Information Aggregation Usage

Information Aggregation Usage

Identifying business usage patterns provides an understanding of application utility both in the short and long term. This lets you better target the desired IT response. Here we classify the uses of the Information Aggregation pattern within a business according to business function. This can be used quickly to identify the user's requirements, the likely application or tool configuration, and the needed Application runtime pattern.

Business intelligence is applicable to all employees, partners, and customers of a company. Employees are usually given the most freedom of access. Partners and customers normally only access the data by pre-built Web business intelligence applications that, while offering business intelligence capability, restrict the user in scope of access and analysis.

Employees and internal users access data using a variety of data manipulation, query, and reporting tools. Often information is applied at a higher level within the company hierarchy, requiring custom-built applications providing information using EIS-like (Executive Information Systems) systems.

BI business usage fits into one of these categories:

• Strategy and planning-oriented: Planning the company's strategy and operations.
• Product-oriented: Managing brands or products.
• Marketing-oriented: Discovering new and better markets for company products and services.
• Sales-oriented: Improving the process of turning a lead into a sale according to the channel that handled the sale.
• Service-oriented: Focusing on customer service, supplier management, and business support.
• Process-oriented: Improving the internal business processes that run a company.

BI usage patterns details

The following BI topics contribute to roll-up information for executives, for example, as items on a balanced scorecard.

1. Strategy and planning-oriented

Use BI to plan the company's strategy and operations.

Characteristics:

User

business strategy planner, senior manager; usually internal use only

Delivery

report (for example, balanced scorecard and standard management reports) or analytical tool based.

Data

summary and detail business transaction reports organized by time, possibly with drill-through requirements to detailed data for event analysis; competitor data

BI role

analysis of trends; storing historical business details or trends; possibly data mining including risk analysis

Likely tool/application

OLAP tool; also traditional BI tools such as Brio, Business Objects, Cognos; (Also COBOL, Excel SAS); "Mart components": DataBase 2™ (DB2) (Universal DataBase (UDB), SP2, OS/390), Visual Warehouse

A typical analysis in this area would be:
Given the trends established from a set of queries previously performed against the warehouse, what will be the result if policy premium is increased for product X?
This type of query is usually performed on local (mart) data, as it often consists of "what if" type analysis in which data values are modified to perform the query. Because it is inappropriate to have users modifying data in a shared data warehouse, these queries are normally run on a local copy of the relevant data. The query tool usually makes this copy.

The actual tool used for business planning might be a specialized tool in which analysis and prediction models already exist. In many cases, this tool is a conventional spreadsheet application. The user imports data using the query tool and builds spreadsheets to show the results of varying business decisions.

Ideally, tools used for business planning are straightforward to use and shield the user from the physical data by using metadata.

Statistical analysis

Statistical analysis is a specialized form of the generalized business planning and querying process outlined above. This is a simple example of a statistical analysis question:

Taking a set of data on product sales performance, determine the statistical significance of the various factors thought to influence this performance and allow business decisions to be made on future modifications to these factors.

Many of the tools appropriate for query processing and business planning can also be considered for the purposes of statistical analysis, though their statistical functions and methods are fairly restricted. For example, the insurance industry uses sophisticated statistical methods to analyze product performance, underwriting performance, and so on. Certain vendors specialize in providing tools capable of handling both this level of complexity and the large volumes of data required as input to the analyses. Due to the complexity of the analysis, these specialized applications are used by highly skilled individuals. The user interfaces are less intuitive than those provided for more general analyses. SAS is an example of this kind of tool.

2. Product-oriented

Use of BI to perform brand or product management.

Characteristics:

User

business developer, product-competitiveness analyst (including actuarial rate-setting); usually internal use only

Delivery

report or visual tool based

Data

competitive product data; marketplace data; detail and summary business results

BI Role

analyze and manage large volumes of business data; perform various data mining activities

Likely tool/application

Relational Online Analytical Processing (ROLAP) tool; also traditional MI tools such as Brio, Business Objects, Cognos; (Also COBOL, Excel SAS); data mining tools; "Mart components": DB2 (UDB, SP2, 390), Visual Warehouse; IM4RM, FAMS, UPA
This is similar to the planning usage but based more on the business behavior of the brand or product set.

3. Marketing-oriented

Use BI to help the marketing department discover new and better markets for company products and services. It is also used to identify prospective customers, to contact those prospects, and to manage the marketing process to develop them into sales or retain them as future prospects. Divisions of responsibility between sales and marketing for this process can vary between companies.

