What’s Next For Image Exchange By Peter Lucas More than a year after Check 21 became law, volume is running at a trickle through check-image exchanges—and mostly ending up again as paper. But that could change soon as the economics of paper processing finally force banks to join networks, network interoperability improves, and the networks themselves launch services that draw on the wealth of data they control. When check-image exchange networks Viewpointe Archive Services LLC and Endpoint Exchange announced late last year that they had linked their systems, the news attracted little notice in banking and merchant circles. The question is, why? Here were two primary image exchanges fulfilling a large part of the promise of the Check Clearing for the 21st Century (Check 21) Act, at last opening a channel by which the banks belonging to one network could trade electronic check files with the banks in the other. As it turns out, since Check 21 became effective in October 2004, actual cost savings from check electronification has not lived up to the hype. Check 21 was supposed to drop check-processing and handling costs from about $1.50 per check to pennies, as merchants and banks exchanged electronic images of paper checks through the networks, rather than shipping paper drafts via airplanes and trucks. Backers saw image exchange generating savings of hundreds of millions of dollars annually for the banking industry. But a number of stumbling blocks have constricted the end-to-end flow of images running through the networks. Lack of interoperability between Viewpointe, Endpoint, and SVPCO, the three major privatesector image-exchange networks, has been one. The reliance on a large volume of image-replacement documents (IRDs), or substitute checks, which are printed out locally from images and shipped to paying banks when for technical reasons they cannot receive an image electronically, is another. While Check 21 enabled IRDs by conferring legal status on substitute checks, it said nothing about end-to-end image exchange. For paying banks, printing and transporting an IRD costs about the same as receiving and processing a paper check, though for the banking system as a whole the process replaces transport costs from the banks of first deposit (no more planes and trucks cross country) and speeds up clearing times. Other stumbling blocks include large variances in the prices charged by the networks, questions about network security, loose data-formatting standards, and difficulty in weeding out duplicate debits (box, page 30) to a consumer demand-deposit account. The result is that, while some services made possible by image exchange—such as the capture of check images at remote locations like retail outlets—have begun to catch on, banks have not rushed to join image-exchange networks. Instead, they’re taking a wait-and-see attitude. “Check-image exchange is a fragmented market, full interoperability has yet to become a reality, and a lot of confusion remains over such issues as pricing and security as the networks are still formulating their marketing message,” says Greg Schratwieser, president and chief executive of Arlington, Va.-based ICI Consulting Services, which tracks image exchange. “Until banks can be shown actual benefits of image exchange in terms of cost savings and know they can exchange images between networks, adoption is going to be slow.” But banks may not be able to take a wait-and-see approach much longer, as the inexorable economics of check processing catch up with them. And the networks are hard at work, not only forging crucial links to each other but also building new services that will draw upon one of their unique strengths: access to a welter of DDA data covering millions of accounts. Already, networks like SVPCO, which was built by some of the nation’s largest banks, are starting to see a volume buildup (chart, page 32). The fact remains, though, that with about 3% of the more than 16 billion transit checks written annually in the United States being processed electronically via image-exchange networks—and mostly ending up as IRDs—the networks are nowhere near hitting the mother lode of volume. Banks using the networks are sending for the most checks of $8,000 or more to gain faster clearing times on high-value items, but it is checks written for $200 or less that make up the majority of check volume in the United States While network interoperability is expected to be a major step in opening the spigot on image-exchange volume, it remains to be seen whether the agreement between Viewpointe and Endpoint will accomplish that goal any time soon. Endpoint, which is also in discussions with SVPCO about a link, acknowledges that the two sides are still far apart when it comes to resolving technical issues. “From a network standpoint, the architecture used to build them does not necessarily work when it comes to connectivity,” says Danne Buchanan, chief executive at NetDeposit Inc., a Salt Lake City-based provider of check-imaging software. “It’s a little overly optimistic to think the networks are going to quickly open the flood gates on image exchange when there are thousands of banks to be connected to.” Tipping Point So far, volume in the networks is nothing to shout about. The Federal Reserve, to which all banks connect, remains well out in front of the private sector, exchanging about 20 million images a month. Endpoint exchanges about 8 million, SVPCO 8.7 million, and Viewpointe, whose five owners process about 70% of all check volume, currently exchanges about 3 million