### The Cornell / Xerox / Commission on Preservation and Access Joint Study in Digital Preservation

## Appendix III–Cost Study Assumptions for Table A

Some of the cost data used in this analysis are based on confirmed measurements and quotations. Others are assumptions or projections. These measurements and quotations, along with the rationales for the assumptions and projections, are presented in this Appendix.

All assumptions are based on costs associated with preserving a 300 page book. The unit costs are calculated in 1992 dollars, and the average annual rate of inflation is figured at 5%. For all equipment, the annual costs are calculated at 47.2% of the capital cost. This figure is calculated to reflect the initial cost of the equipment, a four-year amortization, and a 10% interest rate; 50% is added for space, utilities, and maintenance. This figure represents a very conservative estimate of costs: the interest rates have fallen and the 50% load may not apply to all equipment, but for consistency it is used throughout. The rates of change in the cost of computer equipment and storage are based on historic price declines of the last decade and industry projections. For instance, optical disk storage costs are assumed to decline by 30% per year, scanning equipment by 10%, and printing equipment, which includes a number of mechanical features, to decline by only 5% per year. Obviously, any one of these rates of change could be debated: they should be considered indicative of trends currently underway rather than definitive figures.

Costs were projected for a ten year period. As Chart A indicates, the cost of digital technology will rise slightly over the decade, since the declining costs of technology are dominated by increases in labor and finishing. The labor figure for scanning assumes that there will be no increase in production, which almost certainly will occur as institutions move from prototype to production operations, as service bureaus begin to offer this service, as improvements in automatic and semi-automatic feed mechanisms reduce the risk of a paper jam, as automatic skew correction and page definition become standard, and as bound volume scanners are developed to enable the scanning of two pages at a time.[****]

The numbers below refer to the line numbers in Table A.

**Scanning Costs:**- 1. Labor. It takes a scanning technician 1.72 hours to scan an average 300 page book. The 1992 hourly rate, including benefits, is $12.00/hour. ($12 X 1.72 = $20.64) The ten year outlook does not include an increased production rate, which almost certainly will occur. The costs of labor are projected to increase by 5%/year.
- 2. Equipment. Equipment includes scanner, personal computer, high resolution monitor, application software and network connections. Annual scanning equipment cost is $9,440 ($20,000 X. 472). The equipment is assumed to be operational for 2 shifts. Assuming a standard shift equals 37.5 hour work week, the total number of hours/year that the equipment is in use is 3,300. The hourly cost for equipment ($2.86) is computed by dividing the total number of hours per year by the annual scanning equipment cost ($9,440/3300). Since the average scanning time per book is 1.72 hours, the scanning equipment cost per book is therefore $4.92. This figure declines by 10%/year.
**Transfer to Optical Storage:**- 4. Optical Jukebox refers to the cost of the time it takes to write a disk on the jukebox (20 minutes for 300 page book). This figure is calculated by taking the annual cost of jukebox ($75,000 X. 472) and dividing it by 21,900 (which equals 20 minutes of a year: 365 days/20 hours per day,2 divided by 3 for the 20 minutes). This figure declines by 15%/year, which is very conservative, given declines in technology cost and time taken to write the disk.
- 5. Optical Disk Cost. This figure represents the cost of the 12″ disk ($500) amortized over 4 years, and divided by the number of books that will fit on a disk. $500 X .472 / 44 = $5.36. This figure declines by 30%/year.
- 6. Technological Refreshing. The cost of refreshing, which begins in year 4 and is repeated every four years, is calculated by adding the cost of equipment (#4 above), which is the cost of the time taken to transfer to a new disk, and the cost of that portion of a new disk that the volume occupies (#5 above). Certainly the time between refreshing will increase as data exchange standards are developed.
**Printing and Binding:**- 7. Print Equipment. This figure includes the cost of the Docutech printer, maintenance, set-up for printing, and associated costs. Cornell has set a price of printing at $.0225/side (excluding paper), which is based on full costing of the Docutech, including allowances for space and overhead. A 300 page book would cost $6.75. This figure will decline by 10%/year.
- 8. Acid Free Paper. The cost of acid free paper is calculated at $.01/sheet (a 300 page book would be printed on 150 sheets). This figure will increase by 5%/year.
- 10. Library Binding. The binding figure includes the cost of full cloth binding (Ridley’s Book Binder) and preparation for binding. It will increase by 5’/0/year.
- 11. In-line Finish. The Docutech offers a number of finishing options, including a heat-set tape binding, the price of which is listed here. This figure will increase by 5%/year.
- 12. Unbound / Stapled. A nominal figure is included here for stapling.
- 13. Weighted Binding Cost. Binding costs for subsequent copies are based on the assumption that not all requests will result in a library binding. It is assumed that a full cost library binding will be done for 20% of the print requests, an inexpensive in-line finishing will be used for 40%, and that the remaining 40% will be stapled or left unbound.
- 14. Access Cost / Book. The cost of accessing the book assumes that the optical disk containing the book will be mounted on the jukebox when needed. This cost is calculated as the cost of one tenth of an hour ($1.20) and 5 minutes of optical jukebox time ($.40) (See #4 above). This figure will decline by 10%/year.