American Printer's mission is to be the most reliable and authoritative source of information on integrating tomorrow's technology with today's management.
May 1, 2007 12:00 AM
If you are like 80 percent of today's mailers, then you have not yet invested in a matching system. If you are now looking into this technology, there is an even greater chance that you have yet to master the buzz words and discern which features are essential to your operation.
Inserting systems commonly are used to insert printed products (personalized or non-personalized) into envelopes. Window envelopes are used to avoid the need to match the personalized contents with the exact same address on a printed closed-face envelope.
There is an ever-growing need to market to the masses more effectively by using personalization. As a result, inserting equipment now is being used to insert multiple personalized documents into each envelope, and to insert personalized information into closed-face envelopes.
Of course, standard inserting equipment does not have the automation capability to read each individual piece for quality control or to print the corresponding address on a closed-face envelope. This is where third-party matching systems come into play.
The most basic type of a matching system, which many of us know, is what I call the “pick and pray” method. The operator will take a stack of personalized insert A and a stack of personalized insert B, load the equipment to insert the different pieces into each envelope, then check each envelope. The checking can occur anywhere from every five envelopes to every few hundred envelopes. If the inserter accidentally stuffs a double into the envelope, each subsequent insert will be out of order, and the operator won't know it until the inspection process catches it.
With the introduction of electronic cameras, several companies have introduced systems that can read each document being pulled from an inserter pocket and take action in the event the documents don't match properly.
Here's a primer on the main types of matching applications mailing companies use:
Ensuring two or more inserts are correctly matched before being inserted into a common envelope.
Reading information from the document to be inserted, then printing corresponding information on a closed-face envelope. This is often an address block, but the printed information can be as complex as a variable graphic or logo that corresponds to the inserted document.
In this scenario, invisible ink (already preprinted onto the document by an inkjet device) can be read in the same way as the “read and print” application above, but the ability to have invisible ink can be very attractive when visible barcodes or OCR marks might detract from the appearance of the letter or document.
Other applications can be achieved by mixing and matching the above applications. For example, one of our customers is using a “print-read-print” system that actually prints the personalized custom insert, inserts the document, then inkjets the the customer address onto the envelope.
Other than the obvious need to reduce the labor associated with manual matching methods, our customers have seen other economic benefits to these applications. The prices achieved for all of these applications are greater than the rates for simple insertion of generic material — or even the rates for personalized information being inserted into window envelopes. And while it is easy to envision the cost savings associated with eliminating the labor-intensive step of manually checking matched mailings, the savings compound with the use of read-and-print applications, which combine document matching and envelope printing. Additionally, when the bottleneck of labor and multiple processes are eliminated, the inserting equipment can run at full capacity.
Depending on which applications the system is used for, a printer's customers can benefit from lower-cost envelopes, higher response rates and log data that verifies the accuracy of a mailing project.
The best matching systems on the market can be broken down into four core elements: lighting, format size, decoding and connectivity.
Cameras capture raw barcodes, OCR marks, or data matrix on an insert, document or envelope. The image is used to “match” other documents or provide the look-up identifier to retrieve information from a database.
The newest cameras often have integrated lighting systems, which means the illumination needed for vision is included on the camera, itself. This can greatly simplify the setup and adjustment of the system. There is less time required for installing, connecting and tweaking multiple components. This is particularly important for companies that want modularity and the ability to move cameras and systems. The system does not have to stay permanently mounted on an inserting or finishing device.
If the application calls for reading ultra high-gloss materials — or reading through glassine window envelopes — it still might be necessary to add separate lighting. A good example would be “stadium lighting” that illuminates at a low angle to reduce reflection bouncing directly from the lights off of the glassine window. This type of application usually needs more fine tuning.
The most common application is reading some type of print from a coated or uncoated insert. Usually, this is done particularly well with self-illuminated systems.
The size of the camera can be critical to the viability of using a matching system for a particular application. Very simply put, the smaller the camera, the better — in terms of being able to fit into tight spaces. Compatibility with Pitney Bowes, Bowe Bell & Howell, Mailcrafter, etc., is broadened with a small footprint. Some cameras are small enough to read upside-down in the track of inserting equipment without making structural modifications. This, again, can benefit the modularity of the system — small cameras have more location options.
The decoding of the image (barcode, etc.) can take place at either the camera or the host processor. It is important to know the pros and cons of where the processing happens.
