CHAPTER 5

Alphabet Soup—Big Data, Cloud Computing, DSS, ERP, VoIP …

The amount of information a business needs to acquire, process, transfer, store, and protect to remain competitive is increasing ever more rapidly. Eventually, the size, complexity, and rate of accumulation of data becomes too much for a business to process and manage in a timely and accurate manner using only internal resources. In 2001, Doug Laney described this situation in a Meta Group publication¹ using the terms volume, velocity, and variety to describe the increased amount of data, the reduced amount of time allowed to process it, and the increased lack of structure in that data.

Some large enterprise companies have the technical skills and capital resources to develop their own suite of software applications for managing these huge data sets associated with their operations, accounting, purchasing, planning, forecasting, customer relations, research, and other activities. An example of the typical set of applications used today by a large company is shown in Figure 5.1.

Figure 5.1. The alphabet universe of information applications used by a typical large enterprise company.

However, most organizations do not have either the skill or the resources to develop their own solutions, particularly if their core business offering is not highly technical or information intensive in nature. This is particularly true for small- to medium-sized businesses (SMBs) who might have some staff with the requisite software skills, but what staff they may have is usually needed for other business needs. Small businesses typically use spreadsheets for internal information processing combined with a choice of the many small accounting and invoicing programs commercially available for handling customer and financial information. Another significant limitation for SMBs is the operating expense to support an internal hardware and network infrastructure to handle large sets of data.

As organizations grow in size, their management teams increasingly look outside for more comprehensive software solutions and infrastructure providers to manage their business processes. Transitioning part or all of their information processing needs to external suppliers is often frustrating and more difficult because many employees often prefer using Excel or other spreadsheet software solutions for their personal information processing needs. This is particularly true if the transition requires major changes in how information is entered and retrieved to adapt internal processes to those used by the selected vendors.

Selecting an enterprise software solution best suited for your business is not easy. Some of these software suites are well integrated, that is, they make the best use of information across an organization to reduce errors and costs while also improving performance. Some other suites are a collection of essentially standalone applications that usually work very well within different internal host functions, but, because they are not well integrated, sharing and comparing information among the functions can be very difficult. Other disadvantages of a nonintegrated approach are the creation of duplicate data systems, isolated functional area decisions that often conflict with decisions made elsewhere in the company, and increased operating expenses.

In addition, many businesses do not have a good understanding of what their processes do, what is required by each process, what other processes are dependent on each process, and, in some cases, are unaware that a process critical to their business even exists. The smaller the business, the more likely this is to be true. These hidden processes, I will call them “ghost” processes, often come to light when an employee retires or otherwise chooses to leave the company. Sometime in the past, it is often likely that the employee encountered a problem that occurred often enough that he or she developed a process for dealing with it, incorporated that process in their daily work, but did not document it.²

Even when a business has done a good job of understanding and documenting its more critical processes, problems can arise when the business grows large enough to want to take advantage of outside software solutions and information infrastructure suppliers. SMBs need to recognize that commercially available solutions³ are based on using standard sets of common processes designed for different classes of businesses such as retail sales, manufacturing, food products, service industries, and financial firms. Often, the processes in an SMB are quite different, sometimes for good reason because of the particular uniqueness of that SMB product or service offering. But, consequently, this situation requires either a substantial revision of the SMB’s processes to use an externally developed enterprise-level software solution or considerable expense and delay to have the providers of a solution adapt their software to the SMB’s needs. A caveat to be considered here is that during the initial set-up and growth of a small business, its management would be wise to adapt standard processes for common activities and only develop custom processes for those parts of the business where they can provide a competitive advantage.

A combination of the situations described in the previous two paragraphs often results in one of the horror stories in the business and financial press about a company that tried to implement a new enterprise software solution and the result was disastrous for the company. Sadly, in today’s news, environment disasters are usually more interesting news than successes. More companies benefit from adapting enterprise software solutions to improve their business performance than companies who do not benefit from trying to do the same. That said, there are lessons to learn from such failures and it needs to be recognized that some successful implementations were more difficult to achieve than others.

Big Data

The term “Big Data” used now even by the public news media and cited in some recent governmental programs is still relatively new to the information technology environment where its use is the most appropriate. While its origin is hazy⁴ and its definition is still in the forging stage, Big Data is rapidly becoming part of the IT vocabulary in many businesses. When does the amount of information used by a business grow large enough to be called Big Data? In other words, how big is Big? The simple answer is “It depends.”

