CHAPTER 4

Use and Acquisition of Information

We will begin this chapter with an aerial view of how information may be used by a business, then spend some time discussing specific details about uses in different functions, followed by some methods for acquiring information more easily and with less cost, and conclude with some suggestions for reducing the effort required to handle information in manufacturing and inventory applications.

Many of you should already be familiar with the B2B, B2C, C2B, and C2C terminology describing the four basic e-commerce interactions between businesses and customers where the first letter represents the creator or supplier of data and products and the second letter represents the receiver or user of data and products. For those of you who are not, Figure 4.1 should bring you up-to-date with a few examples of e-commerce businesses whose product or service is best described by each interaction.

Figure 4.1. Interaction matrix for businesses and customers. The use of the Internet as a communication path is assumed in all cases.

Figure 4.2 adds some new interactions to support the growing presence of Big Data applications, Internet-of-Things (IoTs) concepts, the effect of social media, and the increased use of the Internet for interactions between the governments and their citizens and businesses. This has expanded the role of the Internet from that of being a primary communication link between businesses and customers to also becoming an independent information source and a method of direct communication between various types of equipment. The types of interactions have also increased in complexity.

Figure 4.2. Interaction matrix for businesses, customers/citizens, governments, machines, and the Internet. The role of the Internet as a communication path is assumed in all cases. Note that since the Internet is considered to be a single entity, describing interactions with itself is not applicable here.

The term customer normally describes an individual who has a direct financial relationship (money changes hands), that is, acts as either an individual buyer or a seller. Sometimes, however, a customer has only an influential or informative role in the interaction, for example, when the customer is responding to a survey sent out by a business, posting a review regarding a product on an e-tailer’s website, or seeking technical support. Another growing type of online interaction involving individuals is the interaction between citizens and their government. Today, a number of governments have increased their presence on the web by encouraging citizens to submit applications and make payments for licenses, sign-up for governmental benefits, buy postage, and file their taxes electronically. In Figure 4.2, these types of relationships are still represented by the customary C for a customer (or citizen), but it is still important to recognize their differences.

The increased number and complexity of the interactions that now take place using the Internet is another illustration of the increasing growth on information.

Classes of Interactions

Defining each of the 24 interactions shown in Figure 4.2 separately would take more time and space than is appropriate for this monograph. Hence, we will summarize them according to the viewpoint of the primary sender or initiator of the communication in an interaction, recognizing that most interactions are necessarily two-way exchanges.

Business

As more and more information became dematerialized in the past decade, business-to-business interactions primarily moved online to support the growing need to exchange information with suppliers, warehouses, and retailers quickly. This includes sales to other businesses in a supply chain.

While many purchases, particularly material items needed quickly (food, beverages, and gasoline), are still sold to customers in brick-and-mortar stores, the percentage of total sales done online and sent directly to customers continues to increase. This trend requires businesses to review frequently how they collect point-of-sale data, market their product, communicate with customers, and handle warranty and quality issues. Of particular concern to traditional brick-and-mortar retailers is the growing practice often referred to as “show rooming,” where a customer checks out a product in their store and then buys it online from another retailer for a lower price. Easy price comparisons made possible by advances in cell phone technology such as the ability to scan product barcodes are another disturbing trend for such retailers.

The use of the Internet by a business to communicate with a machine is relatively new, but it allows some considerable advantages because communication with such equipment is considerably simpler and can use display and analysis capability that is already in place. This allows remote control and monitoring of manufacturing equipment, security systems, and faster acquisition of information from a variety of sensors.

As the number of Internet users has grown, more businesses are using the Internet for advertising, catalogs, posting product information and manuals, training, and generic technical support (support for a specific customer’s problem is a B2C interaction). The practice of moving from printed manuals and guidebooks to e-books and other downloadable files dematerializes these information products for more rapid and flexible distribution and eliminates printing, stocking, shipping, and revision costs.

Customer

Customers and citizens are, in my mind, the biggest beneficiaries of online communications. Never before in the history of mankind has an individual had the ability to access so much information easily and quickly. The development of easy-to-use and intelligent search engines and the increasing availability of social media sites, blogs, and online education courses covering a wide range of interests allow anyone with a web connection to acquire information from around the world. I am particularly amazed at the range of video content posted by individuals that provides do-it-yourself (DIY) information,¹ musical performances, lectures by leading scientists and scholars, tours of scenic sites and historical monuments, and so forth. As discussed in the previous paragraph, businesses are now learning the value of joining this information-sharing environment, not only to reduce the costs of providing product information but also as a new advertising medium.

The classic example of the C2B interaction in many business textbooks is a customer negotiating with a business regarding the price that the customer is willing to pay for that business’s product offering, most commonly a hotel room or a similar perishable product such as an airplane seat, entertainment event, or a Christmas tree still on the sales lot the day before Christmas. But, given the ease of online access and more immediate response, more customers are also contacting business websites for product information or training, to submit product reviews, ask for technical help, or process warranty claims. In some instances, individuals act as temporary sellers to business. Some examples are selling textbooks back to online book dealers or selling used electronic devices to businesses that refurbish for them for resale in other countries.

Government

Ironically, governmental agencies were the primary users in the early days of what evolved into the Internet. Their presence became secondary during the dot-com growth in the 1990s and then expanded later. Today, many governments are increasingly using the Internet as the primary means for communicating with their citizens and conducting business with them. Electronic filing of tax returns speeds up revenue collection and reduces processing costs. Similar savings and reductions in response time are possible for other activities such as benefits claims for veterans and seniors, provided such agencies update their processes to dematerialize their paper information and enable them to take advantage of the advances in information technology. Governments can also use Internet connectivity to acquire weather and traffic information from monitoring equipment, perform surveillance, activate warning systems, and so forth.

The challenges here are how to handle two significantly different types of users—those who are familiar with using the Internet and those who are not, often older citizens who frequently have a greater need to communicate with governmental agencies. Another complication is providing equal access for all of a country’s citizens. Many rural areas in even developed countries currently have no broadband access, limited Internet capability, or none at all. This is also a problem for small and medium businesses operating in such areas.