Camera-based processing is when the OCR, barcode or datamatrix is decoded into a digital string of alphanumerics at the camera, then sent as decoded information to the host. This type of system was popular when large amounts of data could not be sent rapidly enough to the host computer, and when the processing speed of computers was much slower. In other words, it was much easier for the cameras to do the decoding work and send only a small amount of decoded information to the host computer, rather than large streams of video data.
Now, advanced systems allow for an efficient feed of camera signals to the host computer, which decodes the images centrally.
Bringing the raw data back to the host computer eliminates complex electronics and circuitry from the camera, enabling miniaturization and reducing camera cost.
With sizable amounts of image data pushing their way to the host computer, high-speed communication pipelines are mandatory. Two types of data streams are possible: digital or analog. For matching systems, utilizing digital data reduces the possibility for signal contamination or loss, as 1s and 0s are less susceptible to interference loss. Analog signals are more prone to error. An even more relevant factor in consideration with analog cameras is the need to convert the analog signal to digital values. Inevitably, this is carried out in A/D converter boards within the PC.
The image data captured on the document is most commonly one of the following types:
OCR is easy for the human eye to interpret, but it is one of the most error-prone methods for computer image decoding. OCR can be conducted on traditional “typeface” fonts that are used in documents, or special OCR fonts can be used. An example of those special OCR fonts (such as OCRa and OCRb) are the numbers at the bottom of bank checks. These are designed to reduce the very common decoding errors prone to typeface fonts. While these fonts increase readability and decrease errors, they also are more visually obtrusive.
OCR fonts with traditional typefaces remain popular for some applications simply because they are less noticeable than a 2D optical symbol, such as a barcode.
From an automation standpoint, however, matching applications using OCR have a higher misread rate. A simple toner speck on the paper can cause an error in the decoding — and a resulting misread.
This is where 2D barcodes have some advantages. These barcodes have been gaining popularity for a number of reasons. First, the newer codes such as data matrix and PDF417 have data error correction, which will ensure good reads (and very few erroneous misreads) even when there is damage to some part of the barcode image. Up to 60 percent of a data matrix can be damaged, and it will be readable. Second, the 2D barcode is much more compact than a linear barcode or OCR mark. As a result, less space needs to be used to embed personalized information; conversely, more data can be packed into a 2D data matrix. This is attractive for personalized mailings where a very small data matrix can be used.
OCR marks typically result in many more misreads than 2D barcodes. The reason is that host-based decoding systems can decode barcodes within several milliseconds, on average — and this gives plenty of time to handle even the fastest inserters. These read times are fairly consistent for 2D barcodes but can be much higher for OCR marks. This is because each digit of an OCR symbol will be read independently and more characters will lengthen the processing time proportionally. The host computer will compare characters to a library of standard and learned characters, and if the characters are clear and clean, a match can happen within several milliseconds. Because there is no data redundancy or error correction built into simple OCR marks, however, a damaged mark might result in a misread. The matching system can divert the piece to a tray or stop the inserter.
Utilizing the availability of 2D barcodes along with invisible ink can allow the matching data to be placed on the documents without any detection by the reader. With a UV/IR-compatible matching camera, this data can be decoded and used with a matching system.
With the introduction of the 4-state barcode, mailers now have the capability to embed additional information into a barcode that replaces the PostNet barcode. This data can be the record identifier or sequence number in a matching job and replace the need for a separate 2D barcode, for some applications.
While the cameras capture the image and identity of each piece, the tracking system must know the precise travel and position of each piece at speeds up to 20,000 pieces per hour. In addition, when pieces are not inserted correctly — or mismatched — the tracking system needs to control that misread piece through a divert station, alert the operator to the location of the mismatch, and optionally shut down the system. This type of tracking is necessary to trigger the inkjet printer at the appropriate time.
The electronic tracking of the pieces in an inserter system can utilize one or more tracking sensors to maintain exact positioning of each mailpiece. Each production segment — the inserter, envelope transport, turnover, divert station and inkjet transport — will have its own motor and operate at potentially a different speed.
Depending on the manufacturer, a matching system will use optical sensors, speed encoders and/or hook into the circuitry of the production segment to accurately track the pieces.
A natural use of the electronic tracking system is to extend the tracking process to the end of the production cycle (after the inkjet printer) and record the exit of a mail piece onto the sorting tray or conveyer. This allows for a log to be issued, verifying which pieces completed the process. Both marketers and production staff can benefit from this — internal and external customers want to have quality assurance for “100 percent mailings.”