images per month. Again, most of this volume, though transmitted over great distances as images, winds up as paper again in the form of substitute checks, since most paying banks have not yet geared up to process imaged payments. Most of the images exchanged over ViewPointe’s Pointe-to-Pointe network are traded among the owner banks, Bank of America, JP Morgan Chase, Fifth Third, SunTrust, and Wells Fargo, and the network’s 11 other customers. “There is not enough volume yet to get meaningful exchanges between banks,” says Paul Citarella, an executive vice president for Alogent Corp., an Atlanta-based provider of check-imaging software. Still, end-to-end image exchange is a complex process, perhaps even more so than the more enthusiastic backers of full-scale electronification imagined a year ago. That has tended to rein in expectations as well as volume. The networks point out that key members and owners are still ramping up their imaging systems, which in turn limits their volume. And some industry experts say the networks are proceeding cautiously. Expanding volume at a controlled pace, they say, can limit technical snafus that might turn off potential clients. The payoff in such a cautious approach could be a volume bonanza. Experts expect that once end-to-end image exchange goes mainstream, the networks can potentially be exchanging tens of millions of items, and in the case of ViewPointe and Endpoint, 100 million or more images each per month. A gradual build up to that plateau makes more sense than a fast take-off followed potentially by a daunting crash, says Citarella. Besides, end-to-end imaging is not expected to become a mainstream technology until 2008 at the earliest, so the networks have some room to be cautious, according to industry experts. “Image-exchange volume is starting to ramp up, but there are still several proprietary data formats being used and that requires time to create rule sets to limit and standardize the formats,” explains Sydney Hicks, president and chief executive of VECTORsgi, an Addison, Texas-based supplier of image-exchange software, including the routing system used by SVPCO’s network. The variations in data formats are due to the multiple dialects of X9.37, the ANSI data-formatting standard used in electronic image exchange. Although the data packets that make up an image sent over the networks follow the basic guidelines of X9.37, variations within the format can alter how the data map to the data fields that make up the actual image of the check on the receiving end. These inconsistencies, which are the result of a lack of data-formatting rules, make it necessary for banks receiving the image to scan the file for viruses and worms that can corrupt their database and even crash their servers, since they cannot be certain the variations in the data format were added by the sending bank or a hacker. To avoid the cost of translating the many dialects of X9.37, as well as the potential security risks associated with receiving such files, some banks favor receiving IRDs instead of investing in image exchange applications until these issues are resolved, according to industry experts. “The industry needs rule sets that limit the data formats used because this is a complex end-to-end exchange technology,” explains Hicks. “There are going to be valid reasons why some banks want proprietary data formats in some data fields, and we have a couple of bank customers that require different formats in some fields, so it makes more sense to move forward smartly while volume is ramping up.” What might finally force most banks to adopt end-to-end exchange is the economics of check processing. Paper check volume is dropping, pushing per-item costs higher as fixed-facility costs must be spread over fewer items. The problem only grows worse as the volume of checks converted to image creeps upward. As image volume grows, the per-item cost of processing paper checks is expected to rise so high that holdout banks will have no choice but to get on board. That tipping point should be reached once 10% to 15% of checks are electronically exchanged by the networks, according to Jerry Chambers, former interim chief executive at Viewpointe who now works as a consultant to the network. “Once that volume is taken out of the current system, network volume will move precipitously,” he predicts. Already, Check 21 has opened an enormous electronification market among corporate offices, retail stores, and other non-bank locations that accept checks (chart, page 34). Seeing the opportunity, vendors are starting to cater to niches within this so-called remote-deposit capture market. To serve locations that don’t accept huge volumes of checks, for example, MagTek Inc. and Silver Bullet Technology Inc. have mated Silver Bullet’s API with MagTek’s check scanner, which reads up to 70 checks per batch, to allow lowvolume locations to set up more quickly to begin shipping images. Electronic Railroads Chambers adds that ViewPointe’s owner banks, which include some the biggest in the country, are committed to achieving full image exchange among themselves by early 2007. He acknowledges, though, that that timetable can change as Wells Fargo and Chase are still ramping up. Technical caution has only served to reinforce the banking industry’s historical inertia regarding new payment technologies. So the snail’s pace at which many banks are moving to embrace full end-to-end image exchange should not be surprising, say some bankers. “For banks to become more proactive in deploying image-exchange