For the past decade or so, organizations have efficiently and effectively used a combination of data warehouse based analytics and business intelligence to meet their needs. However, there has been a growing interest in Business analytics and a declining interest in Data Warehouse Business Analytics.
But first, what is the difference between Business Intelligence, Data Warehouse Analytics, and Business Analytics?
- Business Intelligence (BI): BI is the combination of data, technology, analytics and human knowledge and expertise to gather and identify insights that help make better business outcomes.
- Data Warehouse: A Data Warehouse brings together all of an organization’s raw data stored in multiple operational systems, logs, and unstructured data sources into a centralized location. From there, organizations can utilize the data for reporting and analysis.
Benefits of Data Warehouses:
- Easier and more efficient query writing
- Less resource contention for operational systems
- Increased ability to easily query data from multiple sources and query in time snapshots of historical data
- Data quality processes that ensure consistent, trustworthy data
- Business Analytics: Business Analytics take data and theories and mix them to create models that can help provide insights that will allow organizations to better improve their business decisions and results.
There are three types of Business Analytics:
- Descriptive analytics: utilizes reporting and visualizations to create a picture of what happened based on historical data
- Predictive Analytics: uses statistics and machine learning algorithms to predict future outcomes
- Prescriptive Analytics: utilizes simulation and optimization, sometimes from predictive models, to instantly recommend or perform actions
The rise of big data leads to many questions: Is Big Data really that useful? And can it solve a new set of problems, as well as the problems originally and previously solved by conventional analytics? Will implementing big data cost more than a data warehouse? And do organizations have enough expertise and the right tools/technologies (including both open-source and proprietary) available for different use cases?
The question most of us face today is: what is missing in the data warehouse based analytics that we are talking about BIG Data analytics today? What gaps/loopholes the new technology/approach will fix. Or, what new use cases it will cater, which was not possible to address through the conventional approach.
If we look at the data warehouse based approach, we will observe a few disadvantages/limitations:
- More historical data based reporting/analytics
- Inherent inflexibility
- Processing huge data – Time-consuming
- Overly time-consuming/long duration projects
- High maintenance
- High demand for resources / significant consummation of bandwidth, CPU etc.
- Early Binding – Integrating Business Rules Then Analysis
- Data Homogenization