“Web Services” are like a singles bar for computer software: Systems cruise the Internet to find other systems that offer specialized functions, autonomously connect and consummate the transaction, and go their separate ways.
Touted as the future of computing, the concept’s very vagueness sets off a hype alarm. This rings louder when you learn that a major proponent has been Microsoft, which has the obvious agenda of making services most easily available to users of its Windows software. Is this a technology of real value, or just marketing fluff?
After several years, Web Services seem to be finding broad acceptance. Most applications don’t involve scanning the Internet for unknown partners: Like people, smart computers don’t talk to strangers. But the same technology can interconnect a company’s own systems or interact with predefined external resources. Web Services simplify such connections by using common standards including XML, SOAP and HTTP. This ease of integration, and not the ability to find new resources, makes Web Services attractive.
Among direct marketers, a popular Web Services application has been address validation. This is the task of ensuring that addresses are valid and formatted correctly to improve deliverability and earn postal discounts.
Address validation systems have been around for years, and the basic process is well understood: The system breaks an address into its components, corrects any formatting errors, then compares the result with a list of valid addresses provided by postal authorities. Addresses that match are accepted and given postal codes; those that do not match are flagged as possibly invalid. Matching records also may be enhanced with codes for other address-related information such as latitude and longitude, tax jurisdiction, census tracts and telephone area codes.
Address validation requires sophisticated software to account for formatting variations, data entry errors, missing elements and other deviations. Rather than build their own systems, most firms buy a software package or send data to an external service bureau.
Each approach has disadvantages. Service bureau processing requires data extracts and imports while in-house processing requires software installation and maintenance. In-house systems also must be updated regularly with new reference tables from postal authorities.
Web Services avoid the disadvantages of both approaches. Records can be validated in real time rather than batch, and no software needs to be purchased, installed or updated. Of course, it’s possible to set up remote address validation without Web Services, but the integration effort is considerably greater. So applications have been limited to ones where the higher investment can be justified. Web Services make it easier for low-volume users to access address validation remotely.
Several vendors currently offer address validation as a Web Service. They include traditional address software vendors Group 1 (www.g1.com), Firstlogic (www.firstlogic.com), Melissa Data (www.melissadata.com) and Intelligent Search Technology (www.correctaddress.com) as well as Web Services specialists such as ServiceObjects (www.serviceobjects.com) and Better Business Builders (www.bbbus.com).
All handle U.S. addresses, and most can process Canadian and international addresses as well. Along with address standardization and verification, some vendors return parsed versions of the input (broken into separate fields such as first name vs. last name). Several can validate telephone numbers, update telephone area codes, identify address changes reported to the U.S. Postal Service and append various other codes that can be related to a physical address.
Pricing typically involves a fixed fee for a specified transaction volume; higher volumes yield lower prices per unit. But pricing strategies vary by vendor. At Group 1 and Firstlogic, it is cheaper to process low numbers of records on Web Services than to purchase an in-house software license. Both vendors report several customers with low volumes of international data take advantage of this by using Web Services for international processing while still doing U.S. addresses on an in-house system.
In contrast, low volumes of Web Services processing at Melissa Data actually cost more than installed software, while Web Services are slightly cheaper than installed software at higher levels. Melissa Data’s prices in general are much lower than its larger competitors. Given these and other differences, it takes shopping to find the best fit for a particular business situation.
Whatever the business arrangement, the process of using these products is similar and simple. All that’s required is to generate an XML string and send it to the vendor via an HTTP call; what comes back is another XML string with the requested data. That may not sound simple to the uninitiated, and I couldn’t do it myself. But, trust me, as these things go this is about as easy as it gets.
Any instructions, such as the list of data elements to return, are specified in the XML string as well. To help matters along, the vendors provide sample code for different programming languages such as Visual Basic, Java, COM and C. With sample code available, setup should take just a few hours.
Web Services address validation is typically integrated with real-time processes such as order entry. This means that availability and response time are critical to avoid interrupting business operations. To achieve this, most vendors deploy multiple load balancing servers that deliver response in about a half-second.
Though most applications involve sending a single record and receiving an immediate response, some systems can be set up to process records in small batches. Large batches are problematic because HTTP does not allow a continuous connection between systems. Group 1’s system, HotData, includes software that can automatically break a large batch into many small batches to get around the limitation.
Web Services for address validation have been available for several years from the different vendors. Several report data volumes in the millions of records daily, with most volume from a handful of large users. Group 1, Firstlogic and Melissa Data each have 50 to 100 customers for their offerings.