Machine

The idea of connecting machines to computers for data acquisition and control applications is far from new as we described in the history segment of chapter 1. What is new is the ability to connect machines with computers in two or more locations anywhere in the world without using hardwired interfaces. The use of cell phone networks and the Internet allows equipment and sensors in one location to communicate easily with other sensors and equipment in other locations, enabling the use of multiple automatic monitoring systems and the coordination of activities among facilities. This also allows equipment to contact customers or businesses directly to notify them of a changed condition or other event of interest. Many machines can now be programmed remotely over the Internet. The variety of applications made possible by this connectivity, often referred to as the Internet of Things (IoTs), has exploded in recent years. Some more familiar examples, besides the obvious security monitoring applications, are programming your DVR at home to record a favorite TV show while you are still traveling, weather and traffic webcams located in many cities, webcams monitoring baby owls in a nest for scientific study by naturalists and entertainment for the public, sensors monitoring the structural integrity of the new interstate highway bridge in Minnesota, and children checking on the daily health of their aged parents living some distance away. While the real usefulness of the solution is yet to be determined, there are even refrigerators that monitor their content and notify their owners via the Internet when to pick up some more milk or eggs on their way home.

We will talk more about these types of applications later in this chapter and subsequent chapters because they can offer inexpensive capabilities to SMBs and their requirements must be considered when selecting external software and infrastructure solutions.

Internet

We normally do not think of the Internet initiating an interaction on its own, but many businesses have developed applications that ask it to do exactly that. To clarify what I mean, I am not referring to the product suggestions relative to your interests that pop up on an e-tailer’s website when you are viewing it with the intention of ordering something else. That is more correctly described as a B2C interaction with a frequent customer. Instead, the Internet information sites often act as the primary intermediary for a collection of businesses much like old-time print newspapers did. While some would call it acting as a general advertising medium, the difference is that the Internet can target that advertising on an individual basis using the reader’s location (GPS coordinates, zip code, or IP address) and previous browsing history. In addition, the Internet has communication advantages that print media does not have, such as sound, video, animation, and the ability to download information in both electronic and print forms.

The businesses do not know who each customer is until that customer chooses to contact them in response to an advertisement—a C2B interaction after a B2I—I2C set of interactions. Keep in mind here that the potential customer went on the Internet to read the news, not scan advertisements for stuff the customer might be interested in buying. If a customer wants to buy an item, the customer will either go directly to a trusted e-tailer they think has what they want or they will use a search engine to find one. Clicking on a pop-up advertisement is generally not wise unless one recognizes the advertiser and has a strong antivirus and malware detection program monitoring their Internet connections.

Acquiring Information

How we do this is governed by our current level of understanding and some basic measurement rules. That is, you cannot measure or assign attributes to something when you are unaware of its existence. Likewise, if you are aware of the existence of something, but you and other people who are also aware of it cannot agree on how to measure or describe it, sharing what understanding you have is very difficult. These situations occur more often in business situations than one would expect. A common example is a product failure or a poor service result that occurs occasionally for no discernible reason. To prevent them from happening in the future, you need to know the cause. However, until some piece of information is correlated with the undesired outcomes, you have no idea of where to begin your investigation. Big Data analysis methods can help find such correlations when you have accumulated large amounts of data, but you need to remember the fundamental rule “Conclusions are only as good as the information used to derive them.” or more colorfully stated as GIGO, “garbage in, garbage out!” Acquiring large amounts of information can give businesses false confidence that they have enough to solve problems or gain deeper insight into how their businesses work. This also carries over into many of today’s engineering design approaches where many experiments are run under a wide range of conditions to accumulate huge amounts of test data. The data are then analyzed to find correlations, probabilities, other relationships, and optimum sets of results to select the best design tried and perhaps identify what parameters to adjust for better performance.

In effect, they are playing a Battleship game against a small fleet of enemy ships in a very big ocean. The more experimental shots they fire, the quicker they can find the enemy fleet. But if they have limited supplies of ammunition, this is not an effective use of their resources. Consider, however, that they do some intelligence work first to determine what possible ports the enemy fleet could have left from and when. Then making an estimate of the fleet’s speed from those possible locations could help narrow the area of the ocean for their barrages of shots considerably.

There are two important lessons to learn here. The first lesson is that data-mining methods work with known parameters; they are very poor at finding unknown parameters other than maybe indicating that some may exist because no significant correlations are found in a data set. Sometimes you can identify some conditions that your experience tells you are relatively sure to affect your desired results, but you just cannot figure out yet how to measure them. This is an opportunity to gain a competitive advantage if you can. Some useful references in this regard are the book on business measurements by Hubbard² and chapter 5 on seeking for a pattern in noise, from the book by Silver.³

The second lesson is that collecting, storing, and analyzing data is a significant business expense, both in support costs and in diversions of employee time and attention away from the core processes of the business. Occasionally, spend some time to review what information your business really needs and what information is proving to be not worth the cost of acquiring it. Integrating the use of information with other process improvements as discussed in chapter 3 is an effective way of doing such a review.

Some Suggestions for Data Acquisition

There are many ways a business can acquire data. The trick, of course, is to use methods that are inexpensive and provide information most relevant to a particular business. Data acquisition processes that are an integrated part of other business processes are the best choice, particularly those daily operations processes required for producing the services or products a business offers to its customers. Some suggestions for sources of useful information are:

POS data are high-yielding ore for data mining. In addition to its original primary value of providing information for business accounting processes, it is the optimum source for acquiring all sorts of information regarding customer behavior and preferences. By choosing the appropriate format, POS data can also provide information directly to inventory management, staff scheduling, and marketing functions. With some innovation and careful design of customer ordering, shipping, and payment processes, a business can obtain greater informational content at its POSs at a minimal cost. In some cases, a business can even get their customers to do most of the input. One example of how to do this was described in chapter 3, Figure 3.6. Other common examples are the online ordering processes used by many e-tailers.