Characteristics:

User

product marketing or external agent; an emerging trend is to push marketing materials to customers/ prospects through Internet e-mail

Delivery

reports, intranet, extranet, Internet

Data

product details; customer (and prospect) details; company details

BI Role

support campaigns (prospecting); manage large volumes of business data; store and relate relevant externally acquired data; provide data hand-off to sales; provide market segmentation analysis; maintain customer/prospect profile information

Likely tool/application

"Mart components": DB2 (UDB, SP2, 390), Visual Warehouse; applications/solutions: Valex, Insurance C&PO, IM4RM (data mining for customer segmentation)

Marketing leads can be distributed to external agents for follow-up (such as Independent Financial Advisors in the case of Financial Services). Materials can be pushed directly to customers/prospects, based on profiles that are partially defined by the customers and partially deduced/refined based on subsequent response patterns. New trends include using "click-mining" on company marketing Web sites to identify areas that appear to be of greatest interest to the public.

4. Sales-oriented

Use of BI to help improve the performance of the company's sales organization through superior analysis of the process of turning a lead into a sale, according to the channel handling the sale. As cross-selling increases, the need for sales support systems becomes increasingly critical because the "seller" might be trained primarily in customer service, for example.

Characteristics

User

internal or external sales agent (campaign or opportunity driven, using mail, telephone, in-person calls, email, and so on)

Delivery

reports, possibly through an intermittently-connected laptop, intranet, or extranet

Data

product details; customer details; company details; campaign details; competitive information; market information; summary business results; sales process data

BI Role

link from marketing, and provide above details; link to service

Likely tool/application

OLAP tool with drill-through; also traditional MI tools such as Brio, Business Objects, Cognos. (Also COBOL, Excel, SAS); data mining tools (for example, for propensity-to-buy analysis); DE for sales; "Mart components": DB2 (UDB, SP2, 390), Visual Warehouse

Marketing leads, created as above or otherwise acquired, are ideally followed by sales. Leads converted into sales are reported back to the sales department. Leads are assigned to different sales channels and effectiveness is measured. Then service BI measures the true value to the business of the items sold and the process used to do so.

5. Service-oriented

Use BI for customer service, supplier management, and business support.

Characteristics:

User

internal or external service agent (mail, telephone, or e-mail driven); increasingly the customer at home

Deliver

reports, intranet, extranet, Internet

Data

ODS; product details; customer details; company details

BI Role

provide consolidated client details

Likely tool/application

OLAP tool; also traditional MI tools such as Brio, Business Objects, Cognos; (Also COBOL, Excel, SAS); data mining tools; "Mart components": DB2 (UDB, SP2, 390), Visual Warehouse

Service-oriented BI recieves of the results of the marketing and sales area. It is usually the most process-oriented part of the business and has links to process BI. It performs activities such as order-tracking, guarantee-tracking, reponse to customer change requests, warehouse supply, and so on. Increasingly, the service center of the business is also expected to sell products and services.

6. Process-oriented

Use BI to improve the internal business processes that run a company.

Characteristics:

User

manager; usually for internal use only

Deliver

report (occasional and standard management reporting) or visual tool based

Data

workflow-type operational process measures; business and staff effectiveness measures

BI Role

storing historical details or summaries; possibly fraud detection mining; possibly risk management mining

Likely tool/application

traditional MI tools such as Brio, Business Objects, Cognos, and so on, with or without multi-dimensional facilities; (Also COBOL, SAS); "Mart components": DB2, (UDB, SP2, 390), Visual Warehouse

Such information is usually used by internal staff only and is predictable and stable after it is established. However, current trends to move customers to perform more self-service of their relationship with the business, and straight-through processing for lower services costs, will lead to greater flexibility and delivery demands on this pattern. This area can be difficult to distinguish from other department functions. If, for example, you are working with a marketing department, carefully distinguish between BI applications that improve marketing capabilities like Lead Generation (which do not fit in this category) from BI applications that improve the marketing process (detailed within this category). You might still implement a single solution for both problems. Process-oriented BI should consider the implications of the process on other departments that are affected by or involved in the complete business process. However, marketing (for example) also requires BI applications that are specific to the department's particular business activities. Process-oriented BI is designed and implemented best with a global perspective of the business process.