Business Knowledge
Information systems and their components of hardware and software function to enable people and processes to use data to make organizations healthier and support business decisions based on data analytics. This unit examines how data stored in databases, data warehouses, and cloud systems drives that process.
Business knowledge — also known as knowledge management (KM) or business intelligence (BI) — refers to the gathering, organizing, sharing, and analyzing of the data and information to which a business has access. The data stored in a repository can then be organized, shared, and analyzed to support decision-making at every level of the organization.
That data comes from many sources: from within the business itself (organizational memory, workforce experience and skills, documents about customers and suppliers, existing designs and processes) and from outside (market research and reports, customer and supplier conversations, professional associations and trade bodies, trade exhibitions and conferences, and collaboration with partner institutions). Knowledge management may not mean the same thing to every company, however.
Some organizations define KM in terms of productivity gain. Others frame it around data sharing of business intelligence rather than the sharing of knowledge itself. Still others focus on employee portals — sources where staff can access the specific data needed for their jobs. Regardless of definition, the goal is consistent: KM involves spreading knowledge of individuals and groups across the organization in ways that directly impact performance.
Why KM Matters
There are three key reasons why knowledge management is important for a business.
- KM facilitates decision-making capability. Processing an overwhelming amount of information can block high-quality decisions. Scientists estimate the amount of data sent to a typical person in a year is equivalent to every person in the world reading 174 newspapers every single day. A knowledge management system that can make sense of all this data enables better, more informed decisions.
- KM builds learning organizations by making learning routine. Capturing learning from experience builds knowledge that can be used to streamline operations and improve processes over time.
- KM stimulates cultural change and innovation. KM programs can help managers remain open to change and foster an environment receptive to ideas and insight — conditions that lead to innovation regardless of business size.
All of that data needs to be catalogued, sorted, filtered, and linked in order for intelligent data analysis to produce results a business can use. This involves data mining in addition to data analysis. The goal of KM is to provide the business with output that can be used to address specific business tasks and projects. Understanding that goal requires examining the foundations of data storage — databases.
Database Concepts
A database is a computer-based collection of related pieces of data organized so that the data can readily be accessed, managed, and updated. Computers work with the binary system where a bit is the smallest entity represented — either a one (on) or a zero (off). Eight bits combine to form a byte, which represents one character. Characters combine to form meaningful data such as a name or address.
Basic Organization
Databases organize data into meaningful groupings that move from the smallest unit of information upward to the complete database itself.
| Level | Description | Example |
|---|---|---|
| Field | Contains one or more characters, or an audio, video, or image file. Designed to hold a specific type of data (character, numeric, image, hyperlink). Visualized as a column in a table. | First name, last name, street address |
| Record | A collection of related fields, organized as a row. Contains information related to a specific entity — a person, place, thing, or event. | First name + last name + social security number |
| Table | A collection of related records. Contains all records for a particular group or type of thing or event. | All client records for a fund management firm |
| Database | Made up of related tables and other objects such as queries, forms, and reports that help users view data in meaningful ways. | Complete fund management system with clients, managers, and investment tables |
Interacting with a Database
Database management system (DBMS) software allows users to interact with the database at all levels of the hierarchy. After a database structure is designed and fields and tables are created, data is typically entered via a data entry form — a window on screen allowing the user to input values into fields that make up a record.
- Query — a way of retrieving data by specifying criteria that identify exactly what data is to be retrieved and how it might be sorted and displayed. Data can also be modified via a query.
- Reports — used to retrieve data in a format the user specifies. Reports can be displayed, printed, shared, or used for things like mailing labels.
- File maintenance — procedures that keep data current by supporting the adding, modifying, or deleting of records, and creating backup copies of the database.
- Wizard — a software application used to create tables, queries, and reports. The wizard itself is not used to enter or modify data directly. Microsoft Access uses wizards built on the SQL language.
Database Models
All databases are composed of fields, records, and tables. But the organization of those records and tables can differ significantly. Different database models represent different approaches to structuring data, each with distinct strengths and limitations.
