NLS: Digital talking book distribution analysis, Section Three: Key Variables

For: Digital talking book distribution analysis. Task 1 - Distribution system analysis and selection (Final report : September 16, 2005)

Section 3 - Cost Evaluation Model and Estimated Costs of Distribution Options

3.1 Cost Evaluation Model

The cost evaluation model is a spreadsheet tool developed using Lotus v.9.8 and has also been converted to Microsoft Excel format. It enables the user to perform "What-If" analyses when estimating the future costs of DTB distribution. The model addresses both NLS and network library costs, but not USPS costs. As previously stated, the model assumes that: (1) DTB playback machine implementation, (2) the residual RC book circulation system, and (3) DTB narration costs would be identical under all distribution options and are therefore not estimated or considered in the calculations. It was also not an objective of the model to compare the estimated future costs of DTB distribution with current distribution costs for RC books, therefore such comparisons are generally not enabled.

It was determined that the All Duplication on Demand distribution option is not a feasible implementation alternative. This is because it is economically infeasible, from NLS’s perspective, due to all recorded book distribution workload shifting from network libraries to NLS. For this reason, only the two feasible distribution options were developed in the cost model.

The model uses five data tables, each of which is shown for the "baseline," or most-likely environmental scenario, in Appendices 57 through 61. The contents and logic of these tables and the overall logic and assumptions used in the model are documented below. For the purposes of illustration, it was arbitrarily assumed for the Hybrid option that 40% of new titles would be mass-duplicated.

NLS Costs Table (Appendix 57)

The table consists of four columns:

a description of the key variable or factor
one column for each of the two distribution options
a comments field

Data elements that are potentially user-specified are labeled with "user" inside square brackets preceding the element name. Those which should not be changed have no label. Data elements are shown for both DOD center and mass-duplication portions of the DTB distribution system.