In appendix A, “Big Data” is defined as “A collection of data whose very size, rate of accumulation, or increased complexity makes it difficult to analyze and comprehend in a timely and accurate manner.” This means that its use depends on the viewpoint of the user of a given data set. What is huge from a small business viewpoint is likely to be a small daily consideration for a large enterprise organization. For the purpose of this monograph, a better definition would be that a Big Data situation exists when the amount of information used by a business grows large enough, is accumulating fast enough, or is increasing in complexity (becoming more unstructured) enough that the business needs to seek outside assistance and resources to use its information effectively. Hence the scope and size of the outside solution will be commensurate with the amount of information, that is, a wide variety of business intelligence (BI) solutions are available to:

We could really get mired down in the weeds here attempting to describe all of the BI solutions currently available and what are the advantages and disadvantages of each solution. But it would be a futile effort, given their equally rapid evolution to keep up with the increasing amounts of information. In fact, this book required literally continual revision during its preparation to keep up with the changes in information technology and the new management concerns those changes have created.

That said, one application reported recently in the news media illustrated how Big Data methods were used to provide information about the spread of influenza outbreaks more quickly than was previously possible using traditional medical reporting techniques. It is important to note that the individual data entries analyzed were not the result of answers to questions like “Is there a flu outbreak in your area?” or “Do any of your friends have the flu?” Instead, this was done by correlating the number of online searches conducted by individuals using flu-related terms with their location and other data such as posts and tweets about being sick, change in demand at major e-tailers for illness treatment items, and so forth.

Another example might be an analysis of the GPS-related transaction described in chapter 4 where a smartphone user submits a query on his or her phone regarding where they might buy an item they want and what the available price is at each location. Because advertisers like to focus their smartphone advertisements on people likely to respond to them, it would be useful to know the average distance a customer is likely to travel in response to the answers to their query. In addition, is there any correlation between that distance and the price of the item, that is, will the customer on average travel further for a lower price and how much? All that is needed is the customer’s location when they make the request and POS data at the store where they choose to buy the requested item. Obviously, some customers value their privacy more than others and will turn off their smartphone’s GPS locator and pay cash for their purchase, thus reducing the amount of data available regarding their particular transaction. But their request is still processed over the Internet using a cell phone network, providing some data regarding the rough location of the request and when it was made. By processing thousands of such transactions every day, even with parts of each transaction’s data missing, Big Data analysis techniques can provide a reasonably accurate estimate of the average distances traveled and other typical values of interest regarding such transactions.

The solutions for the largest data sets depend on the use of the Internet or specialized networks for connectivity, or both, the sharing of many hardware systems for processing and storage, nonrelational database designs, and software systems designed to manage such analysis using parallel processing. Such solutions are too expensive and often unnecessary for many SMB applications. Depending on the nature of an SMB data set, the information processing solution involves choosing what the business can best do internally and what it needs to export to external providers. One approach growing in popularity is the use of cloud computing for basic data storage and information processing applications; in essence, returning to the basic concept of central mainframes accessed by user terminals used in the 1950s and 1960s. The big differences are that the mainframe is now likely to be a collection of hardware systems located in different geographic locations and the terminals are likely to have more individual computing power than mainframes of the 1980s and early 1990s.

Cloud Computing

There are a number of descriptions regarding what is cloud computing or is not, all of them lacking different aspects of the technology depending on the interests of a particular group of users. Perhaps the use of the term “cloud” came about because of this ambiguity. Cloud services use a combination of computer and network hardware installations that are usually geographically distributed in areas where the utility and other operating costs are low. These installations are connected together using the Internet and managed in a way that appears to each business user as if they were using a dedicated set of workstations and a single storage system. In reality, this management process is dynamically sharing a business’s data storage and information processing across a number of available servers and network connections. This virtualization approach enables the cloud service to use each of their pieces of equipment at near its full capability, reducing the operating expenses and capital investment required to provide the desired level of service to their customers. This allows multiple instances of a word processing program to be running at the same time for different users without conflict.

If we look at the basic services provided by the more common view of the cloud computing concept, it is not unlike the use of mainframe computers in large enterprise companies back in the 1960s and 1970s as mentioned in the previous section. Users accessed these large-for-the-period computer systems via terminals to use them for information processing, retrieval, and storage. Cloud computing is essentially an updated version of that solution with a number of new features and some significant differences. Some of the advantages of cloud computing advertised by its providers are as follows:

While it can be said that all of these claims are true, what is missing is the potential effects on a company’s risk management, security, and operations strategies. What are the disadvantages of cloud computing? To paraphrase an old adage, “One should be cautious about putting all of one’s IT cookies in one jar!” Because the managerial questions that should be asked here are very similar to those that should be asked when considering the use of any information processing or handling solution, we will discuss those questions together with other considerations later in chapter 6.