Hierarchical Model
In the hierarchical model, the structure is predefined. Data is organized in a tree-like parent-child relationship. The top field is the parent and lower fields are children. A parent can have one child; a child can have one sub-child, creating a strict 1:1 relationship at each level.
The key limitation is that there is no way to relate the children of one parent with the children of another — no common key field exists. To retrieve even one piece of information about an item, the entire record for that item must be opened. This all-or-nothing constraint made the model inefficient as data needs grew more complex. If new fields are needed, the entire database must be redefined.
Network Model
The network database is an extension of the hierarchical model. It allows a parent record to have more than one child, and child records to be related to more than one sub-child record. This creates a 1-to-many (1:N) relationship, making it more flexible than the hierarchical structure because new relationships can be established between data. However, because the structure still needed to be defined in advance, it remains fairly inflexible in practice.
Relational Model
The relational database model grew out of a need for greater flexibility in adding new fields and tables, and to retrieve just the information desired for a particular purpose. Instead of each record containing all the information about a person, place, or event, the information is spread across different tables.
These tables are connected by a primary key — a common key field inserted into each record to link records across separate tables. If an identification number is the primary key, each record in the various tables associated with an individual will share the same ID number as one of its fields. New fields or completely new tables can be added without rebuilding the database, as long as the primary key relationship is maintained.
Object-Oriented Model
In an object-oriented database, data itself is conceived of as objects. An object can consist of data (character, numeric, etc.) or instructions on what to do with that data. Whereas a relational database stores all the elements that make up a dog — nose, eyes, mouth, ears, body, legs, tail — in separate fields, an object-oriented database stores all these components together in one object.
Object-oriented databases are especially useful for design, scientific experiments, telecommunications, geographical information systems (GIS), and multimedia such as photos, sound files, and video. These applications rely heavily on images and multimedia data types that cannot easily be stored in relational databases where fields are typically character, text, or number based. In object-oriented databases, tables are linked and accessed using pointers instead of common key fields.
Multidimensional Model
The multidimensional database is built and used optimally for data warehouse and online analytical processing applications. A special database language called online analytical processing (OLAP) or multidimensional OLAP (MOLAP) is used to manipulate data in these types of databases. The multidimensional model is considered the best database model for data warehouses and data mining operations.
| Model | Structure | Best Used For | Key Limitation |
|---|---|---|---|
| Hierarchical | Tree, strict 1:1 parent-child | Simple, predefined structures | All-or-nothing retrieval; inflexible |
| Network | Graph, 1:N parent-child | Complex relationships between record types | Structure must still be predefined |
| Relational | Tables linked by primary keys | Most business applications; SQL-driven | Can be complex to design at scale |
| Object-Oriented | Objects containing data + methods | Multimedia, GIS, scientific data | Less suited to traditional tabular data |
| Multidimensional | Data cubes for multi-axis analysis | Data warehouses, OLAP, analytics | Specialized language (OLAP/MOLAP) required |
Integrity and Validity of Data
What is common to all database structures is the need for ease of entry, accurate retrieval, and the assurance that data in the database is correct and trustworthy.
Data integrity ensures that data can be verified as correct, is up-to-date or timely, is organized in a useful way, is accessible when needed, and is cost-effective (its value is greater than the cost to produce it).
Data validity is accomplished by comparing data being entered against a set of predefined rules to ensure the entry complies with those rules. For example, if text data is entered into a field set up to accept only numbers, the user should be alerted immediately to the mismatch.
Advantages of Databases
All database models support the following advantages over a flat file processing system (in which data is stored in files with no connections between them).
| Advantage | Description |
|---|---|
| Reduced Data Redundancy | Duplicate data is more easily avoided since information is stored once and referenced by key fields. |
| Improved Data Integrity | Changes are made in one place instead of searching through multiple files or spreadsheets. |
| Shared Data | A single set of data can be shared with multiple users. Security settings define who can access, add, modify, or delete records. |
| Easier Access | With appropriate software and access privileges, a nontechnical person can use the database without needing to understand its underlying structure. |
| Reduced Development Time | Database creation tools result in an easier and faster development process compared to developing and maintaining multiple separate files. |
Challenges
- A database system may be more complex than a series of spreadsheets and may require people with special training to design and implement it.
