Payments Watch

Proof of Deposit (”POD”) is the verification of the dollar amount on a check or draft being deposited. This verification is done by comparing the handwritten amount on the check/cheque to the amount written on the accompanying deposit slip. Proof of deposit is an important step in check processing, yet has alluded full automation (unlike the sorting of checks via MICR technology).

There are commonly two ways of performing proof of deposit: conventional POD, where human operators verify the dollar amount of each item -typically when encoding the item’s amount in MICR on the item for later use by automated reading and sorting equipment - and image POD, where optical character recognition technology is used to verify the dollar amount of each item after the item has been scanned into an electronic image.

Unlike a conventional POD process which requires the handling of the physical item (e.g. check), an image POD process (theoretically) does not require the physical item once it has been scanned unless those items need to be physically sent to someone (e.g. a paying bank) later on in the process (in which case, the items can be MICR encoded in a completely automated fashion, sometimes referred to as power encoding). However, even though image POD as a check/deposit processing technology has been around for years, conventional POD has been actually been favoured by most. The reasons are multiple:

• Recognition technology cannot read all (handwritten) amounts. The percentage of items that can be read is called the recognition or reco rate. If the amount cannot be read, a human operator needs to verify the amount. This is usually done by an operator reviewing an image of the item, however, it detracts from the automation being sought. Historically, reco rates of 50% or even lower were common, although today, reco rates are generally much higher.
• The recognition technology may read the amount, but read it incorrectly. In this case, one thinks that the amount has been read, but in actual fact it has not due to the recognition technology substituting one or more characters/digits for another. These are called substitution errors and are particularly heinous in that they are often not picked up on until much later in the check clearing process.
• In the case where the amount cannot be read by the recognition technology, and the image is reviewed by a human operator, the operator may incorrectly enter one or more characters/digits for another. These are called keying substitution errors and are also very troublesome in that they are usually not picked up on until later in the check clearing process.
• The aforementioned substitution errors can make balancing large volumes of items particularly difficult. Deposit items are typically consolidated into large batches (or entries as they are often called) when posting to customer accounts and/or sending items out for collection and settlement). When substitution errors exist, this balancing process can become very difficult and as a result is typically drawn out. In a conventional POD process, human operators are balancing each and every deposit transaction up front in the process, usually resulting in a more efficient balancing process after items have been consolidated into larger batches for posting and/or settlement.
• Incentive pay schemes have lead to extremely high throughput rates by human operators performing the proof of deposit function while MICR encoding items for clearing.
• In many cases, the physical items need to be reintroduced to the process as original items are required for presentment when settling items drawn on another financial institution. While a more automated power encoding process can be employed, the image cannot replace the physical item “end to end” which detracts from the efficiency of the automation being sought via image technology.

In the recent past, however, the tide has turned for image POD due to a number of factors:

• Recognition rates have improved, sometimes dramatically, from the historical norm.
• While recognition rates have improved, the recognition technology today makes fewer substitution errors.
• Check 21 in the United States, and similar truncation schemes in other jurisdictions, has ensured that the original items do not need to be reintroduced later on the clearing process, allowing checks to be processed by the collecting agency as an image end to end.
• The adoption of image exchange, where the images themselves are used as presentment to a paying bank, has allowed the use of images end to end, not just at the collecting institution, but throughout the entire settlement process.
• The use of the internet and other inexpensive communications networks allows items that can not be recognized automatically to be keyed by remote operators. This labour is often characterized by a lower cost and/or higher quality (i.e. higher throughput and less keying substitution errors).