I have written about this in 2012 (link) but this one is different.
One of the most compelling reasons for creating a data warehouse is to reduce cost. Imagine if you were a large investment bank, trading fixed income, equities, commodity and currency. Every month you need to produce hundreds of reports to your clients and to regulatory bodies, and for management. The reporting team currently spends a lot of time creating the reports. They all have access to various systems. Risk systems, trading systems (order management), compliance systems, fixed income systems, finance systems. They copy and paste data from these systems into Excel, and do some calculations in Excel, which as the data source for making the reports in PowerPoint and PDF. Let’s find out how much this operation cost. Let’s say there are 8 systems, costing $200 per user per month, and there are 10 people in the team. The licence cost is 8 x $200 x 12 x 10 = $16k/month x 12 months = $192k/year, make it $200k. The salary is $60k x 10 = $600k per year. Total cost = $800k/year.
If we make a data warehouse, taking 1 year and costing $500k, we will be able to put together the data from that 8 systems into 1 database, and the reporting team won’t need access to those system anymore. That saves $200k/year.
Secondly, we will be able to automate the production of those reports. That will reduce the amount of time required to make those reports, thus reduce the number of people required, freeing them to do other activities. Assuming that half of the reports can be automated, that would save us $300k per year.
So total saving per year is $500k. Project cost is $500k. Surely a project with this level of “cost vs benefit” is well worth pursuing.