technology, they need to see image exchange as the future of check processing,” says Mark Rollo a manager of check-image programming for National City Bank, which is aggressively deploying image-capture solutions among its merchants and in its branches. Once the cost of processing paper checks drives banks to image exchange, the question will be what network to use, not whether to deploy image-capture solutions. Differences have long since emerged in the business models employed by the networks. Endpoint operates a conventional switch, while SVPCO uses a routing system that sends electronic parcels of checks and associated data, or payloads, directly from bank to bank. But both systems in one way or another act as electronic railroads replacing trucks and planes. Viewpointe has chosen to position itself as a network/archive that stores the images for retrieval by both the collecting and paying bank. Paying banks must store their images for seven years in accordance with Check 21. By serving as an archive, as well as a transportation network, Viewpointe is attempting to differentiate itself by providing a value-added service for which it can charge more. Viewpointe built its business model on the belief that banks will prefer to have off-site archival facilities because it is more economical. The archives will be available to the collecting and paying bank. “Our aim is to move payment information over the network and have images available on demand,” says Chambers. “It’s similar to how the card networks move and settle transactions, only banks and merchants won’t have to wait days to receive a copy of the payment draft.” But many banks have already built their own image archives, as has the Fed. SVPCO’s 21 owner banks, for example, have built 16 such archives. The primary reason, according to Susan Long, a senior vice president for SVPCO, is control: Banks are loath to cede direct control over their images to a central authority. In May 2005, SVPCO was handling more than 125,000 check images with a total value of more than $1 billion per day. Daily dollar volume was projected to hit $3 billion by the end of 2005. The Best Deal The differing business models among the networks suggest that banks will likely pick and choose what volume to run through each network, as well as the Fed, and even the automated clearing house. “Banks are going to utilize their options when it comes to settling checks through a particular network or the ACH, based on the economics,” says Rollo. “In a free enterprise system, you look for the best deal.” For many banks, the best deal may not be based on price, but on network connectivity to banks. With more than 18,500 banks and credit unions operating in the country, connecting them all to an imageexchange network will be a monumental task. “The best deal may likely come down to endpoint connectivity to allow for initiation of a particular transaction,” says consultant Schratwieser. Depending on circumstances, the best connection hub may be the Fed, the ACH, smaller regional imageexchange networks, or direct image-exchange links between banks. The role of each varies. The Fed is considered the public sector safety net to ensure that banks without the resources to invest in full-blown image capture and exchange solutions will always be able to conduct electronic image exchange. The ACH, in which banks have a vested interested after helping build its infrastructure, is likely to be a low-cost alternative. During the third quarter of 2005 more than 432 million accounts receivable (ARC) e- check payments were cleared through the ACH, up 62% from the same period a year earlier. In 2004, banks routed more than 1.25 billion ARC payments payments through the ACH, a nearly six-fold increase from the previous year. “The potential is there for ARC to be moved into bank branches to perform remote capture,” says NetDeposit’s Buchanan. Direct connections are likely to prevail when banks are attempting to exchange locally. Austin, Texasbased Frost Bank, which has branches throughout the state, clears most of the checks it receives locally, according to VECTORsgi’s Hicks. “Exchanges aren’t the only way to move images,” she says. “There are going to be multiple solutions based on what the customer wants.” The image exchange networks recognize they cannot thrive offering plain vanilla solutions. ViewPointe has on several occasions publicly alluded to the fact that it was not created simply to electronically exchange and archive check images. Long-term, the network sees an opportunity to offer merchants real-time check-authorization services, a la the credit card networks, that would even include the ability to verify availability of funds in some instances. ViewPointe’s Chambers says that providing such a service is one of the network’s long-term goals. ViewPointe executives insist their network can provide richer data about the behavior of a specific checking account than third-party check-guarantee services because its members and owners have access to most of the demand deposit accounts in the country. In comparison, third-party check guarantee services build their authorization databases and scoring models based on volume running through their merchants. ViewPointe’s database will enable the network to provide more accurate real-time authorizations at the point-of-sale, further minimizing the risk of fraud. Merchants seeking a guaranteed payment can be charged a fee. Beyond Checks? Chambers predicts the service will have tremendous appeal for merchants and banks, which respectively lose $12 