Some cloud computing disadvantages and risks are as follows:

Like all disadvantages, there are strategies to prevent some of them from occurring or to mitigate their effects when they do occur. It is important to develop such strategies before they are needed.

Decision Support Systems

Like cloud computing, there is no single definition that encompasses all of the manifestations of what could be called a decision support system (DSS).⁶ My first boss could easily be classified as a decision support system because of all of the mentoring he provided me. For that matter, a dart-board on the wall of your office with areas labeled with various decision choices could be considered a crude decision support decision by some. However, for the purpose of this discussion, we will consider only DSS software applications.

The term “support” is important because we are talking about applications that do not make actual decisions here. The final choice is left up to the user. The application helps the user in making that choice by processing large amounts of data and providing the results in a summarized format that is easier to interpret. Recall our earlier discussion in chapter 2 about human beings being much better at comparing things than quantifying them; DSS applications often process quantitative data so that it can be presented in a format for comparison. Excel Pivot Tables are one example of this approach. An example of their use for comparing two store locations is given at the end of appendix C.

In the 1980s, a number of software applications were written for individual users and managers as the availability of personal computers in the workplace increased. Among these was a group of programs intended for use in making decisions in areas where there were no defined formulas available that a user could use to avoid their need to make a judgment call for a decision. The terminology often applied to such applications was artificial intelligence (AI) or expert systems based on the expectation that a computer could be programmed to analyze complex sets of often ambiguous data and make a decision that could be supported by users with limited knowledge about the situation. The idea was that by correlating a number of decisions made by recognized experts in a field with the sets of data they used for their decisions, one could come up with a set of rules that a program could use to analyze future sets of data and make a decision for the user entering the data.

The more popular areas of interest for these applications were employee performance evaluations, credit approval decisions, and other processes where an expert’s experience and intuition was often a major part of such decisions. The problem, of course, was that the technology of the day made it very difficult to quantify such experience and intuition in any meaningful way. As a result, such applications faded in popularity and all but disappeared by the end of the decade. Today, the intuitive and experience components of decision making are usually left up to the user.

Recent advances in predictive analytics and the growth of Big Data information have created a new interest in AI and expert system applications. A number of applications are being developed for decision making in situations where the influence of personal experience or intuition is perceived to be quite small. The decision rules in such applications are constantly updated by real-time analysis of the results of previous decisions and the sets of data used to make them. That is, the application learns from its past success and failures and adjusts its decision algorithms to increase the probability of success for its next decision. One place this approach is being put to use is preventive maintenance scheduling.

There are three categories of DSS applications determined by the scope of their coverage:

As the scope and complexity of decision support needed by a business grows, more care must be taken in choosing an appropriate DSS application. For example, an application designed for the decision support needs of a hotel business will likely be a poor choice for a manufacturing-to-stock business and vice versa. In addition, cloud computing users must consider whether or not a cloud-based DSS solution is a better choice than an internal DSS installation, especially if their other processes are using a combination of internal and cloud-based applications.

For more information about DSS concepts and its terminology, see the book by Power,⁹ chapter 2 on collaboration systems in Kroenke,¹⁰ and chapter 9 on decision making in Baltzan et al.¹¹

CRM, SCM, PDM, MRP

A pantheon of standalone software applications for managing the relationships between customers, internal business functions, supply chain partners, vendors, and other business operation areas has come into existence since the use of information technology and databases became available at the individual level. These applications manage the information that needs to be shared and collect and analyze performance data with the goals of identifying potential problems quickly, improving process results, and reducing operation costs. Depending on the company’s business model, some applications may become an integral part of another larger application, such as an ERP program, especially if the information both applications use and share is more effectively stored in a common database.

The following discussion describes a few of the more commonly used applications, the data of greater interest to each of them, and the business environments that are more likely to use them. It is important to recognize that whatever set of applications a given business chooses to use, the databases containing the information they need must be managed in a manner that prevents unnecessary duplication of data and inconsistent methods for entering any shared data such as customer or supplier names, invoice numbers, and filenames. We will discuss procedures and management strategies for achieving this in the data standardization segment of chapter 6.