- A database consumes more memory, storage, and processing power than a simple file processing system.
- Because a great deal of information is stored in one place, if it is lost or the data becomes corrupt, it may affect all users who need to access it — a single point of failure without proper backups.
- Unauthorized access to a database containing personally identifiable information (PII) could result in harm to the individuals whose information is accessed.
Structured Query Language (SQL)
The primary language used in relational databases is called Structured Query Language (SQL). SQL is the most commonly used database language, with most applications using it for analyzing and manipulating data, including in Database Management Systems (DBMS).
SQL is used to create, modify, maintain, and query relational databases. When you use wizards in Microsoft Access to create databases, enter data, and retrieve data for reports, those wizards are built on the SQL needed to manipulate the database.
A simple SQL query follows this structure:
FROM CLIENT
WHERE ManagerID = '1';
The query above would return the contact information for all clients assigned to manager ID 1. SQL allows very precise retrieval — only the fields and records that match specified criteria are returned, with no need to retrieve entire records as in hierarchical models.
SQL offers a platform-independent, web-based database solution that can be embedded in many other computer languages. Microsoft Access is a common DBMS that uses SQL to meet the needs of small to medium-size organizations.
Relational databases are usually built and accessed via a relational database management system (RDBMS) — a program that lets you create, update, and administer a relational database. Most commercial RDBMSs use SQL to access the database, although SQL was invented after the development of the relational model and is not strictly required for its use.
A more recent development is the object-oriented database management system (OODBMS). The development of OODBMSs was driven by the perceived shortcomings of RDBMSs in managing complex, highly interrelated information such as that found in computer-integrated manufacturing environments.
Data Visualization
Data visualization is the graphical representation of information and data. Charts, graphs, and maps can be used to summarize and illustrate data so that it can be used as information for decision-making by organizations. Tableau and Google Data Studio are examples of data visualization applications.
Data Warehouses
As organizations have embraced technology as a centerpiece of operations, most deal with vast volumes of data that need to be stored in a secure, accessible manner. A data warehouse provides an organization the ability to store and extract data from their database for analysis. It is a large database that stores and manages data required to analyze historical and current transactions.
A data warehouse should meet the following conditions.
- Non-operational data. The data stored should be a copy of the original data that can be accessed as needed, not the live transactional data itself.
- Time-variant data. Stored data has a timestamp, which can be used for validation purposes and historical trend analysis.
- Standardized formatting. Data is standardized so it can be consistently queried and compared across sources.
Design Approaches
A data warehouse is typically either bottom-up or top-down in design.
| Approach | How It Works |
|---|---|
| Bottom-Up | Starts with small data marts (subject-specific subsets of a warehouse) which can be combined to create components that satisfy queries and reports. Useful when departments have immediate, specific needs. |
| Top-Down | Starts with an enterprise-wide data warehouse that can then satisfy queries by creating data marts as needed. Provides a more unified and consistent data view across the organization. |
Benefits of Data Warehouses
- Allows the organization to see how data is collected and processed, providing visibility into data flows.
- Acts as a centralized repository that can integrate with enterprise systems such as CRM and ERP platforms.
- Stores copies of data with timestamps, allowing organizations to validate data and information and track changes over time.
Cloud Computing
Cloud computing has emerged as a way of providing IT services over the Internet. Cloud computing services like SaaS, PaaS, and IaaS can reduce service costs compared to the expense of maintaining IT infrastructure and staff in-house. Outsourcing to the cloud is fundamentally a cost-driven business decision.
Cloud service providers allow clients to reduce costs by selecting only the services they require and canceling them as necessary — acting like telephone or utility companies that offer a menu of services clients can choose from. Cloud computing provides certain advantages including ease of access from different locations, scalability, predetermined maintenance, and high security and control.