billion and more than $7 billion to check fraud. “What we intend to offer is nothing unique, but we feel we can execute the service better because of the higher quality of information we have access to through our owners and customers,” he adds. Speculation has also arisen that the image-exchange networks, whose huge pipelines are built to handle bulky data loads, may push into competing for ACH volume and other types of electronic transactions. SVPCO already handles ACH processing and real-time wire transfers and sends returned checks, IRD printing instructions, and account adjustments over its network. “We could even send documents if they are put in the right format,” says George Thomas, an executive vice president. “ACH and wire transfers are a part of our business now, but until we get check imaging down to our satisfaction, we won’t look to expand the use of the network.” If anything, branching off into new types of electronic payments is years, if not a decade, away from becoming a reality. The networks recognize that the immediate challenge is to improve connectivity to banks and between themselves. This is a critical first step to getting beyond the disappointing early volumes running through image exchange systems, says Jeff Vetterick, president of Oklahoma City-based Endpoint Exchange Network. “But our agreement with ViewPointe has brought the industry to a tipping point,” he insists. “The industry is no longer limited in how it can grow volume.” Will Image Exchange Have Some Banks Seeing Double? Duplicate debits to a demand-deposit account have always been a problem associated with check processing, but banks have been able to minimize their mistakes in the paper check world. Now they are facing a more daunting challenge as the industry prepares to make the leap into electronically exchanging check images. The problem: duplicates of image replacement documents (IRDs), the paper printouts made from check images and presented to paying banks for settlement. Until recently, the primary source of the problem was a lack of control by the collecting and paying banks and the Federal Reserve when it came to the presentment of images to be converted into IRDs. The intermediaries had no real controls in place to spot whether an IRD was routed through their network twice or if a duplicate went through an alternative network. At the receiving end, paying banks also suffered from lack of controls by not having an exception queue for questionable IRDs and images. The near-term fix has been relatively easy. The Fed now has tighter controls in place to track what images and IRDs pass through its network and receiving banks are setting up exception queues for questionable IRDs and images to be reviewed manually. No one knows exactly how many duplicate IRDs are now floating around, but the general estimate is that the number is not what it once was. Still, the onset of end-to-end image exchange could cause the problem to rear its head anew. “More controls have certainly been put in place, but the system is not 100% accurate, and as more banks get on board with image exchange many of the problems experienced by the early adopters are apt to be repeated,” says Rod Springhetti, senior director, payments business development for Carekker Corp., a Dallas-based vendor of Check 21 software. The two most common sources of duplicate IRDs are checks returned by the paying bank for insufficient funds and rebate checks using the same MICR (magnetic ink character recognition) lines. In the case of the former, merchants will typically resubmit a returned check on the chance the customer has since properly funded their account. Many times, customers write a check against anticipated float but get caught short when a payment clears ahead of a deposit. Resubmitting a check is simpler than having to go back to customers and wrangle the money out them or hire a collections agency to do it. In the case of the latter, rebate checks typically use the same MICR lines, which means they all have the same embedded sequencing and account data. Given these problems, most banks have concluded they will have to deal with some duplicate IRDs, according to Springhetti. “The key is to have controls in place across all channels in the clearing process to determine which duplicates are valid and which are not,” he says. Carreker, like many check-imaging software companies, has developed an application to enable collecting banks and networks to spot duplicate transactions in real time. In most cases, the transactions are pulled from the system and manually reviewed before they are processed. How good are the detection capabilities of these applications? The early read is that they are an improvement over the previous system, but still have a way to go. The chief criticism is that banks have yet to test them on huge volumes of IRDs and electronic images. “These applications are still new so not all the potential shortcomings are known,” explains Bruce Dragt, senior vice president, check strategy for First Data Corp., which operates the TeleCheck network. If nothing else, the emergence of such applications has helped reduce the error rate. “In 2005, the industry was still learning as it went along, but it is now better off than it was in 2004, and it will be better off in 2006 for what it learned in 2005,“ says Tom Kettell, vice president of marketing for RDM Corp., a Waterloo, Ontario -based supplier of check-imaging software. “The benefit of learning as you go is that safeguards, controls, and best practices are continually put in place.”