Lately, a big red warning flag has been waving high above the mailing industry. At a time when mail budgets are very tight, growing concerns of cost containment and improved efficiencies are heightened. Every year, more than 5.4 billion mailpieces are not delivered to the identified recipients for a variety of reasons. Never before has address quality been as critical for mailers of all sizes, in order to keep costs down and deliverability up. These challenges face all of those involved in the mailing industry, from the U.S. Postal Service and commercial couriers to presorters and large-volume mailers. In addition, postal agencies and large-volume mailers are trying to combat rising operating costs for equipment, fuel and labor. The postal and mailing industries face a continuing challenge to control costs and meet deadlines without sacrificing quality.
Last year, 200 billion mailpieces as well as packages flowed through the US mail stream. As information moves faster through the marketplace, the need for efficient yet flexible document processing becomes increasingly important. Automation is one way the mail industry has sought to control costs while improving turn around and efficiency. Today's mail operators must have automated equipment that can handle the vast expanse of mail being sent in diverse formats, quickly and efficiently.
Automated recognition technology has improved and overhauled delivery processes in the mailing industry. One of the major ways recognition improvements have made an impact is by reducing the reliance on manual keying. Automating the data-entry process speeds up deliverability, and more importantly, increases accuracy and lowers costs. One of the core methods of reducing manual keying is through automatic mail sorting, a process made possible through address-recognition technology. For years, address-recognition has been used to process machine-printed addresses while improving data accuracy and processing efficiencies. Over the last few years, address-recognition technology has made it possible to read hand-printed and cursive addresses while holding error rates to their current volumes. Today, this technology has improved to the point that error rates are dropping noticeably.
Additional improvements in the recognition of addresses have also been made recently. Multiple engines are now being used to process letters in the postal mail stream as well as sort mail for commercial pre-sorters. Today, with Optical Character Recognition, Intelligent Character Recognition and Natural Handwriting Recognition unconstrained cursive and handprint companies are able to process more information on each mailpiece. Using recognition processing has enabled the Postal Service to read more than 80% of handwritten letters that enter the mail stream; that's over 68 million mailpieces read daily using automated methods. This percentage is even higher for typed letters. When handwriting recognition is used in conjunction with machine-print recognition, software can identify an appropriate engine and apply it to the address field, resulting in more mail being read automatically.
Many advances in core recognition technology have led to increased address read rates. One such improvement is the introduction of word recognition technology in addition to character recognition technology. Today, advanced companies use technology that not only recognizes words character by character but also holistically. Holistic recognition treats the word as a single entity and attempts to recognize it from its overall shape as opposed to its individual characters. Holistic word recognition has improved both speed and accuracy.
The third technique employed is the use of in-context information. This information provides data about a particular field, increasing the confidence levels of recognition answers. In-context information includes aliases and dictionaries as well as global and customer-specific databases. ·
Aliases, another technique used to enhance recognition, help identify a particular address field written in a variety of formats, different from the traditional layout. Aliases represent all possible spellings of a name or address and can be applied to a variety of fields including department names, contact names, postal codes, etc. Consider a letter addressed to the marketing department of an organization. This department can be addressed as Marketing, Mktg., Mkt., Market, etc. Aliases help recognition software consider multiple spellings for this group to route the letter to the correct department.
Address verification does not stop there. Automation providers use the Postal Service database to compare address information against the most up-to-date addresses. Commercial pre-sorters also undergo CASS certification, a stringent qualification by the Postal Service, to assure address information is formatted according to the Postal Service standards. With cooperation between the CASS vendors and the Postal Service, software validates, corrects and enhances address recognition to improve mail delivery. Using intelligent search and matching tools, the correct address can be found in the database much more quickly and accurately than with standard address correction tactics. These tools are used to validate addresses and distinguish between residential and business addresses.
To improve the power of automated recognition even further, providers also use parallel processing and voting methods to increase read rates and accuracy. Parallel processing uses recognition software from multiple vendors to read address information. If the engines produce different results, a voting method is applied. Voting technology uses a third engine to verify which answer is correct. These processes allow for multiple verification and a system of checks and balances, essential steps in controlling error rate and maximizing data quality.
In addition to recognition improvements, other technology developments are being made that have expanded the types of mail capable of being automated. In the past, automated recognition had been applied primarily to letter-sorting applications. Recently, automated recognition has expanded into flats and parcel-sorting applications.
Growth into flats and parcels has been made possible with the use of new address search technology called address block locator (ABL). Sophisticated ABL technology can find an address no matter where it is located on a mailpiece. Automation systems of the past could only recognize an address if it was found in a constrained amount of space in a predefined location. Previous technology could recognize most addresses on a simple envelope, but could not recognize the addresses on parcels or flats. ABL has allowed software to identify an address located within a larger area and can capture address information from bigger mailpieces enabling, automation technology to be applied to these applications.
Another implementation of advanced automation technology is address parsing. This process breaks down an address into separate parts to detect the city, state, ZIP Code, address number and street name. Parsing is now used to recognize the many ways street addresses can be written. For example, many people write the address number after the street name or use various abbreviations or acronyms in an address. Once parsing has separated an address into distinct fields, an address database can be used to cross-validate the information and automatically correct mistakes. With this technology, automated systems can capture more than just addresses written in standard format. Today, advanced mail companies using parsing technology can read most mailpieces automatically.
Advances in mailing automation technology have increased accuracy and deliverability for the mailing industry while reducing the costs of processing mail. This has enabled the technology to be applied to new mailing applications including undeliverable as addressed (UAA) mail, business reply cards and airbills.
For example, with returned UAAs, recognition is now being used to automatically capture incorrect information from moves and entry errors. This data-capture service compliments current list cleanup services such as National Change of Address, FastForward and Delivery Point Validation to help reduce the number of undeliverable mailpieces going through the mail stream.
Using technology to automate the data-capture process will allow companies to lower costs, reduce reliance on manual data entry and improve accuracy and efficiency. These advances and others are changing the way mailers look at mail processing and are enabling savings, accuracy, convenience, adaptability and profitability to be realized across the board.
Steven Powers is director of World Wide Postal Business for Parascript, LLC, an industry leader in recognition software. For additional information, visit www.parascript.com, call 888-772-7478 or e-mail steven.powers@parascript.com.