Cloud Service Models
| Model | What It Provides | Example Vendors |
|---|---|---|
| Software as a Service (SaaS) | Provides software applications to clients through the cloud. Typical office applications such as email, word processing, and spreadsheets can be delivered. The only requirement on the user's end is a thin client with a web browser. Users pay for only what they use and can cancel unnecessary services dynamically. | HP CloudSystem, Dell IT Management SaaS, IBM CloudBurst, Symantec.cloud, AT&T |
| Platform as a Service (PaaS) | Provides platforms on which applications or software can be developed. Usually web-based. Allows developers to build, test, and deploy applications without managing underlying infrastructure. | VMware View, CA 3Tera AppLogic |
| Infrastructure as a Service (IaaS) | Provides basic computing services including load balancing, routing, virtualized operating systems, and basic networking. Organizations rent virtualized hardware instead of purchasing physical servers. | Various enterprise providers |
Thin Clients and the Cloud
SaaS delivery often relies on thin client computers — devices designed to reduce expensive hardware and software by using browsers to access backend servers on which most of the processing power and data reside. Users do not own any of the software, hardware, or infrastructure offered by the cloud service providers, which means they can scale usage up or down on demand without capital investment.
Database Security
As data volumes grow and databases store increasingly sensitive information, security becomes a critical concern for every organization. It is the confidentiality, integrity, and availability (CIA) of data that needs to be protected.
- Confidentiality can be lost if an unauthorized person gains access to a database, or if someone authorized to view selected records accesses records they should not see.
- Integrity is compromised if data is altered by someone unauthorized to do so.
- Availability is impacted if those who need to access the database and its services are blocked from doing so.
Why Security Matters
Information stored in databases — client information, payment data, personal files, bank account details, and more — can be hard or impossible to replace. Loss can result from physical threats (fire, significant power outage), human error (data entry mistakes, unintended deletion), or malicious activity (corporate espionage, theft, unauthorized access).
Both businesses and home users should be concerned. If you have used a credit card, made an airline reservation, subscribed to a magazine, been a patient in a hospital, shopped at a chain store, or used a social media site, your information is stored in a database.
Database Breaches
Database breaches are the exposure of database records containing personally identifiable information (PII) or other sensitive information to unauthorized viewers. In 2013 alone, over 822 million records were made available to persons with no authority to view them. Reported incidents likely undercount the true number, since some companies do not report breaches in order to protect their reputations.
| Breach | Records Exposed | Data Compromised |
|---|---|---|
| Adobe (2013) | 152 million | Customer IDs, encrypted passwords, debit/credit card numbers, order information |
| Unknown organizations | 140 million | Email addresses and identification numbers of South Korean individuals (North Korean hackers) |
| Target (2013) | 110 million | Customer names, addresses, phone numbers, email addresses, credit/debit card numbers, PINs, security codes |
| 70 million | User email addresses exposed via a flaw in the site's API |
The business sector accounted for 53.4% of reported breach incidents, followed by government (19.3%), medical (11.5%), and education (8.2%). Hacking was the cause of 59.8% of incidents, accounting for 72% of exposed records.
Customer Relationship Management (CRM)
Customer relationship management (CRM) is a foundation element for business knowledge and intelligence. Also known as relationship marketing or customer management, CRM is a strategy used by companies to manage all facets of business-to-customer relationships. Robust CRM systems are supported by software suites that help manage all the data acquired and used by the system.
The primary applications supported by a CRM are acquisition (obtaining new customers), retention (retaining current customers), loyalty (developing customer loyalty to the company or product), profitability (increasing company profits through customer service), and service (addressing customer inquiries and resolving issues).
Traditional CRM
In traditional CRM, there is relatively little collaboration between the customer and the company. Marketing's focus is to push messages to customers to generate sales. The primary focus areas of any CRM system include:
- Providing the marketing department with information needed to identify and target the company's best customers, design effective campaigns, and give the sales team quality leads.
- Optimizing information shared among departments, resulting in increased sales, better management of existing accounts, and support for the latest communication devices.
- Improving customer satisfaction by supporting the development of individualized relationships with customers, including providing the highest level of service to the most profitable customers.
- Obtaining and sharing with employees the information and processes necessary to effectively build relationships by understanding and identifying customer needs.
Data Analytics
Almost all CRM applications involve using a large relational database, sometimes referred to as a data warehouse. The raw data about customers, products, transactions, demographics, and other information is stored there. The data warehouse typically receives its information in real time from point-of-sale (POS) systems, e-commerce applications, inventory management systems, and other transactional platforms.
