What’s the difference between a Data mart and a Data warehouse? Why are they so often confused.
Simply put, an enterprise data warehouse is the union of all marts being fed by a single source. That single source is a staging area where data is cleansed and harmonized and is often referred to as an ODS or Operational Data Store.
A data mart can be stand alone reporting solution or it can be soundly integrated into an enterprise data warehouse.
Relational Solutions has been building enterprise data warehouses since the mid 90’s and pioneered the concept of an incremental, iterative approach. Ralph Kimball is considered the father of this methodology.
This approach allows companies to get a fast ROI (return on investment) that will address immediate needs of the business users. It will also provide a foundation that will allow companies to get incremental benefits as new data is integrated. The design withstands the test of time and lets the data warehouse grow with your business. With the evolution of new data sources, including Omnichannel data, social media, SOMOLO, e-commerce and other bigdata sources, this flexible methodology is needed in order to continue supporting new data sources.
Most data marts are designed as one off reporting solutions, designed to solve an immediate problem that the business users need to solve. When designed as a stand-alone, they are often referred to as “stove pipes” or “silos” of information.
Analysts new to the space think these are new terms, but Data warehousing consultants have used these terms since the 90’s. They’re used to describe stand-alone reporting solutions. Typically these stand-alone solutions are developed by individual teams or departments.
These groups develop “silo’s” or “stove pipe” reporting databases to achieve a specific goal that they were unable to get financial approval for. If they have a need for something that you can't get approval for, you resort to building something on your own. It happens in every company and every department.
That said, all data marts are not created equal. Some are in access, some in spreadsheets, some are in SQL Server or Oracle. Some are silo's and some are not. Data marts do not have to be silos. Designed correctly, a data mart can be integrated and should be fed from a single staging area where business rules are applied. Thus, a sound data warehouse is the union of all marts being fed by a single source.
Having an infrastructure that stages the data, cross references the data, cleanses it, harmonizes the data, and feeds it into a data model that then feeds subject specific marts offers the best growth potential. The shared dimensions from one data mart to another provides consistency from department to department. We describe the data warehouse process in several educational video's on our YouTube Channel.
The right data model is also critical to the success of a data warehouse. Data modeling techniques vary depending on the database target. Designing the data model correctly allows business rules to be applied and data to be accessed easily by the users. This design also maintains consistency from department to department. It also provides IT with a manageable solution that is designed to evolve over time to accommodate new data sources and new user requirements.
Companies who have a properly designed data warehouse can integrate internal data, outside data and even new big data sources. Request Your Consultation