claims payment data and other internal and external sources to identify trends or anomalies. Once an anomaly has been detected, the system triggers an alert to the CFS team specialists for further investigation. And since the system is designed to run in near real-time, it provides up-to-the-minute fraud detection and prevention.

CFS says its SAS implementation was prompted by recent recommendations in the UK Government's Fraud Review Report that advocate a professional, holistic approach to tackling fraud. Counter-fraud activities by the CFS team have helped the NHS to cut its losses from fraud by 55%. It expects an even greater return once the new SAS predictive analysis system kicks in.

Insurance firm Norwich Union implemented Teradata's active data warehousing to support the commercial launch of its newly conceived motor insurance policy offering called Pay As You Drive, which tracks driver behaviour with in-car geographical positioning systems (GPS) technology. The novel usage-based policy is designed to work like a mobile phone bill, with customers being charged according to usage, including time of day, type of road, and mileage. Norwich Union says that Pay As You Drive insurance begins from as little as a penny per mile.

The company has implemented a 100 terabyte Teradata warehouse to create a high performance and highly available operational platform for analysing driving information gleaned from in-auto GPS devices: geospatial data such as car position, routes taken, time of day and mileage.

Teradata's signature data scalability was a prime factor in Norwich Union's choice of vendor. "It was the first time we had dealt with this sheer amount of data. Storing the data, turning it into a monthly bill, and the necessary summaries and calculations meant a new infrastructure was required," says Norwich Union's product development manager Sue Rowland.

Norwich Union has been piloting the project on 5,000 customers for the last three years, recording car usage data over 100 million miles and 10 million trips. The data is combined with other information such as traffic data to customise the product to consumer needs.

Norwich Union expects data collection rates to increase more than 15-fold in the first full year of the policy scheme's operation. Each day, every trip made by each customer will be analysed in the Teradata warehouse and the intelligence passed on to the billing system to produce individualised customer premiums.

Teradata's analytic capabilities also help Norwich Union to continue rating and managing a much larger set of customer trips on a daily basis, while better managing the associated risk. It will also be able to take advantage of this increase in information by combining with other data, such as traffic information, to customise the product to meet the needs of different customer segments.

Interactive reporting
Ford Motor company has tapped into the Ajax-like interface capabilities now included as part of Information Builders' WebFocus enterprise reporting system, to ease the transition from staid mainframe-based tabular reports written in Cobol to richer and more interactive web-based reports.

One of the initial reporting application areas that WebFocus targets is Ford's warranty business, to help consultants quickly identify and resolve repair and claims trends problems through better reporting of warranty expenses and trends. Unlike the static older 'rows and tables' warranty reports printed in plain text, the new WebFocus reports are comprised of colourful and dynamic charts that users can drill into or interactively tweak dimension views and measures with the click of a mouse. For example, users can click on a graph and you can drill down into the data, moving from the average cost of repairs or the number of repairs per thousand vehicles, to charts and data showing repairs by components, such as engines, transmissions, suspensions, or electronic systems.

It was important for Ford to deploy the system to its dealerships without forcing them to install any special software on their computers. Instead, WebFocus' thin approach offers rich interactivity that could be accessed from an ordinary web browser. The use of Ajax technologies helped immensely with performance. The WebFocus system is capable of dipping into the huge database to pull out the records, report the complex statistical calculations, and draw the graphs with drill-down links, all within 60 to 90 seconds.

Ford estimates that its WebFocus implementation has cut about $25m a year in costs through the fast and accurate delivery of business data in a visual, interactive and flexible reporting environment. The company also says that because WebFocus managed to reuse about 80% of the legacy reporting code it resulted in a speedy implementation. Ford's dealer warranty information comprises more than 80 million detailed records, spread across Oracle and Teradata databases stored on mainframes. In addition, the diagnostic reports and claims lists that supported the analysis were built with sophisticated code that Ford was reluctant to pull apart and rebuild.

Point-of-sale analysis
In a classic example of a company struggling with "spreadsheet hell", cosmetics manufacturer L'Oreal UK has replaced its Excel-based sale and customer forecasting processes with performance management software from Cognos. L'Oreal wanted to get a closer look at how customers at major retail outlets like Tesco, Asda, Boots, and Morrisons are buying its various product brands that include Garnier and Maybelline.

Working with performance management consultancy BI Inform, L'Oreal deployed Cognos Planning to improve its sales planning and forecasting processes and fix internal financial reporting inefficiencies caused by its previous reliance on Excel spreadsheets, which often required national account managers to manually reconcile inconsistent data. Every time a new planning or forecasting scenario was introduced, such as the success of retailers with L'Oreal-led promotions or seasonal variations, account managers had to effectively start from scratch with hours of reworking spreadsheets or inputting new data.

Cognos Planning alleviated this admin-intensive spreadsheet process by securely taking electronic point-of-sale data from the retailer system that the L'Oreal national accounts team uses to develop unit, sales and profit margin projections, including built-in key performance indicators that are tailored to each retailer's individual requirements. The software also allows various business scenarios to be modelled in minutes.  I would like to receive Application Development News
 I would like to receive B2B & B2C News
 I would like to receive Business Applications News
 I would like to receive Business Digest
 I would like to receive Business Intelligence News
 I would like to receive Computer Business Review Daily News
 I would like to receive CRM News
 I would like to receive Data Management News
 I would like to receive EAI & Middleware News
 I would like to receive IT Services News
 I would like to receive Microelectronics News
 I would like to receive Mobile Technology News
 I would like to receive Networking News
 I would like to receive Operating Systems News
 I would like to receive Outsourcing & ASP News
 I would like to receive PCs & Workstations News
 I would like to receive Security News
 I would like to receive Servers & Mainframes News
 I would like to receive Storage News
 I would like to receive Systems Management News