Customer Support Data

Customer support data are particularly useful for preventing future problems and obtaining information for directing efforts to improve products and production processes. These data should include customer returns and warranty claims. There are several advantages in combining these processes with other customer requests for help and information as part of a business’s website. The information is obtained in an easily processed dematerialized form, the customer does most of the work to enter it into a company’s database, and the response to a customer’s concerns is more readily available.

Performance Data

Most businesses collect some data about how well their more important processes are performing, usually with regard to cost, throughput, yield, downtime or uptime, and utilization or capacity. In large enterprise companies that can afford to use automated processes, a significant amount of these data is relatively easy to acquire. In many SMBs, most of these data are manually acquired and summarized by managers or senior staff at the end of a shift, day, or other reporting period most suitable for a particular business.

While useful for planning, accounting, scheduling, and monitoring purposes, these data are often inadequate for use in preventing problems, resolving problems quickly when they do occur, or for identifying more specific areas for process improvements. What is often missing is adequate time information. Just knowing exactly when different activities in each functional area are occurring and which person, machine, or information system is performing them at that time can be of significant help. This enables easier identification of trends for use in preventative action decisions and maintenance. The ability to correlate events across more than one company function or department helps quickly identify less obvious causes of process failures, such as a change in a material supplier or a modification of a support process. Most of this information can be added simply and inexpensively in dematerialized information systems by including their internal timestamps and identifiers in the data set.⁴

Monitoring Data

Monitoring information is a necessity in some businesses, particularly those handling food items, providing security systems, growing plants in greenhouses, supplying utilities such as electricity and natural gas, refining oil, or running any other operation where proper environmental conditions are critical to their success or safety. Monitoring such data in the past allowed us to respond quickly to an undesirable condition. It also was necessary to provide documentation to regulators that we had complied with the necessary requirements to ensure that our customers received a safe-to-use product. Many of these systems were either primarily mechanical or electrical in nature. Some of the monitoring devices allowed a possibility of recording their measurements versus time on paper, such as the circular temperature or pressure chart recorders that some older manufacturing workers are likely to be familiar with. However, many of them required manual reading and recording at specified time intervals, a process that was subject to errors in entering both measured values and the times they were taken.

Thanks to advances in measurement and information technologies, an increasing majority of such monitoring needs can now be recorded directly in electronic form. Such records not only satisfy the safety and regulatory requirements, but the data now can also be processed in real time to predict the possibility of an unsafe or other undesired condition early enough to prevent it from happening. Many of the preventive maintenance software solutions now available depend on such monitoring data.

GPS Data

The ability to output location information directly in digital formats has not only made navigation much easier, but has also enabled the easy use of location information for controlling processes, analyzing potential markets, targeting more appropriate advertising, and directly labeling images with their location. As a result, correlating data to locations has become an important business tool. This is especially evident in many decision support systems and logistics strategies. The expanding use of GPS data not only for internal business use but also for services and products such as enhanced mapping applications, survey equipment, and personal navigation systems has created a subdiscipline in the IT community called GIS (Geographical Information Systems).

The rapidly growing use of smartphones worldwide⁵ has created a corresponding growth in targeted advertising and other services based on location data. One common example is displaying nearby businesses on a person’s smartphone in reply to their search for a place to eat or stay or as part of a query about prices and nearby stores for something they want to buy.

Frequency Data

While it can be argued that frequency information is really part of the data sets we have already discussed, it deserves separate attention because of its importance in evaluating risks and making yield management choices. Most businesses have to cope with demand variations and external factors that are difficult to predict. For physical products with a reasonable shelf life, they can use a safety stock inventory strategy, but for products with a very limited shelf life or services offerings they need to anticipate these effects on their business as well as they can. One way to improve their chances of success is to improve their ability to predict what will happen next. Determining possible seasonal variations for customer demand and the probability distributions for the occurrences of external factors that might affect that demand or the business’s performance is usually a good place to start.

Consider a simple yield management problem where you own a hotel. If you do not have enough reservations by the time the rooms should be occupied, you lose money on the empty rooms. The incremental overhead expense for servicing each room must now be recovered with less revenue. The situation is the same for empty seats on an airplane flight. The common solution for such a problem is to begin offering discounts or other incentives to bring in more customers to occupy the unreserved rooms or fill the empty seats before takeoff. A common complication is that some of the reservations are likely to be cancelled at the last minute.

To cope with these last-minute cancellations where there is little time left for offering incentives and discounts, many hotels and airlines overbook their capacity when the demand is high by the number of customers they think will cancel at the last minute. Where do they get this number? Furthermore, what happens if they overbook too much? How do they handle the customers who must be turned away? What’s the probability of that happening?

This is why collecting frequency data is important. By counting the number of times in a given period (a time long enough to include all or at least most of the possible values) there were no cancellations, just one cancellation, just two cancellations, and so on, one can calculate the probability of each cancellation level by dividing the number of times it occurred by the total number of nights or plane flights covered by the period. If one is lucky, the resulting range of discrete probabilities may correlate closely enough to a standard probability distribution such as Poisson, normal, or binominal that the mathematical function describing the distribution can be used in future calculations regarding the best choice for an overbooking level. If not, then the set of discrete probability values can be used directly to determine the optimum overbooking level. In a similar fashion, frequency data regarding sales volumes for perishable items can help a retailer determine the optimum number to stock of each item.

Of more general interest to businesses is the value of frequency data to identify seasonal factors for more accurate forecasts of demand. The granularity of such data will vary, of course, from business to business depending on the nature of the business and the intended use of the data. For example, staff scheduling will likely require daily or weekly values and production of make-to-stock items may only need monthly or quarterly values. Such decisions regarding the collection of frequency data and the desired level of detail are critical for success when a business decides to incorporate the use of an enterprise software solution.

To do this using internal POS data and other information collected from past events, that information needs to include whether or not some of the external factors occurred at each event and more specifically which ones. Then, for some defined levels of performance such as lost money, broke even, made money, or really cleaned up, count the number of times each level of performance occurred for each combination of demand and possible factors. Then, by dividing the total number of times you lost money for a given combination by the total number of times you encountered that combination, you can obtain a rough estimate of the probability of losing money if you encounter that combination in the future.

