Archive digitization costs: budget preparation and indexing alongside scanning

A scanner's capture rate is only one input to the cost of a usable digital archive. Staff may need to locate records, prepare them, preserve their order, capture images, name files, enter metadata, review results and resolve exceptions. Estimate those stages separately before agreeing a price per image or a completion date.

The U.S. National Archives' digitization strategy describes digitization as a series of activities, including preparation, metadata and quality control, rather than scanning alone. That workflow distinction is useful for a commercial estimate, while the requirements for a particular collection must come from its owner and applicable standards. A federal archival strategy is not automatically the governing rule for every private project.

Define the promised output: images, searchable files, a document index, structured extracted fields or a combination. Specify the approved quality and completeness requirements. A price for image capture should not silently become a commitment to verify every extracted data field or organize an archive that arrived without a usable inventory.

One physical sheet can produce two images, while one logical document can contain many sheets. A folded map may need a different capture method from a letter. State the counting unit for each stage so the estimate does not multiply a per-document indexing task by images or charge a per-image task against boxes whose contents are unknown.

Sample the actual mix before applying a throughput assumption. Include bound material, fasteners, mixed sizes, fragile originals, annotations and pages needing special handling where those occur. Have qualified staff determine how material can be processed without damage. The financial model should follow the approved handling method rather than assume everything can pass through an automatic feeder.

Record what the sample excludes. A small selection of clean loose pages may not represent a collection dominated by bound files or handwritten indexes. If the quantity or condition remains uncertain, use a transparent range or a defined discovery stage with an agreed decision point instead of presenting a precise total unsupported by the intake evidence.

Assume a hypothetical project contains 10,000 required images across 1,000 logical documents. Eight thousand images can be captured through an approved process at an observed 1,000 images per hour, requiring eight hours. Another 2,000 need a manual process at 200 images per hour, requiring ten hours. Capture therefore totals eighteen hours, not ten hours at the faster rate.

Suppose the reviewed workflow also needs ten hours of preparation, two minutes of indexing for each of the 1,000 documents, twelve hours of the agreed quality review and five hours of supported rework allowance. Indexing takes 33 and one-third hours. Total modeled labor is 78 and one-third hours, costing $2,350 at an illustrative loaded $30 per hour.

If the job sells for a hypothetical $0.30 per required image, revenue is $3,000. Adding $300 of attributable equipment, transfer and administration cost leaves $350 after the modeled $2,650 delivery cost, before other expenses outside this scope. Pricing from capture labor alone would have counted only $540 and materially overstated the apparent room in the quote.

Use a short reason record for exceptions such as an unreadable original, ambiguous document boundary, missing page, unusual size or capture failure. Record the time and the required decision. A growing exception queue is unfinished work, even if the scanner has produced the expected number of files.

Distinguish correcting a production error from resolving uncertainty in the source material. The contract may allocate those costs differently. Preserve the evidence rather than treating every difficult page as an extra charge or absorbing every new interpretation request into a fixed image price.

Set a threshold for reviewing material changes in the observed mix or exception rate. The threshold and authority should be agreed before the project consumes the full budget. If the collection contains substantially more manual work than the sample suggested, the parties need a concrete revised scope and estimate while the remaining work can still be planned.

Faster capture may not shorten the delivery schedule if indexing or review is already the limiting stage. Compare each stage's workload with the staff and equipment available for it. A large folder of unindexed images can be work in process rather than the completed archive the customer needs.

Plan handoffs in traceable batches so the team can reconcile originals, images and document groups without waiting for the entire collection to finish. Preserve the original relationships and the approved naming method. Batch size should balance handling efficiency with the ability to detect and correct a repeated issue before it affects too much material.

Include the agreed review and correction cycle in the delivery calendar. Customer questions or approval delays can affect completion even after production staff finish their part. Describe the required turnaround and what information the customer must provide instead of assuming immediate acceptance in the schedule.

Reconcile the required document and image population with accepted delivery, authorized exclusions and unresolved exceptions. Keep reworked images linked to their originals so a second scan does not appear as a second new customer document. The invoice's quantity basis should follow the approved agreement and accepted output.

Compare actual preparation, capture, indexing and review effort with the estimate. If the job exceeded budget, identify the stage and the changed input that caused it. A single overall productivity figure cannot show whether the problem was poor intake information, a manual-handling mix, ambiguous indexing or preventable capture errors.

Use that evidence to price comparable future collections and to distinguish fixed setup work from volume-driven work. Preserve custody, access and original-record disposition under the owner's separate approved requirements. A successful production budget supports a usable archive at a known cost without treating the scanning machine's speed as the whole service.