[user] Annual Systemwide DTB Circulation (Copies): a default value of 20,000,000 copies per year is used for the baseline scenario, as cited in Section 2.
Percentage of Systemwide Circulation Handled by DOD Centers: calculated by the model for the Hybrid option using the Percentage of New Titles Mass-Duplicated variable and the Workload Split Table (both described below). The value for the All Mass-Duplicated option is always 0%.
Annual Circulation Handled by DOD Centers: calculated by the model as the product of circulation and the percentage of DOD circulation for an option.
[user] Unit Cost to DOD DTBs: $0.66 per copy is used as the default, the derivation of which is described in Section 2. This includes cartridge reconditioning and labelling, assumes the use of a reusable envelope, and excludes the cost of both cartridges and reusable envelopes, both of would be GFE.
Annual Costs for DOD Centers Excluding Flash Memory and Reusable Envelopes: calculated by the model as the product of the unit cost to DOD a DTB copy and the circulation.
[user] Loss Rate as a Percentage of Circulation: 3% is used as the default, the derivation of which is described in Section 2.
Number of Cartridges Lost Annually that are Replaced by NLS: calculated by the model as the product of circulation and the loss rate.
[user] Wholesale Unit Price of Flash Memory Cartridge: $6 is used as the default, as is a 256 MB capacity, both of which were described in Section 2.
Annual NLS Costs for Replacement of Lost Cartridges: calculated by the model as the product of the unit cost for a cartridge times the number of cartridges lost annually.
Number of Reusable Envelopes Lost Annually that are Replaced by NLS: calculated by the model as the product of circulation and the loss rate.
[user] Wholesale Unit Price of Reusable Envelopes: $0.50 is used as the default, which is the current price of an RC container.
Annual NLS Costs for Replacement of Lost Envelopes: calculated by the model as the product of the unit cost for an envelope times the number of envelopes lost annually.
[user] Average DTB Turnaround Time (Days): 35 days is used as the default, the derivation of which is described in Section 2.
Average Circulation per Cartridge per Year: calculated by the model as the quotient of 365.25 days divided by the average turnaround time in days.
Working Inventory of Cartridges Required for Operations: calculated by the model as the quotient of annual circulation provided via DOD divided by average circulation per cartridge per year.
Initial Investment in Cartridges Required for Operations: calculated by the model as the product of cartridge unit price and the number of cartridges required for start-up.
Working Inventory of Reusable Envelopes Required for Operations: calculated by the model as the quotient of annual circulation divided by average circulation per envelope per year.
Initial Investment in Reusable Envelopes Required for Operations: calculated by the model as the product of reusable envelope unit price and the number of envelopes required for start-up.
Estimated Initial Investment in Cartridges and Reusable Envelopes: calculated by the model as the sum of the initial investments required for both cartridges and envelopes.
Estimated Annually Recurring Operations Costs: calculated by the model as the sum of DOD operating costs and the costs of replacement cartridges and envelopes for those lost.
Estimated Total Annual Costs for DOD Center Operations: calculated by the model using a 10-year, straight-line depreciation of initial investment costs summed with annually recurring costs.
[user] Percentage of New Titles Mass-Duplicated: the value for the Hybrid is user-specified, while that for the All Mass-Duplicated option is always 100%.
[user] New Titles Produced (Mass-Duplicated) Annually: 2,000 titles per year is used as the default for the All Mass-Duplication option, as described in Section 2. The value for the Hybrid is calculated as the product of the percentage of new titles mass-duplicated and the number of new titles for the All Mass Duplicated option.
Copies per Title Required: calculated by the model for both options using the Percentage of New Titles Mass-Duplicated variable and the Copies per Title Table (described below).
Copies Mass-Duplicated per Year: calculated by the model as the product of the number of titles mass duplicated and the number of copies per title.
[user] Mass-Duplication Unit Cost per Copy: $0.804 per copy is used as the default, which consists of $0.55 for cartridge labelling and the duplication process itself, $0.09 for container labels and $0.164 for labelling the container. It excludes the costs of the cartridge and container both of which would be provided GFE and are separately calculated in the model, and the costs of reconditioning used cartridges and containers, which are also separately calculated in the model.
[user] Container Price per Copy for Mass-Duplicated Copies: a default of $0.50 is used, which is the current price of a 4-cassette container.
Cost per Year for DTB Mass-Duplication Before Reuse Offset: calculated by the model as the sum of the cartridge, container and duplication costs multiplied by the number of copies mass duplicated.
Percentage of Mass-Duplicated DTB Cartridges Reused: 47% is used as a default which was calculated from FY 2004 data provided by NLS for RC recycling operations, with 68% as a maximum (as described in Section 2) and probably 25% as a reasonable minimum.
Number of Mass-Duplicated DTB Cartridges Reused: calculated by the model as the product of the cartridge reuse rate and the number of cartridges used in mass-duplication.
Cost Offset from Reuse of Mass-Duplicated Cartridges: calculated by the model as the product of the number of cartridges reused and the cartridge unit cost.
Unit Cost to Recondition DTB Cartridges: $0.109 is used as a default, which was calculated in analyses described in Section 2.
Cost to Recondition Mass-Duplicated DTB Cartridges: calculated by the model as the product of the number of cartridges reused and the unit cost to recondition them.
Net Cost Offset from Reuse of Mass-Duplicated DTB Cartridges: calculated by the model as the difference of the cartridge reconditioning cost subtracted from the gross cost offset.
Percentage of Mass-Duplicated DTB Containers Reused: 24% is used as a default which was calculated from FY 2004 data provided by NLS for RC recycling operations. The maximum is likely 30%, the minimum 20%.
Percentage of Mass-Duplicated DTB Containers Disposed: 23% is used as a default which was calculated from FY 2004 data provided by NLS for RC recycling operations.
Number of Mass-Duplicated DTB Containers Reused: calculated by the model as the product of the container reuse rate and the number of containers used in mass-duplication.
Number of Mass-Duplicated DTB Containers Disposed: calculated by the model as the product of the container disposal rate and the number of containers used in mass-duplication.
Unit Cost to Recondition DTB Containers: $0.33 is used as a default, which is the current unit cost to recondition RC containers.
Unit Cost to Inspect/Dispose DTB Containers: $0.06 is used as a default, which is the current unit cost to inspect and dispose of RC containers.
Cost to Recondition Mass-Duplicated DTB Containers: calculated by the model as the product of the number of containers reconditioned and the unit cost to recondition them.
Cost to Inspect/Dispose Mass-Duplicated DTB Containers: calculated by the model as the product of the number of containers inspected and disposed and the unit cost to inspect and dispose of them.
Cost Offset from Reuse of Mass-Duplicated Containers: calculated by the model as the product of the reuse rate for containers and container unit cost and the number of copies mass duplicated.
Net Cost Offset from Reuse of Mass-Duplicated DTB Containers: calculated by the model as the difference of the container reconditioning cost subtracted from the gross cost offset.
Estimated Net Costs per Year for Mass-Duplicated DTBs: calculated by the model as the cost per year prior to the reuse offset less the net cost recovery associated with reused cartridges and containers.
Total Annual NLS Distribution and Duplication Costs: calculated by the model as the sum of DOD center and mass-duplication total costs.