For example, consider that your business is providing snack foods and beverages for outside events. The number of expected customers can vary for a number of reasons—the popularity of the event, the weather, other events at the same time, current economic conditions, and so forth. The types of snacks and beverages the customers would expect to be offered can also be affected by the nature of the event, the weather, economic conditions, and so forth. So, how would you choose what types of food and beverages to bring to an outdoor event when the weather forecast says there is a strong probability it will be rainy and cold? What quantity of each chosen food and beverage item should you buy in order to obtain the best return and satisfy your customers well enough that they would want to do business with you again?

One answer is to just give such choices your best guess based on what you can remember from the past experiences. A better strategy is to add some information to your business’s POS data when it is collected. The normal sales receipts provide enough demand data to answer questions about product mixes, but it will not explain, for example, that the reason hot coffee purchases were greater at some events than others is that the number of such purchases is typically higher on a cooler day or when it is raining. While you can add such information to future POS data, what about the large amount of past POS data collected by your business? This is where clever use of external data available from other sources and Big Data concepts can help give you a competitive advantage over other event providers.

External Data

The first seven suggestions relate to information that is best collected internally by a business to best capture the nuances in such data that are characteristic of that particular business. To supplement internal data to answer questions posed in the previous two paragraphs, you need to consult some of the many external sources available. Some are free such as US census data available online at http://www.census.gov/#, data from the European Union online at http://europa.eu/index_en.htm, and data from many other US government agencies online at http://www.usa.gov/Topics/Reference-Shelf/Data.shtml and http://www.data.gov/. An easy web search looking for information about other countries can direct you to a number of governmentally supported websites that not only tell you about their country and its government, but also provide useful business data regarding transportation capabilities, customs regulations, their primary business strengths, and so forth.

A number of businesses provide information for other businesses as their primary product. Marketing firms can do customer preference and satisfaction surveys or evaluate the effectiveness of a business’s advertising among their other capabilities. In the United States, a relatively inexpensive reference that has been useful for SMBs in assessing market potential, location considerations, transportation networks available, and other startup or expansion information is a commercial atlas such as the one published by Rand McNally.⁶

Two of the most significant sources available regarding customer preferences are the databases accumulated by online search engine and social network providers. Their databases are a result of the ultimate use of customers inputting information regarding their interests, preferences, dislikes, what they have questions about, referrals to other potential customers, and in many cases their personal contact and location information in real time. While many of the individual data inputters are not yet fully aware of the extent of the data they are entering, some are beginning to express concerns about the privacy of some of their information in social network applications. Such concerns also apply to SMBs that use these applications for communication with their customers. We will discuss these concerns further in chapter 6.

External data sources are especially useful for start-up businesses needing background data for their business plans. Because they obviously will not have accumulated much internal data yet, reviewing appropriate data already collected by other representative businesses can be useful when setting up their internal processes. Such data for more common processes are often available in local small business support organizations.

Methods of Handling of Information

Considerable effort is spent by businesses to input and otherwise handle the information they need for decisions, process instructions, and record keeping for a variety of accounting, tax reporting, and regulatory requirements. The information for businesses involved in supplying services such as medical examinations, haircuts, lawn care, tax preparation, house painting, and other offerings that do not require tracking of completed physical inventory⁷ has become more automated with either direct computer entry by the employees performing the service or by the customers requesting it. But, for many back office activities supporting those services and for businesses producing, distributing, storing, and selling physical products, there is still a need to physically track the status and whereabouts of those products and easily link that information to other business processes like inventory management, procurement, pricing, and POS systems.

A plethora of books, articles, and other references exists regarding how to best handle information for most common business processes. Here we concentrate on more recent technological advances in information handling applications for manufacturing, retail, and customer interactions with the observation that many of the advances in these areas are now being applied to other processes such as record keeping, tracking workflows, and providing information to employees and supply chain partners more quickly. Some of these advances that have made material handling and related information processing much more efficient, less expensive, and more accurate are:

Barcode

The use of barcodes is now ubiquitous in modern societies primarily because they reduce the time required to enter information accurately. In addition, that information is in a dematerialized form that can be readily processed, stored, copied, and transferred. Some examples of commonly used barcodes are shown later in this chapter along with a discussion regarding how the coding works and associated applications.

This technology had its beginnings in the late 1940s and early 1950s when a food chain president expressed an interest at the Drexel Institute of Technology regarding having an automated system developed for reading grocery items at the checkout station. In response, two graduate students at Drexel developed a method using a code consisting of alternating lines of varying widths configured in both circular and line patterns and read with an electronic device also of their design. Their invention was submitted for a patent in 1949 and it was approved in late 1952.⁸

However, their reading method required a very bright incandescent bulb and physical movement of either the detector or the code. In addition, it required the development of a sufficient number of unique code patterns to represent the various grocery items for it to be practical. This could have been done on a store-by-store basis with each food chain developing its own code, printing it on adhesive labels and affixing them to the respective product in their inventory. But the remaining roadblock was the scanning system: the bulb emitted a lot of heat, the scan was slow, and the overall size was much larger than a cash register. It required advances in electronics from vacuum tubes to transistors, the development of the laser, agreements by several different food chains and their suppliers on a standardized code for grocery items, and the development of some internal barcode applications by some major industries before the first grocery store installations occurred in 1973.

At first, the technology only provided a list of the items selected by a customer, although it certainly was faster than the cashier entering the list manually. Prices still needed to be entered separately. Stock boys at that time had to stick two labels on grocery products, one label for the code, and another label for the price. Later, the labels were improved to have both the code and the price on the same label and food manufacturers began printing the codes for their products on the cartons, cans, and bottles they came in.

Transportation industries began using barcodes to track their rolling stock such as rail cars, semi-trailers, and large shipping containers. However, because the reading distances were greater and the environments dirtier, such labels were often misread or were not readable at all. Some of these industries such as railroad companies changed to using RFID devices for this tracking.⁹ RFID devices were more expensive, but they were not discarded after use like the grocery store barcode labels; instead, they could be used a long time and their reliability was much better.