Data mining is the process of looking at data stored in a company's database to determine if statistically relevant trends exist. By identifying these trends, companies can develop strategies to better serve customers and increase sales. Using complex statistical analysis software known as data mining tools, analysts can query the data warehouse in many targeted ways.
| Open Source Tools | Commercial Tools |
|---|---|
| R — The most popular big data analytics tool. Integrates well with big data platforms. Known for a steep learning curve. | SAS — Long considered the leading data analytics tool. Versatile and easy to learn, with specialized modules for IoT, anti-money laundering, and midsize business analytics. |
| Python — Covers a host of statistical and mathematical functions. Also used as a data-gathering tool through web scraping. Supports social media analytics. | Tableau — Excellent for creating visualizations and dashboards. More robust than Excel for large datasets. |
| Apache Spark — Focuses on unstructured data or huge data volumes. Integrates with Hadoop, a Java-based open-source framework for large data sets. | Excel — The most widely used analytics tool. More accessible for non-analytics professionals. |
| Apache Storm — Used for moving data or when the data is continuous. Works well with real-time analytics or stream processing. | QlikView / Splunk — Popular visualization tools with user-friendly interfaces. |
| Pig and Hive — Used by companies that work with Big Data and leverage the Hadoop platform. | — |
Social CRM (SCRM)
Customers can now communicate with companies through website chat and social media accounts such as Facebook and Twitter. CRM systems that include social media integration are now essential for many companies. These systems are known as social CRM (SCRM).
In contrast to traditional CRM, SCRM invites the customer to collaborate with the company in solving business problems, primarily through interaction with online social media sites. This empowers customers to shape their own experiences and build relationships directly with the company. Dell reports that customers posted over 18,000 new product ideas and almost 100,000 comments through their social platform, with nearly 500 ideas actually implemented.
SCRM is also a vehicle for direct and indirect advertising. Participation by companies in the social media environment has led to customer experience management (CEM) — a broader approach in which the customer is no longer a passive recipient of the company's services but an integral part of the customer experience and relationship development.
Social media advertising strategies available to companies include direct advertising via banner ads, sending ads through networks of friends or contacts, increasing brand awareness through groups and fans, using the company's own social networking site, and viral marketing in which individuals pass along embedded marketing content to others.
CRM Trends
CRM and the software that supports it is not a static product. To remain effective, organizations must be sensitive to the following trends.
- Customers expect more. Generation Y customers increasingly prefer web chat, smartphone applications, and social media over telephone support. Meeting customers on their preferred channels is now a baseline expectation.
- Keeping valuable staff. Experienced customer service staff are leaving their positions at higher rates. Training, support, and up-to-date tools are needed to retain CRM staff as their duties expand beyond phone calls to include social media and mobile interactions.
- Privacy. Companies are learning more from digital customer interactions. Users may be willing to share data if they understand how it is used, but they also expect it to be protected. Companies need to foster trust.
- Mining social media inputs. Customers post honest insights about products and services on social media — data that companies would ordinarily pay to gather through surveys. Social media analytics tools help marketers and sales personnel observe and respond to customer sentiment in real time.
- Cleaning up data. Inaccurate or duplicative CRM data can hamper call center and sales employees. Organizations need to invest in ongoing data quality efforts.
Conclusion
Databases are the foundation of modern information systems. Whether organized hierarchically, relationally, as objects, or as multidimensional structures, they provide the storage and retrieval capabilities that make business intelligence possible. Understanding how data is structured at the field, record, table, and database level is essential for working with any information system.
SQL gives organizations precise control over relational data, allowing them to query exactly what they need without retrieving unnecessary records. Data warehouses extend that capability by providing a centralized, time-stamped repository for historical and analytical data. Cloud computing further expands access to these capabilities through SaaS, PaaS, and IaaS models that eliminate the need for costly on-premises infrastructure.
Security is not optional — the CIA triad of confidentiality, integrity, and availability must be protected at every layer. And CRM systems demonstrate how all these components come together to serve real business goals: acquiring customers, retaining them, building loyalty, and driving profitability through data-driven decisions.