Matching systems vary widely with respect to what data they use as well as what logging or audit information can be provided back to the operator.
In a simple matching application without print, it is possible to perform a match without the use of a data manager or database system. In this case, just a sequence number is read from each piece and, if they match, the system gives them the OK to proceed and it is considered a successful match and insert. If they don't match, of course, they are diverted or removed.
In a more sophisticated system, a database management system performs a “lookup” on each sequence number and provides the operator the full information of each record that is passing through the system. Rather than just seven digits or a barcode, some systems can show the progress of “John Smith” and his address and record information. This is more user friendly for the operator ensuring the job recovers from mismatches.
The database management system is mandatory for a read-and-print system, because the address and other information needs to be printed on the outside of a closed-face envelope.
A database management system also is needed for effective job audit logs. A log that records the time and status of each mail piece is much less valuable when it only contains a sequence number as a record ID. When used in conjunction with a database management system, the log can deliver a file with full record information that is valuable to the mailer and its customer. In its fullest extent, this logging system can provide a full file of the customer's data along with a record of each event through a four-step production system, ending with delivery of each piece onto a conveyer.
Matching systems offer an opportunity to add revenue and profits, as well as throughput and capacity, to a mailing operation. By spending some time understanding the principles of matching systems, which features are relevant to your needs, and how to decode the buzz words, you will have good start at evaluating whether this technology can benefit you. When evaluating camera or matching systems, it's best to start by developing a good understanding of your customer's needs and your own applications' requirements. Learn which factors are important to you — barcode type and placement, data management, type of envelope, etc. With this understanding, you can discern which features are important and start to compare vendors.
Glenn Toole is vice president of sales and marketing for MCS, Inc. (Gaithersburg, MD). Contact him at firstname.lastname@example.org.
AMERICAN PRINTER asked Glenn a few key questions on mailing equipment and matching systems for small to midsize printing companies.
What is the average cost of a matching system?
“The cost of complete systems — starting with vision systems that simply verify the correct matching of inserts — cost, at the low end, between $20,000 and $30,000. It can be lower, but the level of automation, functionality and support starts to drop. Integrating inkjet systems on top of that (for applications such as reading a code on an insert, doing a database lookup and printing the associated address on a closed envelope) adds another $30,000 to $60,000.”
Can you describe the differences between low-end and high-end systems?
“Most of the differences are between the different vendors, in terms of whether they have certain features (such as job logging, invisibility capabilities, camera capabilities, etc.). Some companies seem to do a piecemeal integration, whereas we manufacture the cameras, write our own software and manufacture the inkjet system.”
Why would a print business running a smaller mailing operation need this technology?
“As printers are moving into mail services, they are purchasing the latest generation of high-speed inserters such as Pitney Bowes Flowmasters, Bell & Howell CMC, etc. The high-speed units justify the automation provided by automated matching systems. The older inserters (which are very prolific in established direct mail companies) don't force the automation issue as much, because manual labor can more easily keep up with the slower speeds. Once you invest in the new equipment, you really need some automated matching systems or you will end up running your high-speed inserter at well below the rated throughput.”
What other applications are there for matching systems?
“Plastic card printing, such as tip-on cards for applications like sticking a membership card on top of a letter, and printing the personalization of the membership card. To achieve this, the letter is fed through a card attacher, a camera reads the database serial number of the letter, and then it does a lookup and prints the name, membership number, etc. on top of the plastic card after it is attached to the letter. I think printers are getting more into printing cards.
“I have seen a number of applications that are linked with the growing personalization provided by digital presses such as HP Indigo, iGen3, Xeikon, etc. Printing companies that personalize these high-value documents now have a need to make sure they go into a secure envelope with a personalized address. The way to do this is to read a small number or code on the brochure and track/print the address after it has gone into a closed envelope. It is the need for closed envelopes with targeted internal information that has driven the need for automated matching and read-and-print systems. In the old days, you would have 300,000 identical letters or mail pieces go into closed envelopes, where you could print the addresses and not worry about matching to the internal contents.
“So, two factors are at work: Printers are moving into mail, and increasing personalization is driving the need for automation. Both of these factors are pushing at mailers — even the smaller ones. Also, the margins are healthy for mail services and the price of equipment is low, so the payback often is very good, even for a smaller printer.”