The advances in computer technology and decreasing cost of standalone computer systems combined with greatly enhanced barcode standards around the globe provided much more capability for all businesses to cut their material handling costs and increase productivity. Now the readers could not only provide the cashier with a quick list of the items to be purchased, but also the computer in the checkout station could correlate each item’s barcode with its current price in the inventory database, debit the inventory stock for that item, add up the total cost for the customer, calculate correct change, and print out time and day on the receipt. This also eliminated the need for stock boys to attach new price labels to existing stock when the price changed for a sale. But the cashier still had to manually handle each item to pass it over the scanner, enter the data by hand for smudged or missing barcodes, collect payment, handout any change, and bag up the items. Today, self-service checkout lines with automated payment capability allow the customer to do most of the manual activity with cashier assistance only needed for smudged or missing barcodes.

Another advantage is that a business or an individual customer can easily print their own barcode labels using equipment already available in businesses and households capable of using the Internet. This ability was important in the early days of the technology when manufacturers had not yet established the practice of printing bar-coded information on their packaging. Today it is used by consumers to print their own boarding passes for expedited check-in at airports, tickets for events that they have purchased online, and so forth. An application useful for SMBs is printing a custom menu of barcodes for employees to use when inputting inspection data, machine settings for a particular process, specialized conditions at checkout stations, customer complaints, or other defined sets of information used by the business.

A new application is the use of smartphone cameras for reading product barcodes. This allows users to check on a price for an item before reaching the checkout station and even compare it with the price offered at another store.

Like most things in life, there are some disadvantages to consider in using barcodes for data input. Some of them are as follows:

The biggest disadvantage and one that is hard to overcome is the barcode has to be visible in its entirety to read it correctly. This means no significant smudges or wrinkles and the package carrying the barcode has to be positioned so the scanner can see all of the barcode. Coupled with this requirement is the maximum distance from the scanner which is usually less than a few inches. In some industrial applications, a lens-and-camera system can be used to acquire an image of a barcode from a greater distance and that image can be analyzed by a computer process to decode the information. However, this is not usually a practical solution for many barcode applications.

At one time, the orientation of the barcode relative to the scanner was much more critical for a correct reading. Advances in how the laser beam is scanned in different directions across a barcode have eliminated this disadvantage in more recent scanner versions.

Another disadvantage is that the time required to scan a group of items is still controlled by the time to orient and scan each item. Shipping departments in large warehouses, express mail operations, and airport baggage-handling systems reduce this time as much as possible by specifying where to place the shipping label with the bar-coded information on a package and then to put the package on the sorting conveyor belt with the label facing up or to one side, depending on where the sorting system mounts its barcode scanners. Similarly, it is important for a warehouse to have a policy regarding initial placement of items on shelves so that their barcodes can be easily scanned without the need for rearranging the items to do so.

Many of the standard codes only provide identification information such as the manufacturer, the country of origin, and a product identifier. If a manufacturer wants to track other information such as product versions, date of manufacture, and expiration dates, additional custom-designed barcode labels are required along with requirements for their placement relative to the standard identification label.

Finally, barcodes cannot be changed once printed. This disadvantage can be partly offset by the practice of associating the relatively simple code on the label with more easily modified information in a company’s database such as the latest price of the item or its description. It becomes awkward when the same standard identification barcode in a warehouse inventory represents both older and newer versions of a product using this approach.

Development of Barcode Standards

The use of barcodes in processes related to consumers or external suppliers requires some form of standardization so that all of the participants are able to access and share the same information. Some examples of common barcode standards are shown in Figure 4.3.

Figure 4.3. Examples of some commonly used barcode standards: (a) UPC, (b) ISBN, (c) POSTNET (top) and Intelligent Mail (bottom), and (d) QR codes for the ISBN of the author’s earlier book on waiting lines (left) and the URL regarding this book on the publisher’s website (right).

While barcodes used only for internal operations can be any code the business wants to develop, it is best to keep that code as close as possible to standard formats in order to take advantage of commercially available scanning solutions. To clarify this statement, the UPC barcode standard specifies what each of the numbers of the code can represent when a merchant prints a UPC code on their product. Okay, some of you already familiar with the UPC code are likely to say, “Wait a minute! Each manufacturer is responsible for at least part of the UPC code, what about that?” (We will go into more detail about this later). In any case, when the UPC format is used only for an internal business need, those numbers can be used to represent anything the business wants. A part of the UPC standard supports such internal applications by using one value for the first number to indicate that the following numbers do not correspond to any values in the standard UPC database. Some common internal uses for the UPC format are part numbers, filing labels, lab test requests, and so forth. This allows the business to use a standard barcode reader to scan the code internally with the results correlated to the business’s internal database instead of the standard UPC database.

The UPC standard uses 12 digits to represent the country of origin, the manufacturer, a description of the product, and a check digit. The actual values represented by the various lines and spaces in the machine-readable pattern are printed at the bottom so that a human can read and enter the data if the pattern is too smudged or otherwise damaged for an accurate scan.

When a manufacturer applies to the Uniform Code Council (UCC) for registration of its products, the UCC uses the first six digits to assign an ID number to that manufacturer. The first digit is used to indicate particular classifications like random-weight items (2), pharmaceuticals (3), in-house use (4), and coupons (5). A particular useful classification (0) is used for abbreviated UPC labels to indicate that the code contains suppressed (not shown) zeros in the number sequence. This allows the use of UPC information on products with limited space for labels, such as beverage cans and bottles.

