Why do we need Data Warehouse?

Question

I was wondering about why and when we need a data warehouse, I mean the main goal of data warehouse is to provide a reporting from multidimentional view, but in some case there is a way to build a report using DBMS, it can produce a report from multidimentional view, about the size of the database, the size can be made incrementally, and about integration DBMS also can make integration from many data source and if they have a different key from many data source, dbms can fix it (althought we're not using data warehouse), from the peformance side, how much time it will take to generate a report from operational data store and generate a report from a data warehouse, are the processing time is having a big gap?

Sorry for bad english and weird question, I'm still learning about data warehouse.

Answer 1

Why do we need data warehouse?

I will share experience of me and company I work for, how and when we have decided to build data warehouse.

1. Historical Data

   Our ERP back-end is designed to cover one year of data (because of high volume data). All predefined reports are built to watch one year data. So, if you want to examine sales growth per customer or product you have to run report twice (from two sources - you pick source on login page) and than lookup values in excel.

2. Multiple source

   At some point we bought third party ordering solution (tablet order -> web service -> database) for business enhancement. ERP provider took a fortune from us 'to integrate it in system'.

3. Bad data

   Our CFO and his team needed a whole week to generate good P'n'L report. Why? Analytic data was wrong, so they have to go row per row (in excel export) and to search for abnormalities.

4. Reports performance

   Sometimes it took over two hours to generate yearly report of sales grouped by product, customer and day.

5. Adding intern grouping

   When management decide to separate customers or products by another criteria, we have to pay ERP providers, and also we have wait for 1-3 weeks. And guess what, Sales Manager gives quit, and new one is asking new grouping types. It costs (time & money).

What to do? To change ERP system, or to create data warehouse? We spend months and months to create custom system from theirs standard product. We spent money and money to pay theirs extra hours. We have more than 250 employers to teach how to use new system. We have to wait for at least one year for new ERP provider to develop our custom needs.

Then we bought company which is the competition. We wanted to track our progress in comparison to their sales value. We decided to build data warehouse!

And we made our lives easier. After that, we bought Warehouse management system, Geo-Tracking Logistic Management System, another company, and guess what? We have data we want, in way we want, with great scalability for no giving money to ERP provider.

Now, our data warehouse is providing:

1. Historical Data from our company, and two others company with perfect keys mapping.
2. Multiple Source, our current data, our historical data, bought companies historical data, WMS, TMS, Ordering System and Excel Sheets (KPIs, targets, estimation, ect)
3. We made a rules table, so CFO and his team have only to administer on rules table, we always get good data, and wrong data is automatically reported (on daily bases) to specific department in order to change it.
4. Oh, managers just love OLAP. It's blazing fast.
5. Using Excel tables as one of source, we can easily map products or customers if management wants new grouping.