What is the most important question to ask before a document scanning project?

How the scanned files will be indexed and named. Image quality matters, but an accurate scan nobody can find delivers no value. Agree the metadata fields, document separation rules and naming convention in writing before any paper is handled, because retro-fitting an index means processing the documents twice.

Do I need a data processing agreement with a scanning provider?

Yes. Under UK GDPR and the Data Protection Act 2018 you remain the data controller and the scanning bureau acts as your processor, so Article 28 requires a written contract covering the scope of processing, security measures, sub-processors and breach notification. Ask whether any work, including offshore index keying, takes data outside the UK.

What OCR accuracy should I expect from a scanning provider?

Clean laser-printed text commonly exceeds 99% character accuracy, but faxes, carbon copies, poor photocopies and handwriting fall well below that. Ask the provider to run a free sample using your most difficult documents rather than accepting a headline figure based on their own demonstration set.

What happens to original documents after they are scanned?

You choose: secure destruction to BS EN 15713 with a certificate, return to your premises, or transfer into managed off-site archive storage. Keep a quarantine period of at least 30 days after you have verified and accepted the scans, and retain originals where wet-ink signatures are legally required, such as property deeds and share certificates.

Should I run a pilot before scanning my whole archive?

Yes. A small paid pilot of around 10 representative boxes tests image quality, index accuracy, file naming, delivery format and turnaround against written acceptance criteria before you commit the full archive. It is the cheapest way to discover a mismatch between what you specified and what you will actually receive.

What Questions Should You Ask Before Starting a Document Scanning Project?

Before you send a single box out of the door, you need answers on six things: what you are actually scanning, how the files will be indexed, who touches the paper and where, what image and OCR quality you will accept, what happens to the originals, and how the digital files reach your systems. Scanning projects rarely fail on the scanning itself — they fail because nobody asked how the output would be found, verified, or handed over. The questions below are the ones that separate a project that pays for itself from a hard drive full of unsearchable PDFs.

Scope: what exactly are we digitising, and why?

“Scan the archive” is not a scope. A UK business with 400 standard archive boxes is looking at roughly 800,000–1,000,000 images at typical fill rates (2,000–2,500 sheets per box). Scanning all of it when only 15% is ever consulted is the single most common way to burn a digitisation budget.

What is the business trigger? Office move, audit, subject access requests, a team that wastes hours hunting files — each implies a different priority order.
Which records are past their retention period? Anything beyond retention should be securely destroyed, not scanned. Scanning expired personal data creates a fresh GDPR liability rather than removing one.
Does everything need the same treatment? Live client files may justify full OCR and detailed indexing; dormant 1990s ledgers may only need box-level off-site storage with scan-on-demand.
What is the document mix? Stapled bundles, treasury tags, A3 plans, carbon copies, double-sided forms and photographs all change prep time — which is usually the largest single cost in any scanning job.

Indexing: how will anyone find a file afterwards?

This is the question most often skipped, and the one that decides whether the project delivers value. A scanned image with no index is a digital equivalent of an unlabelled box. Agree the metadata fields before scanning starts, because retro-fitting them means handling the paper twice.

Questions to settle in writing

How many index fields per document, and which ones? Three to five is typical — for example client surname, matter or account reference, document type, and date.
Is indexing keyed manually, captured from barcodes, or pulled by OCR from a fixed zone on the page? Manual keying is the most accurate on messy documents and the most expensive.
What defines a “document” versus a “page”? If a 60-page file is delivered as 60 separate PDFs, your team will hate the result. Separator sheets and barcode splitting need agreeing up front.
What is the file naming convention, and does it match how staff already search? Consistency matters more than cleverness.
What accuracy level is guaranteed on index data, and how is it measured — per field or per character?

Security and compliance: who handles the data, and where?

Under the UK GDPR and the Data Protection Act 2018 you remain the data controller throughout. The scanning bureau is your processor, and Article 28 requires a written contract covering what they may do with the data. The ICO can issue fines up to £17.5m or 4% of global turnover, so this is not a paperwork formality.

Is there a signed data processing agreement, and does it name any sub-processors?
Does any part of the work — including offshore keying of index data — leave the UK? If so, what transfer mechanism applies?
Are staff DBS-checked and vetted to BS 7858, and is the scanning floor access-controlled and CCTV-covered?
Is the provider certified to ISO 27001 for information security and ISO 9001 for quality, and can you see the current certificates with their scope statements?
What is the chain of custody from collection to delivery — sealed containers, GPS-tracked vehicles, barcode scan-in on arrival, and an audit trail you can be shown?
How would a data breach be reported to you, and within what timeframe? You have 72 hours to notify the ICO, so a processor promising “prompt” notification is not good enough.

Quality: what does “good” look like, and who checks it?

Ask for the provider’s written quality control procedure rather than a promise of accuracy. For records that may be needed as evidence, BS 10008 is the relevant UK standard for the legal admissibility of electronic information — worth asking about specifically if you handle contracts, medical records or anything likely to end up in a dispute.

What resolution and colour mode — 200 dpi bitonal is adequate for clean typed text, but 300 dpi is the safer default and colour is essential for signatures, highlighting, and coloured stamps.
Is every image checked, or a sample? Ask what percentage, and what happens when a batch fails.
How is page count reconciled against the original file so missed or double-fed pages are caught?
What OCR accuracy is realistic on your documents? Clean laser-printed text can exceed 99%; faxes, carbon copies and handwriting fall dramatically. Insist on a sample test using your worst paper, not the provider’s demo set.
What file format is delivered? Searchable PDF suits day-to-day use; PDF/A is the archival choice for long-term retention.

Originals, delivery and cost structure

What happens to the paper?

Decide before collection, not after. Some documents must be kept in original wet-ink form — property deeds, certain wills, share certificates and some HR records. Everything else can typically be securely destroyed once scans are verified and accepted. Ask for a quarantine period of at least 30 days, then certificated secure destruction to BS EN 15713, or return to managed archive storage if retention rules still apply.

How do the files reach us?

Encrypted portal download, encrypted drive, or direct import into your document management system or SharePoint?
Is there an index file (CSV or XML) that maps metadata to filenames, so a bulk import is possible?
Are files delivered in batches as work completes, or in one drop at the end? Phased delivery lets you spot problems while they are still fixable.
How long does the provider retain a copy after handover, and how is it then deleted?

What is actually included in the price?

Without quoting figures, the line items to force into the open are: collection and transport, document preparation (staple and fastener removal — often 40–60% of total labour), scanning per image, indexing per field, OCR, re-assembly if originals are returned, media or portal delivery, storage or destruction of originals, and any minimum-volume charges. A rate that looks low per image while prep is billed separately at an open-ended hourly rate is the classic way a quote doubles. Our guide on comparing scanning quotes covers this in more depth.

Run a pilot before you commit

The most useful question of all is simply: “can we start with 10 boxes?” A small paid pilot using genuinely representative documents — including your most awkward paper — tests image quality, index accuracy, file naming, delivery format and turnaround before you commit the whole archive. Agree acceptance criteria in writing, review the output with the people who will search these files daily, then scale up. Providers confident in their process welcome a pilot; those who resist are telling you something useful. See our document scanning service overview for more.