Decision Support Systems

Taking right decision at the right time in business lead it to peak of glories and in this digital age business analyst cannot rely on old ways of analysis to take decision to drive industry. When it comes to plan new market strategy to overcome the effect of competitors campaigns and hold the customers with trust then within hours or in some cases within minutes decisions are required to make. Here comes decision support system that helps the industry to forecast future and make the right movement in the right direction.

Decision support system is any interactive software that captures a picture of business and shows negative or positive indicators of it which helps the individual to take decision based on these patterns. These patterns which can include revenue of the company or sale of some product are derived from raw historical data and other data repositories of the organization. We’ll discuss decision support system in the light of data warehousing and business intelligence tools.

Data Warehouse

When data is extracted from various sources and assembled in a meaningful way on a wide centralized database then this process is called data warehousing. Various sources of database may include flat files, OLTP databases, logs etc. Data Warehouse teams identify important attributes of each source data that can value to business and extract them to warehouse and discard the remaining part of data which is not required. Cleaning up data and discarding unnecessary information is called transformation. After which data ware house loading tools load the data into target tables. This all process of loading is called ETL.

1- Extraction

2- Transformation

3- Loading

Evolution of Data Warehouse

Different business units of an organization possess their own transactional processing systems. When organization grows to an enterprise level then it becomes difficult to perform analysis on each of these systems separately. Moreover reconciliation of these systems becomes a big challenge. Another disadvantage of OLTP (Online transactional processing) is that they are not designed to maintain history and they usually contain data of few days. A centralized data repository then becomes necessary to interlink these sources of data to make OLAP (Online Analytical Processing) system for analysis. Data Warehouse is that centralized OLAP system that extract data from these OLTP and maintains history. This is first evolution of data warehouse which is focused towards reporting from single view of the business. Questions are usually known in advance like weekly sales reports, monthly revenue report etc. When warehouse starts maintaining history then trend analysis of the business is performed like why revenue dropped in the first week of the month or segmentation of customers by analyzing buying behavior. Loading of data in such ware houses performs usually in the form of batches in night window. The reason to populate data in night is due to off business hours hence less load of data on the source systems. Forecasting is the next stage of data warehouse in which questions like what will happen next in the business are asked and analyst by seeing historical patterns of the market through data warehouse proactively manage the organization strategy. This technique is called data mining and separate tools for data mining are available to integrate them in data warehouse environment.

Active Data Warehouse

Data warehouse finally evolves into operational phase also called active data warehouse. Active data warehouse implementation allows near real time decision making. Data from sources loads into ware house on real time basis which allows the management to take vital decisions based on actual market situation.

Active Elements

The definitions of the ADW active elements are quoted from Teradata: Active Data Warehousing.

• Active Access: Front-line operational decisions or services supported by NRT access; Service Level Agreements of 5 seconds or less.

• Active Load: Intra-day data acquisition; mini-batch to near-real-time (NRT) trickle data feeds measured in minutes or seconds.

• Active Events: Proactive monitoring of business activity initiating intelligent actions based on rules and context; to systems or users supporting an operational business process.

• Active Workload Management: Dynamic management of system resources for optimum performance and resource utilization supporting a mixed-workload environment.

• Active Enterprise Integration: Integration into the Enterprise Architecture for delivery of intelligent decisioning services.

• Active Availability: Business Continuity to support the requirements of the business (up to 7x24x365).

Example

Mr. Johnson is a regular customer of dwhtech telecom. He usually calls in the evening and likes to text his friends a lot. Mr. Johnson makes a phone call to the call center of dwhtech to get some information. When the call arrives in the call center, the number goes to Decision Support System or data warehouse to fetch some product which can be advertised to him before agent picks up the call. Data Warehouse runs a script to see last six months call details of the customer. After performing analysis on the call details of Mr. Johnson, Data Warehouse picks night package (say Superb Night) which allows hundreds of SMS daily from 8 PM to 6 AM on minimal subscription charges. The package also includes 20 free minutes of talk time daily. Data Warehouse pass on this package name to call center application which instead of playing music in the background, plays an already recorded IVR advertising this package to customer while he is waiting for an agent to pick his call. This is an example of active data warehouse in which customer behavior is analyzed on real time basis and appropriate campaign is advertised.