Pioneers in check imaging, a technology that many bankers expected would revolutionize one of their core operating burdens, have had to rethink its economics.
When a few of the nation's largest banks first bought the systems that convert paper checks into more manageable computerized images, they anticipated dramatic cost reductions-particularly in the labor-intensive proof-of-deposit, or POD, area.
But in the half-decade since those pioneering installations, the savings in POD have been lower than expected. As a result, they are seeking out fee-generating opportunities that make use of stored check images.
"People started down this dark hallway called image POD, and they're only now finding that the payback just isn't there," said J.D. Carreker, president of Carreker Antinori Inc., a Dallas-based consulting company.
"The new, enlightened approach is to capture the image with the idea of using it for a whole string of things."
Banks' experience with check imaging sends a cautionary message about any use of experimental technologies. The economics of untested systems rarely play out exactly as expected, observers said, and early adopters should be as conservative as possible in predicting financial benefits.
Image POD got a warm welcome when vendors unveiled it around 1990. Check processing had seen few if any major breakthroughs in the 1970s and 1980s, and commercial banks in cost-cutting modes were open to any tools that might help them handle the ever-rising tide of checks more efficiently.
Image POD, championed primarily by Unisys Corp., appeared a "killer application" for check processing.
Though the systems cost millions of dollars, they promised to cut overall processing costs by 20% to 30%.
Such savings would come largely through staff reductions. Using character recognition software, POD systems automatically read numbers from check images, obviating the need for human operators to do the same.
Some POD improvements have played out favorably. For example, executives at Barnett Banks Inc., Comerica Inc., and Huntington Bancshares (which along with Signet Banking Corp. were the first to install Unisys' POD application) said they have met their expected "read rates" of around 50%, meaning data are automatically read from about half the checks they process.
But bankers and other observers said character recognition technology has not reduced labor costs as expected, because banks often overlooked new requirements introduced by imaging, such as the physical preparation of paper checks for scanning.
"I think we all underestimated the value of operations people in the proof department," said Doug Halvorsen, worldwide marketing manager for payment solutions at International Business Machines Corp. In traditional processing operations, "people sitting behind unit encoders would take off paper clips and staples and straighten the work," said Mr. Halvorsen. "This prep work ... still has to get done" in an imaging environment.
Mr. Halvorsen and several bankers said balancing functions also have proven less efficient than expected when imaging came along.
"POD continues to be a good application but probably not one you'd want to lead with," Mr. Halvorsen concluded.
The new wave involves trying to do more with images once they are captured. Banks want to take advantage of the electronic portability of images to streamline "day two" functions, such as research, adjustments, returns, and reject-repair.
Instead of looking at these as individual events "and trying to optimize each on its own, banks are now trying to find commonality and connections, and that's where they are going to find benefit," said H. Douglas Ewbanks, executive vice president and managing director at Carreker-Antinori.
Rick Sellers, president of Huntington Bancshares' technology unit, added, "The only way that you can justify check image in a large bank is to look at the whole package-including POD with character recognition, archive, statements, and the export."
Mr. Sellers said Columbus, Ohio-based Huntington, like most of the other imaging pioneers, expected its investment could be justified by POD improvements alone, but "experience has made us wiser."
He said Huntington recently participated in a study of the check- operations costs at image-based banks versus those with traditional processing systems.
"We were more expensive than the best paper-based banks-but we were close," said Mr. Sellers.
He expects more significant financial benefits will emerge in coming months as Huntington constructs an archive system that will let the bank create fee-generating services. The archive should be completed at the end of next year.
Comerica Inc. executives agree with Mr. Sellers' emphasis on functions beyond proof of deposit.
Detroit-based Comerica processes 2.5 million items a day and is weeks away from putting its check image archive into full production.
The archive has enough capacity to house 35 days' worth of on-us checks. "This is designed to be a microfilm replacement," said Don McKenzie, first vice president of information services.
He said the potential financial benefits of the archive are split pretty evenly between cost savings and revenue generations. The bank immediately would become more efficient as it eliminates microfiche, which is both expensive and unwieldy to handle.
When images are stored in the archive, they can be easily shuttled from computer to computer. Bank customers could be charged fees for access to check images on-line, and statement printing could be customized more easily. Further, internal research functions can be done more quickly.
Mr. Halvorsen noted that many banks already are using images for fee- generating applications-particularly in wholesale areas. But many are rescanning checks several times for each application, which is wasteful. Image archives will right this situation.
"The revolution in imaging is happening in the archive," said Mr. Halvorsen. "That's the area to watch."