Disclosure: FileTally sells a local document-counting product. This guide explains where a count helps and where another measurement or professional decision is needed.
Start with the billing unit
A vendor may quote hosted data, processing volume, review hours or a combination of services. Read the quote's definition before multiplying a quantity by a rate. For a volume-based line, confirm the GB definition, the included representations, any minimum and the billing period. For a page-based line, confirm which rendered pages and copies are included.
One documented example is Everlaw. Its January 5, 2026 Billing and Pricing FAQ bases pricing on hosted data, describes monthly billing using peak native and processed sizes, and explains that generated productions add storage. This is Everlaw's published model, not a rule for every provider. Confirm current contract terms.
Our analysis: a source folder's bytes and pages can help explain what was handed over, but neither automatically equals that model's billable size. Processing outputs, included copies and timing can change the quantity. A discrepancy may reflect different scopes rather than an error.
Why pages per gigabyte varies
Pages and bytes describe different properties. Adobe's PDF optimization documentation identifies images, fonts and other stored elements that affect file size. Its compression guide documents desktop and web workflows. The narrow lesson is that storage size depends on how content is represented; it does not provide a pages-per-GB ratio.
A ratio from a named sample can be a labeled planning assumption for a similar sample. It is not a measured count of a different delivery. Preserve original documents; do not compress or convert evidence just to make a quote smaller.
A worked receipt with one exception
Synthetic example: these four entries and values are invented for arithmetic, not taken from a customer folder or a FileTally scan. Assume the first two files have been independently confirmed to be byte-identical; equal names or sizes alone would not prove that. Assume three PDFs were successfully measured and the fourth could not be read.
| File | Bytes | PDF pages | Outcome |
|---|---|---|---|
| packet-a.pdf | 800,000 | 12 | Measured |
| packet-a-copy.pdf | 800,000 | 12 | Measured; exact copy |
| packet-b.pdf | 400,000 | 8 | Measured |
| locked.pdf | 300,000 | Unknown | Unreadable |
- Receipt inventory: 4 files, 2,300,000 source bytes, retaining every copy. That is 2.3 MB using decimal units (1 MB = 1,000,000 bytes).
- Successfully measured pages: 12 + 12 + 8 = 32 PDF pages across 3 readable files. One file has an unknown count. Report “32 measured pages; 1 unreadable file,” not “32 pages in the entire delivery.”
- Pages with confirmed exact file copies counted once: 12 + 8 = 20 PDF pages within the readable subset, if the confirmed copy is counted once for this calculation. This does not establish unique pages or near-duplicate content, cover the unreadable file, or authorize removing a copy.
The byte total includes the unreadable file; the measured-page total does not. Dividing 32 by the full byte total would mix scopes and hide the exception. None of these numbers establishes that every relevant document was collected, reviewed or produced.
This example does not claim duplicate detection in the publicly downloadable FileTally version. Byte-identical files can have distinct source paths and evidentiary context. Similar text, email families and near duplicates require different comparisons.
A checklist for the vendor handoff
You can keep this record in a local spreadsheet or text file. For a small delivery, a filesystem inventory plus counts checked in the relevant document applications may be sufficient. A secure review platform is useful when you need processing, family relationships, privilege review or production controls.
- Freeze the receipt scope. Record the receipt date, source folder, sender's list and selected subfolders. Preserve the source. Compare what arrived with what was expected; keep missing and extra entries as open questions.
- Record files and bytes. State whether archives are still containers, whether extracted children are listed separately, and whether bytes describe the received container or expanded contents. Avoid adding a ZIP's bytes to its children as if both were independent source content.
- Record measurements by unit. Keep readable PDF pages separate from spreadsheet page estimates, other estimates and file inventory. A workbook's print area and page layout answer a print question; they do not count populated cells or verify formulas. See Excel printed-page measurements.
- Keep the exception list. Identify failed, unreadable, password-protected, unsupported and excluded items, and any provisional or estimated results. Unknown is not zero. Record the next action and owner, then retain the original report when remeasuring.
- Describe copy treatment. Keep totals containing every copy. If a separate unique-content figure is calculated, state its matching method, unit and coverage. Keep source paths and context; duplicate findings do not authorize deletion.
- Reconcile the quote. Ask the vendor to state billable quantity, GB definition, included source and processed representations, deduplication scope, minima and period. Compare like units and scopes. Use the agreed rate only after those quantities are defined.
Where FileTally helps
The current download is FileTally 1.0 for Windows and Mac. It counts supported local files and exports CSV or JSONL. Review counted pages, estimates and excluded files before using a total. Those exports support the receipt record; they do not establish the vendor's hosted-data size or replace legal review. See the formats and counting methods and current release and checksum.
In FileTally 1.0, Excel pages are estimated from each sheet's used range on a standard page, not rendered from the saved print layout, and populated cells are reported separately. Keep estimates separate from counted PDF pages and check the reported method.
FileTally processes document content on the user's computer. A cloud-synced folder may already have left that computer, and license activation is a separate one-time online request on each computer. Local counting does not secure the whole computer or guarantee confidentiality. Read the licensing terms before relying on offline availability.
Correction, October 6, 2026: this guide previously suggested that page counts should replace file-size estimates for vendor quotes. The right quantity depends on the quote's billing unit and scope. This revision adds separate-unit arithmetic, exceptions and a handoff checklist.
Questions people ask
How many pages are in a gigabyte of documents?
There is no universal conversion for a mixed collection. File type, images, fonts and compression affect bytes. Measure readable PDF pages separately from source bytes, and disclose files whose page count remains unknown.
How do I estimate the cost of an eDiscovery production?
Start with the actual quote: its billing unit, scope, rate and time period. Record source files, bytes, measured pages and exceptions separately. Ask the vendor to reconcile these with its billable quantities; a source-folder page count is not a hosted-data estimate.
Does FileTally replace an eDiscovery platform?
No. FileTally measures supported local files and exports a count report. Legal review, privilege decisions, production rules and the vendor's billing measurements require their own workflow.
Can I count pages without sending data to a vendor?
FileTally processes document content on your computer. Local counting does not require uploading those documents to FileTally. License activation is a separate one-time online request on each computer, and a cloud-synced source folder may already have been transmitted by its sync service.