20 times ROI from data warehousing initiatives

2 mins read

Almost all manufacturers of any size should be investing in a data warehouse initiative: it would deliver “20 times” the cost of implementation over time, and outweigh savings from a re-engineered ERP project. Brian Tinham reports

Almost all manufacturers of any size should be investing in a data warehouse initiative: it would deliver “20 times” the cost of implementation over time, and outweigh savings from a re-engineered ERP project. So says data warehouse veteran and advocate Allen Messerli, formerly global data warehouse project leader at 3M, and now principal of consultancy Messerli Enterprise Systems in the US. “The returns are huge.” He ought to know: his work at 3M from 1996 to 2001 on a massive Teradata-based system is credited with delivering that company more than $1bn savings over the last two years in procurement and indirect cost reductions alone. Messerli says that while that was achieved via a $50m investment, “first benefits came in the second year, although they got much bigger as the project progressed.” He also argues that for smaller companies, the cost goes down (“It could be $5m.”) according to numbers of source systems, data tables and so on. “Data warehousing technology is available to all manufacturers. I’ve identified 23 areas of business benefit… They have the potential to save an average of 10% of their annual revenue over a five year period.” That value is obtained by getting cohesive views of information across multiple systems and sites, and querying the data warehouse rather than the individual operational transactional systems. “That just doesn’t work,” he insists. “You need a data warehouse that’s up 24/7/365 and can deal with tens, hundreds, or thousands of concurrent users running complex queries.” He also indicates that, while there are now plenty of providers of data warehousing technology and associated business intelligence applications, it’s key to go for a logical data model, data integration – and granularity. “You want KPIs and summary data, but you need your users to be able to drill down to the specific related issues, products, geographies, functions, whatever. That’s what leads to success, not complex cubes, or non-standardised data that mean you need the IT department to go away and write another query.” His point: immediacy, detail and self-help are everything. “I’ve seen plenty of systems, and that’s what is successful. It enables companies to centralise their data and decentralise their operations because they can share real time information and make decisions from wherever they need to.” He concedes that data warehousing projects aren’t cheap, but maintains that they’re a lot cheaper than going for new ERP systems that won’t necessarily deliver the desired visibility. “Companies can waste a lot of money re-engineering their ERP and not get the benefits they’re expecting. Some companies do need new ERP systems, but most don’t have problems with their transaction processing – it’s access to information they want. “Some ERP companies misrepresent what their systems will do. They won’t necessarily solve your information needs.” Incidentally, he also says that data warehousing, done properly, can contribute to a supply chain integration project. “With all the systems feeding data into it and taking data out, they’re connected.” And he says that at 3M, everything from demand planning, through supply planning, forecasting, inventory planning and order management, was done through the data warehouse.