Benefits of Data Warehouse (Part3)
The relationship of market types and technologies of data warehousing
Common idea that the amount of stored data such as measured in Tera bytes (TB) for Data warehousing is necessary to forget. Also if Information system has 4 TB of data for 1.5 million customers in some way be justifies being proud on the company information system, but also recalls for caution since it could also mean hundreds of generated reports and none of them actually were not read. Value of generated reports is not only in number of pages. Information about using 4 TB of data itself is probably useless, ie, it could only mean a large amount of numbers that are in the warehouse from which it has little value if they are of low quality. The primary question that business leaders need to know is what the data warehouse has to serve. For what type of questions data warehouse must provide answers.
- Data Warehouse may be a way to display data from production systems that are already adapted to the analysis of data entry – Reporting function.
- Another often neglected aspect is the fact that the data warehouse as a system serves for data storage. Of course data warehouse is usually both, but the question that has a primary role. In production system data is deleted periodically, and if it is not then it slow system. If production system is not slow dispate large data quantities this could mean purcased hardware was at the time of instalation overpaid and unnecessary. It is certain that it does not have to keep your payment receipt from permanently disconnected subscribers for decades. On the other hand, these data can be useful in statistical terms, and then it comes to other essential functions of a data warehouse to the prediction of customer behavior based on past behavior. The production system can satisfy most of reporting functionalities (they are more complicated to do than if they are not generated in data warehouse technology) but without historical data. Such reports from production systems can not never supply historical context for reporting. The fact is also that the data storage is somewhat neglected in recent years in the IT industry since the development of explosive speed that has brought it mobilized a lot of IT experts to develop this segment. In following list is presented an overview of usually stored data – frequently deposited information.
- Detailed sales figures for the financial and tax purposes.
- Data on trends of use of services and purchase of goods that have strategic value.
- Information required by various state offices (eg statistical office, etc.).
- Medical records generally must be kept very long time.
- Documentation of toxic substances, etc. which also must be kept forever
More information in Data Warehousing PDF.
- Reporting in the classical sense and prediction of customer behavior (such objective analysis generally) justify the investment in data warehouse. Prediction of the behavior of client that production system can not efficiently provide.
Predicting behavior of customers will depend on the categorization and on two important parameters from market: the interaction between consumer and product selection coincidence. Looking at these two parameters can be analyzed 4 combinations or 4 types of markets that are really environments for business companies:
- Linear market (interaction of consumers and product selection coincidence are low). An example of waste enterprise in which most customers do not care too much for the service (and indeed do not think of company if it performs as expected (if such waste is collected regularly). On the other hand the number of users is also subject to low fluctuation. Fluctuation, customers churn is almost neglectable because population is relatively constant in analyzed area. Predicting customer behavior is very accurate and possible, because last year’s behavior can tell a lot about the behavior of consumers in current year.
- Statistical market (small interaction of consumers and HIGH accidental selection of products). Sales of household goods is a good example of this. Although some consumers know who will probably buy detergent at the same time are not so sure if they need windshield cleaner. Time of the year still affect whether customer will be better to sell air conditioning or snow chains for cars. Correlation of different sizes in data stock is definitely a useful and new information about consumer behavior. Especially if we have a well developed attributes of the buyer (which is usually not the case because most sellers may know that the buyer is a man of 42 years but whether the household has 3 or 7 members, it has been something else which definitely defines its purchase)
- Chaotic markets (HIGH interaction of consumers and small product selection a coincidence) A good example is the high fashion market where the taste is formed on the basis of interaction between consumers and what was at one point “in”. Data warehouse in such conditions do not provide any new information.
- Real-time market (HIGH-consumer interactions and randomness LARGE selection of products) stock financial markets is a good example of real-time market. By itself it is clear that, unfortunately, data warehouse will provide answers to questions where to invest in the past years, although dwh has great historic data and data about customers and everything. The market is still complicated as chaotic market, the number of parameters that influence the purchase are large, crucial parameters are often secret.
The trend that CRM and Internet imposes is the quality of communications between manufacturer and a buyer. Internet is increasing interaction of consumers (see for example the reader reviews of books on www.amazon.com or hundreds of sites that deal with high end electronics http://www. hi-fiworld.co.uk /). This trend contributes to the growing interaction between consumers (growing and chaotic real-time market) which is one of the factors why today’s data warehouse terms are less effective in their role as predictive of behavior in the future. Terabytes of data will remain just terabyte of data for the named markets.
Benefits of Data Warehouse (Part1)