CRM

Customer relationship management uses POS data, warranty return information, technical support records, customer satisfaction survey entries, and other customer transaction information to maintain and possibly improve the relationship between a business and its customers. Sometimes the CRM application is managed by a company’s marketing function; in other businesses it may be a tool for the sales organization and expanded to cover the customer ordering processes.

SCM

A close cousin of CRM (in fact, they may be one and the same in some businesses), supply chain management applications can range in scope from just managing the interactions between a business and its suppliers to a more full-blown application managing all aspects of the supply chain, including vendor selection, supply quality monitoring, purchasing process management, shipping, warehousing, and retailing (particularly if the company is an e-tailer). Information of particular interest to SCM applications is lead-time information for purchasing, shipping times, in-process losses such as theft and handling damage, shipment errors such as incorrect quantities and wrong addresses, pipeline inventory, and economic factors.

PDM

Product data management applications are more common in manufacturing businesses because of the substantial amounts of information required by today’s fabrication processes. Tracking changes in specifications, bills of materials (BOM), and machine settings is a common need along with managing process control and performance data and maintenance information.

They also can be used by businesses whose products are based on information to track document revisions, inventories of printed materials, clip-art collections, and image files.

MRP

Materials requirements planning applications were among the first information processing applications to be developed for businesses. First introduced in the early 1960s, they were used by sales and operations planning (SOP) functions to manage their production schedules and inventories of materials, purchased parts, and finished items using demand estimates and summaries of the bills of materials for each product to be produced. Today, many small manufacturers are able to duplicate the capability of earlier versions of MRP by using a combination of linked Excel spreadsheets. I developed an example of such a solution using an imaginary company making two products to create a series of assignments for operations management students. One Excel sheet contained a monthly demand forecast for the coming year, and another contained the bill of materials (BOM) for the two products with associated assembly times or purchasing lead times for each item on the list. A third sheet contained inventory information for each item, and purchasing data such as lot sizes, minimum safety stock levels, and order placement frequency for each item that need to be purchased. A fourth sheet contained the master production schedules for each product on a monthly basis and the fifth sheet showed all of the purchasing or production schedules for each item in the BOM. Students were shown how much of the schedule information on the last two sheets could be obtained automatically using appropriate formulas linked to selected parts of the information available on the first three sheets. For a more detailed example of how such information is used in a simple MRP program, see the work posted by J. E. Beaseley¹² on the Internet. The URL for his posting is listed under his name in the References.

MRP-II

As time went on and businesses became more aware of the usefulness of an MRP application, more capabilities were added to include financial information, human resource requirements, and capacity constraints. To differentiate from earlier versions, the newer versions are now called manufacturing resource planning applications (one of those all-too-often situations where the meaning of an acronym used in business depends on the context of the associated discussion.)

Enterprise Resource Planning (ERP)

Enterprise resource planning applications integrate many of the applications previously discussed with other business functions such as accounting, transportation, training, quality, maintenance, and marketing to form a single comprehensive business knowledge management solution. While many services businesses consider ERP to be useful only to a manufacturing business, many of the commercially available ERP offerings are readily adaptable for larger service providers. Some of the advantages of an ERP application are:

Okay, it looks like ERP solutions can be pretty useful for a business. If so, how is it that some businesses, even large companies that would seem to benefit the most, have not yet implemented one? The answer is complicated because there are a number of reasons, some valid, some not. Some of these reasons were briefly discussed at the beginning of this chapter, but deserve repeating and some additional clarification here as they relate to ERP. Many of the reasons also apply when considering the implementation of standalone applications such as CRM or SCM.

First, ERP solutions have a poor reputation in some circles because of a number of high-profile implementation failures reported in the business media a few years ago. For an article about some of the more spectacular ones, see Wailgum.¹³ This history coupled with more recent news about other failures has caused many companies to hold any consideration of implementing ERP solutions until the track record appears to be improving. Some other reasons are as follows:

Many of these difficulties can be eliminated or at least reduced if a business begins looking forward to a time when the business may need an ERP system to remain competitive. This means keeping future ERP requirements in mind when improving existing processes or establishing new ones. This also means increasing the level of understanding regarding how information is used, stored, transferred, or accessed in the business. Developing an integrated information and process management methodology for the business will help support those strategies and is in alignment with most ERP approaches.