Workload Split Table (Appendix 58)

This table is used by the NLS Table in the model to select the cumulative circulation associated with a user-specified percent of new titles mass-duplicated for the Hybrid option. Titles are ranked in descending order of popularity, and cumulative circulation estimated for both the DOD centers and network libraries are shown.

Copies per Title Table (Appendix 59)

This table is used by the NLS Table in the model to select the copies per title required for the mass-duplication portion of the Hybrid option. Titles are ranked in descending order of popularity.

Library Cost Table(Appendix 60)

The elements contained in this table are taken from the NLS LCC model, with FY 2004 values used. The first step in the calculations was to estimate the current costs attributable to recorded book operations versus those for playback machine operations (which the total LCC values include). Factors available in the LCC model which contain this percentage (shown in the "Percent Recorded Book" column in Appendix 60) were used to make this allocation.

The next step was to focus upon the costs of those library recorded book functions that are directly distribution related and extremely likely to change under the different distribution options. The costs for these functions are shown in the "Vary with Distribution Option" column in the table.

Finally, the costs for distribution functions for the All Mass Duplication and Hybrid options are estimated. The All Mass Duplication option is calculated relative to the costs of current RC operations, and the Hybrid option costs are calculated relative to the costs for the All Mass Duplication option. Below are listed the distribution-related recorded book functions that are expected to vary under the two DTB distribution options.

The mass duplication option is assumed to require approximately the same cost as that for the current system, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries.
Network Labor: Inspect Books, Cost of Book Inspections Performed Upon the Issuance or Return of Books: The mass duplication option is assumed to require approximately 50% of the effort as that required for the current system, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries.
Network Labor: Duplicate Books, Cost of Reproduction of NLS Books: Based upon information provided by NLS, it was estimated that the All Mass Duplication option would require 75% of the current effort for this function, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries.
Network Labor: Build and Maintain Collection, Cost of Weeding / Shifting of Book Collections, etc.: The mass duplication option is assumed to require approximately the same effort and cost as that required for the current system, and for the Hybrid the model applies a reduction commensurate with the reduction in the number of titles managed in the library collections.
Network Labor: Repair Books, Cost of Repair of Recorded Books, etc.: The mass duplication option is estimated to require approximately half the costs as that required for the current system, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries.
Network Labor: Check Out and Delivery, Cost of Retrieval of Materials from Storage, etc.: The mass duplication option is assumed to require approximately the same costs as those required for the current system, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries.
Network Facilities: Facility and Occupancy, Costs of Facilities and Occupancy: It was assumed that 40% (based upon current RC storage at a sample of seven regional libraries, ref. Appendix 56) of total facility space costs are associated with recorded book collection storage. The All Mass Duplication option is then estimated to require 50% of this space (and costs) for the storage of DTBs using smaller media and storage containers, and efficient storage and space management practices. For the Hybrid option, the model applies an additional reduction which is proportionate to the reduction in the number of titles that would no longer be stored in the library collections.
Cost of Supplies: The All Mass Duplication option is assumed to require approximately the same costs as those required for the current system, and for the Hybrid the model applies a reduction commensurate with circulation workload reduction in the libraries. It is estimated that the cost of flash memory cartridges themselves will be approximately half the total amount shown based upon data supplied by NLS for duplication of RC copies in libraries for NLS titles.
Total Annual Library Distribution and Duplication Costs: This value, calculated for each option, is the sum of all the elements listed in Appendix 60 for the relevant option.