The next five digits are used for the item number and their assignment is the responsibility of the manufacturer in answer to the possible reader question given earlier. The last digit is a check digit that is used by the scanner to verify its accuracy in scanning the first 11 digits. Readers interested in more details about the UPC code and how it works are referred to a useful online reference by Brain.¹⁰

International Standard Book Number (ISBN) was developed initially as a 9-digit number more than 30 years ago by a statistics professor¹¹ at Trinity College in Dublin, Ireland, to simplify the identification of books for libraries and book retailers. Its value to the information processes used in this business segment increased its use to other countries, and it has been an international standard for identifying books and their publishers. Today, each country has its own ISBN agency for assigning numbers to books published in that country. Earlier, ten digits were used to represent the book information. On January 1, 2007, the ISBN was expanded to 13 digits to handle the growing publication volume. Older 10-digit ISBNs are still in use, but can be converted to the 13-digit format by using a calculator on the ISBN agency’s website at www.isbn.org. The agency emphasizes that an ISBN is not a barcode; it is the identification for a book. That said, these numbers are easily coded into an optically scanned format using the same methods as for UPC labels and are generally printed together with the ISBN on the book cover for easier data entry and tracking. Note: The barcode observed on the front of magazines and other periodicals is not an ISBN; UPC codes are typically used for these items. However, sometimes a special issue of a periodical will be sold as a book and it will have an ISBN assigned to it.

Books are identified by a 13-digit ISBN using five groups of information.

A new challenge is in deciding how the ISBN classification can be applied to electronic publications such as e-books. Because they are dematerialized information, there is no need for physical inventory tracking. Normal software file management solutions can handle distribution and storage needs. But, unlike most electronic files, there is a price associated with their distribution. So, how do we handle POS data for such transactions? Since an e-book is still a book, it would be simpler to use the same identification system as for other books.

Further complicating the situation is that an increasing percentage of the e-book publishers are individuals rather than the traditional big printing houses most of us are familiar with. An individual e-book author acting as one’s own publisher can currently get an ISBN assigned to his or her book by applying to the agency in their country and paying an appropriate fee. But, given the explosive growth of e-publications, this ability is likely to undergo significant changes in the near future. Stay tuned!

The information processing and handling problems encountered in delivering mail and packages to the right address are an excellent example of the issues involved with handling primarily unstructured data. Just obtaining the correct city and state information is difficult when one considers customer spelling errors, bad handwriting, or printing. The US Postal Service began using codes to enable accurate sorting more quickly when they introduced the 5-digit Zip code to be used by their customers as the last entry in an address. The rules regarding what each digit in the Zip code represented were somewhat fuzzy at times, but, in general, the first three digits represented one of the numerous sorting centers distributed across the country. The last two digits represented one of the local post offices served by that center. This code added some structure to the address data and later could be read with acceptable accuracy by automated sorting equipment scanning POSTNET (Postal numerical encoding technique) barcodes representing the Zip code values. Unlike the lines and spaces with varying widths used in the UPC barcode, POSTNET uses combinations of five bars of two different heights to represent each number as shown in Table 4.1, variation similar to the dot–dash patterns used in Morse code.

Table 4.1. Bar coding for each numeral in a POSTNET barcode

An additional tall bar is used at each end to mark the start and ending of a POSTNET code.

The POSTNET format has had four versions over time to expand its coverage from a six-digit format for the original Zip code to its most recent version of 12 digits shown in Figure 4.3c representing the Zip+4 code plus the two last digits of the street address to allow further sorting of the mail at the receiving post office in order of delivery for the mail carrier. Version 4 is often referred as the Delivery Point Bar Code, or DPBC.

Implementation of the POSTNET code was not easy. While large-scale business users could be encouraged to use it for their mailing purposes by offering a discount if they did, individual users did not have the capability for printing POSTNET codes on their correspondence. Some word processing programs began offering the capability to their users, but a substantial amount of mail volume still did not have even a Zip code with the address. As a result, the post office began developing systems to print POSTNET codes on the submitted mail that did not have them and, as a result, was one of the first large-scale users of optical character recognition (OCR) technology in 1982, which we discuss in more detail later in this chapter. They used OCR to recognize typed or printed Zip codes in an address and then print the associated POSTNET code on the envelope for easier sorting as the mail moved through the distribution system. Handwritten address information still required manual reading by an operator to input the desired code to be printed on the envelope.

As the population has grown larger, the need for more information in the address code has resulted in the development of the USPS Intelligent Mail Barcode. This 31-digit barcode now uses four types of bars of different heights and ascending and descending portions instead of two different heights, as shown in Figure 4.3c, to provide more precise mailing information. Eventually, its use will be required to mail information at automated handling discount prices.

In the past decade, a new optically scanned code format developed in Japan in 1999 has come into use. Called QR for quick response and shown in Figure 4.3d, this code is two-dimensional with large targeting squares in three corners and a smaller square in the fourth corner to allow the optical reader, typically a smartphone or tablet computer camera, to orient the QR code for correct reading. The initial international standard for this new code, ISO/IEC 18004, was approved in June 2000, and later updated in 2006. I first encountered its use during a visit to Japan in 2005 where QR codes were displayed on billboards and other advertising on posters, train cars, and newspapers. Some even had simple pictures or graphics embedded in the code pattern to suggest the general nature of the information—a Mickey Mouse outline for a local Disney attraction, a bowl of noodles for a restaurant, and an octopus silhouette for a marine aquarium. The Japanese residents interested in the information just held up their smartphone to take a picture of the QR code; the smartphone then decoded the image and connected its owner to a website displaying further information or dialed a phone number for tickets or a reservation.

One QR code can contain up to 1,817 Kanji characters, 4,296 alphanumeric characters, or 7,089 numeric characters. To give you a better impression about what that capacity is, it is roughly equivalent to two pages of text from this book. In addition to the earlier advertising examples I observed in Japan, this capability could be used, for example, to display a menu of instructions for different work procedures on a single sheet of paper that could be quickly and accurately accessed by a machine operator and directly used to program a piece of equipment to machine a new part. It also means that a lawn tractor manufacturer could use just a few QR codes to print the entire user manual inside a cover panel or under the seat for the tractor, including servicing and trouble-shooting data, for easy access by the owner’s smartphone or tablet computer. If additional help is needed, the QR codes could also provide the necessary information for accessing the manufacturer’s website, for calling technical support, or to order a replacement part. Appropriate QR codes could also be printed on the back or front panels of home entertainment equipment, household appliances, and so forth, for the same purpose.

