Backfile Conversion Pilot Project: Sample Pricing (US)
The short answer
Run a paid pilot of one to five boxes with two vendors, using your hardest records, and choose on cost per correctly indexed page — not on the per-page quote. Quotes for the same backfile can differ by a factor of five, mostly because vendors include or skip preparation, indexing depth and QA. The pilot is where that hidden scope shows up. Figures below were checked October 20, 2026.
This is the pilot spoke of our best OCR companies for large-scale backfile conversion; QA rules and per-page ranges for full projects are in backfile conversion QA, metadata and per-page rates.
Sample pricing for a pilot
Using the industry range Access cites in its 2026 cost guide (8–12 cents per page for straightforward letter-size scanning, OCR about one cent more) and roughly 2,200 pages per banker’s box:
| Pilot size | Pages (approx.) | Scanning at $0.08–$0.12 | + OCR (~$0.01) | What is NOT included |
|---|---|---|---|---|
| 1 box | 2,200 | $176–$264 | $198–$286 | Prep, indexing, pickup, QA report, setup |
| 2 boxes | 4,400 | $352–$528 | $396–$572 | Same |
| 5 boxes | 11,000 | $880–$1,320 | $990–$1,430 | Same |
| 10 boxes | 22,000 | $1,760–$2,640 | $2,200–$3,300 (Access’s own example) | Same |
Add for: document preparation (staples, folds, sticky notes), field-level indexing (the fastest-rising cost), oversized or bound items, chain-of-custody pickup, and rush turnaround. Microfilm is priced per roll or per fiche instead — see microfilm to digital conversion providers.
How to run the pilot, step by step
- Write the spec first. Resolution, colour or grayscale, file format, document breaks, index fields with examples, OCR requirement and the QA sampling rule. Give the same spec to every vendor.
- Pick representative boxes. Different years, departments and record types; include the worst condition material. A pilot of only clean letter-size paper predicts nothing.
- Ask for the pilot quote in three lines: per-page production rate, setup and pickup, and whether the pilot fee is credited against the main contract.
- Inspect a sample of every batch. NARA’s rule for federal paper digitization — at least 10 records or 10% of each batch, whichever is larger, checked on a calibrated monitor at 100% — is a good private-sector benchmark. Have someone other than the vendor’s QC staff validate, as the federal rule requires.
- Reconcile counts. Pages scanned against pages in the box; every missing or blank page explained.
- Score index accuracy. Check every index field on a few hundred documents; for search-driven use, measure whether real queries find the right file.
- Record turnaround and rescans. Days from pickup to delivery, and how rescans were handled and charged.
Compare vendors on the right number
Cost per usable page = (pilot invoice + rescans + your staff time fixing errors) ÷ correctly indexed, QA-passed pages. A vendor at 9 cents with 6% index errors can cost more than one at 12 cents with 0.5%, once your staff correct files. Extrapolate the full project from that number, then negotiate volume pricing — volume is the most reliable lever on unit cost.
Red flags
- No written QA sampling method or rescan policy.
- A per-page rate with “prep and indexing billed as needed”.
- Refusal to run a pilot on your own boxes, or a free pilot on hand-picked easy pages.
- No chain-of-custody documentation for regulated records (HIPAA, IRS, legal).
Last verified: October 20, 2026.