KavarneDiscovery for law firms and legal departments
KAV · 300001

Native vs TIFF: eDiscovery format risk planner

Converting native files to TIFF is how most productions are made, and for some file types it silently removes content. Tick what is in your collection and see what images cleanly, what loses evidence, and what to put in your ESI protocol before you agree a format you cannot undo.

Updated September 6, 2026

What's in the collection?

Tick everything you expect to encounter.

Documents & mail
Messaging & mobile
Scanned & media
Specialist & structured
Known trouble
2
Image cleanly
1
Lose content
0
Need special handling

Word documents and text PDFs

Images cleanly
→ Image (TIFF or PDF) with load file and extracted text

Converts reliably. Bates numbering, redaction and confidentiality endorsement all work as expected on the image.

Email and attachments

Images cleanly
→ Image with load file, or near-native EML/MSG

Images well. Family relationships between parent messages and attachments must be preserved in the load file, and the time zone used for date fields should be stated.

Spreadsheets

Loses content
→ Native, with a Bates-numbered placeholder image

Imaging discards formulas, hidden rows and columns, hidden worksheets and cell comments. This is the category behind the 484-page file that emerged as 12 pages. Formulas are frequently the evidence.

Suggested protocol language

How the planner scores a collection

Five representative collections, each run through the same assessment as the planner above, from formats.json. Green images cleanly, amber loses content when imaged, red needs special handling.

Imaging risk counts for five representative collections
CollectionImages cleanlyLoses contentSpecial handling
Word documents and email only200
Typical office collection220
Office collection plus scanned paper230
Chat and mobile messaging added212
Everything in formats.json258

Format is fixed in the ESI protocol, and the choice drives several checks on the production checklist.

Why format is decided before collection, not after

The form of production is the most consequential clause in an ESI protocol, and the most expensive one to get wrong. If you collect and process in one form and opposing counsel successfully insists on another, the material may have to be re-collected, re-processed and re-reviewed. That figure dwarfs any per-gigabyte conversion fee, and it is entirely avoidable by settling the question up front.

What imaging actually removes

A TIFF is a picture of a document. Anything that is not visible on the face of the page does not survive the conversion: spreadsheet formulas, hidden rows and columns, hidden worksheets, speaker notes, tracked changes, comments, and embedded objects. Images also arrive stripped of metadata and searchable text, which then has to be produced separately in load files and reassembled at the other end.

Craig Ball has made the point bluntly for years: TIFF productions run many times larger in byte volume than native production of the same material, which makes them significantly more costly for the receiving party to ingest and host, while working poorly for spreadsheets and presentations, and not at all for video and sound.

Recommended form of production by file-type category
CategoryRecommended formWhy
Word, text PDFs, emailImage with load file and extracted textConverts cleanly, Bates numbering and redaction are straightforward
Spreadsheets, presentationsNative with placeholder imageImaging discards formulas, hidden content and notes
Audio, video, databasesNative, or agreed report or exportNo meaningful page-based representation exists
Scans and photographsImage with OCRNo extractable text until OCR is applied; cost is per page
Chat and mobile messagingNative or near-native, unitization agreedNo natural document boundary: what counts as one "document" must be agreed

Exceptions are billed as hours, not volume

EDRM defines an exception as a document the processing tool cannot completely interpret, requiring reporting or remediation. Password-protected, encrypted and corrupt files accumulate in almost any collection, and each one is resolved by a human rather than a pipeline, which is why they land on the invoice as hourly work.

Container files deserve particular attention. EDRM treats an exception encountered while processing a mail or forensic container as critical, because a single failed PST or NSF can silently omit everything inside it. That is not a cost problem, it is a completeness problem, and it is the kind that surfaces in a sanctions motion rather than on a bill.

What this tool deliberately does not do

Format conversion is itself a billed line: the Fee Field Guide sets out how TIFFing is charged, and our pricing page includes production in the flat rate. It does not estimate an exception rate or predict your conversion bill. No reliable published figure exists for how often files fail processing (it depends entirely on the collection, the sources and the tool) and inventing a percentage would make this look precise rather than useful. What it does is tell you which categories carry risk, so you can price them with your own vendor and write them into the protocol. For the money side, use the cost calculator with rates from your actual quote.

Common questions

Should spreadsheets be produced native or as TIFF images?

Native. Converting a spreadsheet to a static image discards formulas, hidden rows and columns, hidden worksheets, comments and cell-level formatting. Practitioner consensus and most model ESI protocols treat spreadsheets as a category that should be produced in native format with a Bates-numbered placeholder image, precisely because imaging removes content that may be evidentiary.

What file types cannot be produced as TIFF images?

Audio and video files cannot meaningfully be imaged at all, and databases and other structured data have no natural page-based representation. CAD drawings are often held in structured storage formats requiring specific discussion between the parties. These categories are normally produced natively, or by agreed report or export in the case of structured data.

What is a processing exception in eDiscovery?

An exception is a document the processing tool cannot completely interpret and which requires reporting or remediation. Common causes include password protection, encryption, file corruption, unsupported or proprietary formats, and files without extensions. Exceptions accumulate in almost any collection and are usually resolved by hand, which is why they are billed as hourly work rather than by volume.

Why is agreeing production format early important?

Because reversing the decision is expensive. If a party collects and processes data in one form and opposing counsel successfully demands another, the material may have to be re-collected, re-processed and re-reviewed. That cost typically exceeds any conversion fee, and it is avoided entirely by settling the form of production in the ESI protocol before collection begins.