RFID

The idea of using radio waves to identify objects has been around much longer than optical barcode concepts. It can be said that the first radio frequency identification occurred when radar was invented. A radio pulse was sent out and was reflected back when an aircraft intercepted it. To be fair, the identification information was only that something was out there. When radio transmitters were added to friendly aircraft so that the transmitters could send back a coded reply indicating that they were a friend when interrogated by a radar pulse, the rudimentary concept of RFID technology was born.

The application of radio waves for more definitive identification was considered by a number of agencies during the 1950s and 1960s. One development was the idea of a single-bit device installed in a package or attached to a unit of merchandise. The device is turned on by the nature of its fabrication process for use in detecting someone leaving a building with the package or item without paying. After payment, it is deactivated, destroying it for further use. This initial passive design is still in use today for verifying tickets to events and for preventing theft of items from an exhibit. These devices are essentially a tiny radio antenna that is tuned to the frequency of the interrogating transceiver. Connecting the two halves (poles) of the antenna is a tiny capacitive diode. When a radio pulse of the proper frequency at a close range strikes the RFID device, it reflects that pulse back, sometime modifying that frequency slightly depending on the design of its antenna. It can be deactivated by subjecting it to a strong electromagnetic field like that used to demagnetize old cassette tapes, causing the tiny capacitive diode to overheat and burn out, much like an overloaded electrical fuse.

There are two basic types of RFID tags: passive and active. Each type has a range of similar capabilities for storing and transmitting information depending on the complexity the microchip used. Passive RFID devices have no internal energy source, such as a battery, making them much less expensive but not as cheap as a barcode label, which is also a passive device. Hence, their operation depends on the energy transmitted by the interrogation pulse from the radio transceiver attempting to read them. When these devices receive an interrogation pulse from the transceiver attempting to read them, the minute amount of energy in the pulse is used to power the small microchip in the device long enough for it to respond with the necessary information. This response is typically accomplished by the microchip acting as a variable capacitor to modulate the reflected part of the interrogation pulse in simpler passive devices. More sophisticated versions can send more information using a set of interrogation pulses that not only supply enough energy for the device to operate, but also instruct the device as to what information is wanted. In batch reading applications, the instruction set usually includes one telling the device not to reply to any further interrogation after the reader has obtained the information it wants for the moment.

Active RFID devices have their own energy source in the form of a battery or solar cell. This allows them to broadcast their information independently at programmed intervals, acting like a location beacon,¹² or enables them to perform more complex information storage and processing applications. These capabilities come at a higher per tag cost and, as a result, these tags are not typically used for routine retail or inventory tracking applications unless the items of interest are expensive or require substantial amounts of information or both for their management or maintenance.

At the current state of the art, simple passive RFID tags are still more expensive on a per tag cost basis than their barcode counterparts. From a strict IT perspective, this added cost has precluded their use in many retail and inventory management applications. But, when viewed from a more integrated information and process management viewpoint, the overall cost to a business is likely to be less because of the added process capabilities and efficiencies provided by RFID technology.

Let’s compare some of these capabilities against the disadvantages of barcode applications. First, the scanning requirements are much more flexible. A line-of-sight view between an RFID tag and a radio transceiver used to read it is no longer necessary and orientation is not important. In fact, the tag can be out of sight inside a package container or beneath a pile of other items. While the communication between RFID tags and transceivers operating on UHF bands can be obscured by metals and liquids, this is largely not a problem for RFID systems operating at higher and lower radio frequencies. An RFID tag does not have to be in as close proximity to its reader compared to a barcode; typical scanning distances can range from one to two meters for smaller passive devices to more than several hundred meters for active devices.

If the RFID tag is an active device, it can even initiate the communication with readers based on some internally programmed schedule, something not possible with optically read barcodes. Some places where this capability is useful is tracking transportation equipment or notifying the maintenance department in a factory when a piece of equipment requires routine maintenance or calibration.¹³ The tracking readers could also reply to tag the current location of the reader so that the tag itself could contain a history of where it has been.

A unique capability provided by not needing line-of-sight access to an item is finding specific items much more quickly. Imagine that your inventory shows that your business has an item in its warehouse that is needed quickly by a customer, but your warehouse supervisor cannot remember where it is physically located since it was purchased some time ago. A barcoded inventory would require an item-by-item search to find it. By setting an RFID reader to the desired part number in an RFID tagged inventory, the supervisor only has to scan each section of the warehouse for a minute or so to quickly find it. In fact, some large RFID tagged warehouses have one of these readers installed for each section so the entire warehouse can be searched or inventoried in just a few seconds.¹⁴ Using a similar approach, items that have been incorrectly shelved or stored can be quickly identified and put in their proper locations. This capability is especially useful for proper maintenance of library book collections and other material record facilities such as medical record archives.

These advantages compared with barcode applications allow for much faster scanning of a collection of items in a grocery cart (note that the cart does not now have to be unloaded at the checkout station), on a pallet being moved by a forklift, or shelves in a warehouse. While it will take a further decrease in passive RFID tag costs to encourage supermarkets to convert to using RFID tags, the potential advantages in reducing checkout space and checkout personnel as shown in an IBM television commercial are getting us closer to that point.¹⁵ By adding the ability to sort out expired items in a store’s inventory quickly because the RFID tag could also contain date and other desired information, additional operational cost savings are possible. This last application may even have value for consumers in the future where a refrigerator capable of reading RFID tags could alert them of out-of-date food items in their fridge, or that some items such as eggs and cold beer have run low and need restocking.

Now at this point, some of you are likely thinking “Whoa! I understand how barcode scanning can read a bunch of items at a checkout stand since each one is passed in turn across the scanner. But how can an RFID scanner read a bunch of items seemingly all at once and without having to pass each one in turn in front of the transceiver?”