Total Costs Table (Appendix 61)

This table simply shows the calculated costs for NLS-borne distribution and duplication costs, those borne by the libraries, and total costs for each of the distribution options.

3.2 Estimated Costs of Distribution Options

Analyses were performed which examined the sensitivity of NLS, network library and total systemwide costs to the split in workload in the Hybrid option, i.e., for various combinations of proportions of new titles mass-duplicated and their associated workload splits between DOD centers and network libraries. The results of these analyses for the baseline scenario (i.e., using the expected values of all key variables and factors in the model), are shown in Appendix 62.

In every instance, total systemwide costs and network library costs are lower for the Hybrid option than for the All Mass Duplication option, while for NLS a minimum cost is attained for scenarios which call for mass-duplication of titles in a specific range. At very low levels of mass-duplication of titles, NLS takes-on a relatively large proportion of total system circulation and would incur costs above what it would incur if the All Mass Duplication option were implemented; with only 5% of titles mass-duplicated, NLS would incur annual costs of about $14.5 million versus about $8.3 million for the All Mass Duplication and about $5.5 for comparable current RC functions. As noted, however, both library and systemwide costs are minimized at these levels, with total system costs approximately $17.5 million versus about $24.7 million for the All Mass Duplication option; libraries would have costs in the $3-to-$5 million range versus about $16.3 for the All Mass Duplication option. But it is simply unrealistic for NLS to assume costs well above current costs and estimated costs for the All Mass Duplication option.

On the higher end of this scale, although NLS costs are minimized with 60% of titles mass-duplicated (in this baseline scenario), systemwide costs increase by about $3 million from their lows, with library costs increasing from the $3-to-$5 million range to about $13.7 million (although this is still about $3 million less than libraries would incur in the All Mass Duplication option). However, another possibly more restrictive consideration is that, with 60% of titles being mass-duplicated, DOD centers would only circulate an estimated 6.6% of systemwide demand. This would almost certainly mean that only one DOD center would operate, increasing the degree of operating risk. It may be that no more than 50%, or even 45% or 40%, of titles could be mass-duplicated in order to have sufficient workload for two centers to operate efficiently, realize economies of scale, and operate within the bounds of the conceptual operation developed and described in this report.

In considering the totality of the economics of the system, the economics of its components individually, and the constraint regarding sufficient workload for two centers, it is recommended that if a title-based Hybrid business model is implemented that the number of new titles mass-duplicated fall into a range from 30%-to–50%, with 35%, 40% or 45% being more likely candidates for implementation than either 30% or 50%. At these levels: (1) significant systemwide savings occur; (2) NLS incurs at most and probably less cost than the amount that it would incur for the All Mass Duplication option; (3) libraries realize significant savings relative to an All Mass-Duplication mode; and (4) sufficient DOD workload would exist to support two centers employing efficient production techniques and realizing adequate economies of scale.

Some additional sensitivity analyses of the baseline scenario model yielded the following general findings:

Changing the unit cost to DOD produces a moderate change in costs for the Hybrid and no change for the All Mass Duplication option assuming a range of variation within expected bounds.
Changing the unit price of flash memory has an impact upon both options, with the greatest relative impact upon the All Mass Duplication option.
Changing the unit price for mass duplication has little to moderate effect upon the Hybrid option, and moderate effect upon the All Mass Duplication option assuming a range of variation within expected bounds.
Changing the percentage of cartridges that are reused (for mass duplication) has a minor impact upon the Hybrid option and a significant impact upon the All Mass Duplication option assuming a range within expected bounds.
Changing the loss rate has no effect upon the costs of the All Mass Duplication option (losses are absorbed by libraries and this attrition is effectively a type of unplanned and uncontrolled weeding of the collection) and a minor effect upon the costs for the Hybrid option assuming a range of variation within expected bounds.
Changing turnaround time has no impact upon the All Mass Duplication option costs, and only a minor impact upon costs for the Hybrid option assuming a range of variation within expected bounds.

NLS: That All May Read

Digital Talking Book Distribution Analysis