Scheduling Applications

Scheduling back office activities, customer appointments or reservations, shipments, workforce assignments, vacations, preventative maintenance, and job shop projects are a challenge for many small businesses that cannot afford enterprise scheduling solutions such as those used by major airlines and transportation businesses. Even determining how long to make the average appointment period for a hair salon or doctor’s office is complicated because of all of the variables involved—varying times required per client, some clients may arrive late, how much time to allot for tidying up between clients, and so forth. I addressed this problem in my last book,¹⁵ using an example of a typical doctor’s appointment schedule. It is reproduced here for your convenience.

In the example, the fact that some of the time elements vary in value, that is, they are stochastic in nature and are characteristic of many values when dealing with customers, makes even this relatively simple determination somewhat complicated to perform. Imagine how much easier it would be if the time to take care of each client never varies, that is, it is a deterministic value and everyone always arrives on time.

For job shops making an assortment of standard products, the scheduling of which product to make next and by who is often made easier by assuming standard times for making each product. However, what do you do if two of the workers have to use the same machine at the same time? How do you handle a priority job for a preferred customer? How do you handle custom jobs that may take more or less time to do, are your workers equally capable or do they operate at different levels of proficiency?

There are a number of ways small businesses can deal with such scheduling problems and some of them will be discussed in more detail in a following book about process modeling and improvement. In the meantime, a useful reference for those of you interested in more details about developing your own scheduling solutions is the book by Pinedo.¹⁷ You should be warned that his book requires some knowledge of higher-level mathematics to take full advantage of what he has to say.

Communication Applications

The use of landline telephone systems for business communications is declining in favor of online communications and cell phone service. Recent surveys reported in the news media indicate that more than half of adult Americans now own a smartphone and roughly one-third of them own a tablet device. A number of factors are contributing to this change.

A number of VoIP (voice over Internet protocol) communication applications are available to help SMBs replace their clunky PBX (private branch exchange) installations in areas where there is adequate broadband access to Internet. Combined with a BYOD policy to reduce a business’s investment in communication devices, these solutions can save money and increase a business’s capability to adapt quickly to changing communication needs. By combining them with wireless networks for a company’s computer systems and peripherals, office and other process arrangements can be changed easily to accommodate new employees or reorganize groups for new projects without having to move phone lines and Ethernet cables and reprogram the PBX.

Video communication applications can help businesses with more than one location or a set of geographically dispersed supply chain partners conduct face-to-face meetings between staff at different locations without the need for the delays and costs of travel. The use of these applications for online education allows students to have conferences with their professor and work with other students on projects.

Some large global companies have their own video conferencing solutions and lease dedicated fiber-optic cables to transfer the large amount of data required by the multiple displays and audio channels used to create an environment that simulates a real meeting room for the participants. For one demonstration of what this level of technology can do see the online promotional video¹⁹ for the Halo solution developed by Hewlett-Packard a few years ago. A quick search on the web yields a number of similar demonstrations because a number of similar solutions have been developed by major communication service providers.

Although video conferencing solutions such as Halo are currently much more expensive than a typical SMB can afford or justify needing that level of performance, the seemingly never-ending pace of technological development is likely to make that level of capability in the future as affordable for an SMB as a high-definition flat-screen computer monitor is today. In the meantime, SMBs can take advantage of a number of less-expensive video communication solutions for one-on-one video chats, small video conferences among business colleagues at different locations, and members, and online training seminars, often referred to as webinars. Some of these applications are now used worldwide. One example familiar to the most of us is the use of Skype by the news media for interviews with correspondents in far-off places.

Before closing this discussion of video communication options for businesses, it should be noted that the interest in video communication is not new as evidenced by articles published as early as the 1950s and the development of the PicturePhone²⁰ by Western Electric during the 1960s and 1970s. The infrastructure needs of the PicturePhone were ahead of the technology of the time, a situation that still plagues many new product offerings today, and its size and overall awkward appearance did not attract enough customers to justify its further development.

Custom Applications

A number of unique information processing and management applications have been developed to address the needs of special situations taking advantage of the advances in information technology. One familiar situation is the monitoring, control, and decision processes used to manage a nation’s power distribution grid. Other not-so-familiar situations are the management systems for two large river systems in the United States, the Colorado River Basin Project,²¹ and the Office of Columbia River.²²

The scope and complexity of these river management systems make the large ERP implementations for our major corporations look like child’s play when one considers the enormous variety of customer groups interested in the management outcomes, the wide range or regulations that they must operate under, the resources they have to manage are limited and the demand for them is growing, and they have little control over their variability. Furthermore, many of the demands on the different interest groups and the various state and national regulations conflict with one another. It is the ultimate make-the-best-of-a-bad-situation management challenge.