This is where the added cost of an RFID tag comes in. The small microchip connected to the tiny antenna in the tag contains not only the attached item’s identification information (in fact, that information could be exactly the same number as displayed on the UPC barcode label for that item) but also the ability to make some simple decisions. There are several schemes for dealing with “tag collision,” the situation where a number of tags are addressed at the same time by a reader. One scheme takes advantage of the natural situation that no group consists of tags that are exactly alike, that is, one of them is bound to reply a tiny fraction of a second earlier than the others’ response. The reader accepts that data and then sends a pulse that tells that first responder to shut down and not reply to further inquiries. This process is repeated for the remaining unread tags until all have been accounted for, a sequence of readings that typically takes less than a second to complete. Another scheme is for the reader to send out a pulse that tells all of the tags to generate a random number out of a large set of possibilities for their individual temporary ID. Given a large enough range of random numbers, it is highly unlikely there will be two random numbers in the group that will be the same. Then the reader starts down the list of possible ID numbers to read each tag in turn and tell it to shut down. The complete process is likely to finish in less than a second because each communication takes only a few microseconds.

The preceding discussion already points out that RFID technology has more data capability; it is only a matter of how complex the microchip is. For expensive items such as large flat-screen televisions, automobiles, home appliances, high-end digital cameras, and yard equipment, it could be useful to have the item’s RFID tag accompany it after purchase. At point-of-sale, the store’s reader could record the purchase date, store location, and even the customer’s ID on the tag for ready warranty information later, should it be needed. If in-warranty repairs are required, the serviceperson could use a reader to record the call and reason for it and also program the tag with that history.

RFID applications have one disadvantage that is worth discussing. Because the tags can be read at greater distances, especially the active versions, and without the need for line-of-sight access, the ability of an unauthorized user to access their information is increased. This disadvantage has led to concerns about people reading your personal information stored on an RFID tag in your credit card or passport, a thief driving by your warehouse or a customer’s residence and scanning it for desirable items to steal. While these concerns have created an aftermarket of security protection items for consumers such as metal-lined holders for passports and credit cards, there are a number of more effective methods to deal with this concern. These methods are typically designed for active RFID devices because they are more likely to be used on expensive items and to have a greater communication distance. The interchange of data can be encrypted for warehouse inventory and tracking applications and the reading distance can be reduced at the item’s point-of-sale by instructing the active device on the item to switch to a low-power transmission mode.

Like other aspects of information technology, the development pace in RFID applications is also growing. For a short introduction to RFID technology and an example of the enterprise level software solutions being developed to support its use, see the online reference by Holloway.¹⁶ For SMB supply chain managers and operations professionals, the book by Zelbst and Sower has a more recent overview of RFID technology with a number of business application examples and implementation suggestions.¹⁷

OCR

Optical character recognition technology has come a long way from the capability of its early days. It deserves attention here because of its value in allowing businesses to dematerialize not only some current information in paper form more easily, but also to convert some of their historical documents into a form more easily processed and shared.

Those of us who attempted to use OCR technology a couple of decades ago may recall that it often took longer to proofread the result and make the necessary corrections than it would take a trained typist to just transcribe the material directly. Processing tables of numbers was particularly time consuming, especially when the accuracy of the data was critical. As a result, the small amount of dematerialization of paper-based information was limited to making electronic copies of important documents.

In the intervening years, many scanning solutions incorporated into commonly available office printing equipment combined with document feeders has made it easier and faster to scan large numbers of documents. Many businesses use this capability to at least reduce their older records to a form more easily stored and retrieved. However, these records cannot be edited, searched for text content, or their numerical content accessed for analysis unless the information shown in the document images is converted into digital text and numbers.

OCR capability has become faster, more intelligent, and, most important, more accurate. Enough so, that when combined with fewer transcribers that have the requisite typing speed and accuracy, it deserves another look by businesses previously discouraged from using it as a viable option for inputting data that are not in a digital form.

As an example, I printed out this sentence as you see it here with the following list of numbers, letters, and symbols—3.14159Aa Σ ^® Δ λ 22C75S. The printed copy was scanned by an inexpensive (less than $149) OCR application (2010 version) for its conversion back into digital text. The scanner used was part of a common all-in-one office printer with a document feeder. The result without any editing is shown in Figure 4.4.

Note that the only problem the OCR program had was converting the Greek alphabet, a not uncommon problem when converting mathematical equations. More advanced OCR programs can be set up to handle more than one language set at a time, but the inexpensive one I used can handle only one language at a time out of a choice of number of languages. Language choice is important since part of the OCR accuracy is based on appropriate spelling to help, for example, recognize the difference between a lower-case I or a lower case L.

Figure 4.4. Unedited OCR printed text conversion result. When comparing this with the original sentence, the conversion process had the most difficulty with the Greek alphabet.

A consequence of using OCR for dematerialization of information is the need for concomitant policies for secure disposal of paper documents to be discarded and for storing those documents that currently still must be retained such as deeds, original contracts, birth and death certificates, other document forms required for proof of originality. We will discuss such policies and some information disposal methods later in chapter 6. Eventually, of course, some universally recognized proof of originality for digital documents will be developed and accepted. There are some digitally signed approaches used now by some organizations, but until there is some reasonable assurance that they won’t be tampered with by the many hackers out there keeping a paper document safely stored in vault somewhere significantly reduces access for such mischief.

GPS

GPS data collected in real time is usually in electronic form with location coordinates expressed in dimensions of longitude and latitude on the earth’s surface. Occasionally, one will see the coordinate information for a store or restaurant printed on a business card or advertising brochure to allow a customer to enter that destination in their personal GPS navigation device. Because this entry method is awkward and prone to error, more efficient methods for users to enter these data have been developed. Most navigation devices will accept more familiar postal address information because their internal database is capable of matching that data to the more precise latitude and longitude values for that address. In other cases, the former printed location coordinates have been replaced by a QR code that can be read by the user’s smartphone to input the data.

GPS systems that are more sophisticated can also determine altitude above sea level. Knowing altitude can be important in collecting business data if your product, service, or information values are affected by atmospheric pressure or the typically drier and cooler conditions at higher elevations. Accurate location data are important for transportation industries for navigation and for selecting routes that get their shipments or passengers to their destinations quickly at a minimum expense. Other useful business operations data are the current locations of capital assets, for example, trucks, forklifts, portable emergency generators, and rented equipment. An example of the use of GPS data for locating and controlling equipment and managing business operations is given below.