The Massachusetts Department of Revenue (MDOR) undertook an ambitious leap into 2 dimensional barcode technologies by processing Residential Personal Income returns (Form-1) and Corporate Excise returns (Form 355, Form 355-S & Form 355-C) in 2003.Our experiences are presented in respective approaches and conclusions

FORMS APPROACH

Starting with Income, the Form-1 and its associated schedules (Schedule B, C, CB, D, E, INC & XYZ) were 2-D enabled and 7 software vendors chose to support it. As a fallback, we also required ‚ exact positioning‚ of data fields on the forms & schedules so that OCR / ICR technologies could be used to read the imaged returns. 22 total software vendors (including the 7 2-D vendors) support exact positioning for income forms.

To help insure return integrity, we chose to have each form and schedule have it’s own 2-D barcode on it.The barcode is printed on the top right hand corner of the first page of the return or schedule.We initially looked at a cover sheet approach where the 2-D barcodes are printed on a separate piece of paper.We felt that approach would lend itself to data integrity issues between the printed and encoded 2-D data as well as the possibility that the cover sheet would not be submitted.

Corporate returns were the next concern. Our 3 Corporate Excise Forms and 15 associated Schedules were 2-D enabled and 3 software vendors chose to support it. Also new this season for the corporate forms was making them OCR / ICR “exact positioning” (image enabled) compatible. 18 total software vendors (including the 3 2-D vendors) support exact positioning for corporate forms.

To emulate the income forms, we chose a multiple 2-D approach, but with a slight twist. On the corporate forms we are looking for a 2-D barcode on EVERY page. There are several multiple page schedules, for example schedule H has 6 pages, so there would be 6 2-D barcodes for that single schedule.

SYSTEMS APPROACH

Systemically, PIT and Corp are handled the same.

Our manual document preparation process, among other things, search for returns with 2-D barcodes and separate them from non 2-D returns. Those 2-D returns with any manual changes applied to the printed data are out-sorted to the non 2-D pile.

The 2-D batches are scanned using IBML high-speed scanners. The non- 2-D” batches are scanned using Kodak high-speed scanners.

Based on our approximate average of 95% success rate in reading the 2D barcodes, its admittedly redundant for the 2-D batches. The process is very quick and therefore does not slow the overall processing time at all. In fact the recognition process is as quick and sometimes quicker than the scanning process. The redundancy gives us the luxury of not having to rescan failed 2-D returns into our conventional imaging system.

The 2-D batches have the barcode read, decoded, imported and validated. The validation at this step is limited to 2-D formatting. Those validations are as follows:

Invalid 2-D barcode data is ignored and recognized data is used in its place.

All data is run through a second validation process that verifies the data in terms of limits and mathematical calculations. Those returns that fail the second validation process gets passed to a data verify process. A data verification operator will review the return fixing obvious recognition errors.

All the data is then formatted for upload to our DEC back end for further validations after being merged with data farmed from all other data collection technologies. (E.g.: Electronic Filing, Manual Data Entry, etc.)

The images of all scanned documents are also uploaded to a folder system and stored on high-density EMC storage devices.

EDUCATION APPROACH

During the first 2 weeks of January, key representatives (technical, legal and executive management) from MDOR conduct Practitioner Seminars at different locations throughout the state. These seminars provide an opportunity to inform the practitioner community of new issues and services that will be in place for the coming filing season. Given that this was our pilot year, 2D barcoding was the 1st new item on the agenda.

Anytime that the MA practitioner community needed to be addressed about a potentially serious/noteworthy issue, a large directed e-mail campaign was undertaken using the respective database developed/maintained by MDOR. This provided a platform on which MDOR could address concerns as well as an opportunity for the practitioners to feedback/ask additional questions directly to MDOR.

The MDOR website was also a vehicle in which we would announce and/or educate (to the best of our ability) the advent of 2D barcoding to PIT return processing (emphasis on positives derived from using 2D supportive software).

INITIAL CONCLUSIONS

It is a little early in the processing year to fully compile our lessons learned list. However, early results from this endeavor seem to indicate a success. The confidence level is high enough in the department to warrant expansion of both income and corporate systems by introducing additional forms and schedules in both arenas. We will be examining why certain returns failed the 2-D processing and, equally important, those returns that went through the system completely without human intervention. Finally, we will be polling our business partners to determine where we did well and where we failed in terms of implementing the 2-D technology.

By the numbers (Income)* as of June 2, 2003

2-D paper returns processed - 498,944

Conventional paper returns imaged - 580,971

Electronic Filing returns** - 785,832

Overall returns processed all methods - 2,164,744

Valid 2-D Barcodes*** - 96.26%

Form 1 - 96.33%

Form INC - 98.57%

Sch B - 98.29%

Sch C - 98.27%

Sch D - 91.04%

Sch E - 94.87%

Sch X/Y/Z - 96.50%

*Corporate numbers are not included here because that system is just coming up live in our production environment.

** Electronic Filing included technologies used by a paid preparer and software sold off the shelf to taxpayers directly. (In some circles differentiated as Professional and Home software).

*** Read rate data for Schedule CB unavailable pending verification