A filing cabinet you cannot find anything in is obviously a problem. A hard drive you cannot find anything in looks tidy.
What actually improves
The honest list is shorter than the marketing one, and the real wins are worth being specific about.
| Claim | Reality |
|---|---|
| Faster retrieval | True — if documents are searchable |
| Less physical space | True, and usually the biggest win |
| Better backup | True — paper has no backup at all |
| Easier sharing | True |
| Environmentally better | Partly — storage and devices have a footprint |
| Cheaper | Eventually; scanning time is the cost |
| More secure | Neither more nor less; different risks |
That last row is worth resisting. Paper burns and can be stolen from a cabinet. Digital files can be copied silently at scale by anyone who gets access. Going paperless changes the risk profile rather than reducing it, and pretending otherwise leads to skipping the encryption and backup work.
Searchable, or it is not filed
This is the single decision that separates a working archive from a hoard.
A scanned PDF with no text layer cannot be found by searching. It is a picture in a folder, and finding it depends entirely on the filename and where you put it — which is exactly the constraint a filing cabinet had.
Run OCR on everything as it goes in. Every operating system indexes PDF text, so a searchable archive is findable by content: type "invoice Wickes 2024" and it appears regardless of where it sits. That is the actual benefit of going paperless, and it is entirely conditional on this step. Image to Text handles it, and the sequence — scan high, OCR, then compress — is in how to scan documents to PDF properly.
A naming convention beats a folder tree
Deep folder hierarchies feel organised and fail in practice, because every document could belong in two places and you will not remember which you chose.
A consistent filename does more work than any structure, because search matches on it and it sorts usefully.
- Start with the date, ISO format:
2026-03-14. It sorts chronologically as text, which no other format does. - Then the counterparty:
2026-03-14_Wickes. - Then the type:
2026-03-14_Wickes_invoice. - Then anything distinguishing: a reference number, an amount.
- No spaces, no "final". Underscores or hyphens. Version by date if you must version at all.
With that, a flat folder per year is genuinely sufficient for most personal and small-business archives. The convention is doing the filing.
Retention: decide before you scan
The temptation is to keep everything, since storage is cheap. That is how digital clutter accumulates — and unlike paper, nothing physically overflows to force a clear-out.
Different documents have genuinely different retention requirements, and they are worth looking up for your jurisdiction rather than guessing. Tax records commonly need several years; some employment and property records considerably longer; a great deal of everyday correspondence needs none at all.
Keeping personal data longer than necessary is also a compliance question under GDPR and similar regimes, so for a business "keep everything forever" is not a neutral default. Decide the policy first, then scan — deciding afterwards means reviewing thousands of files.
What to keep on paper
Not everything should be scanned and shredded, and the exceptions are specific.
- Anything requiring a wet signature to be valid — wills, some deeds, certain court documents. The categories are in are e-signatures legally binding.
- Original certificates — birth, marriage, qualifications. A scan is a convenience copy, not a replacement.
- Anything with a physical security feature — a seal, a watermark, an embossed stamp. Scanning discards precisely the thing that makes it authoritative.
- Documents an authority may demand in original form.
Scan these too, for retrieval and as insurance against loss. Just do not shred the paper afterwards.
A workable starting point
The common failure is starting with the back catalogue, which is thousands of documents, boring, and abandoned in week two with the archive half-migrated — the worst possible state, since now nothing is anywhere reliably.
- Start with new documents only. Everything arriving from today gets scanned, OCRd and named. The habit forms in a fortnight.
- Set up the backup before the archive matters. Two copies in different places, one of them offline. An archive with no backup is worse than paper, which at least does not fail all at once.
- Do the back catalogue by exception. When you need an old document, scan it then. Most of the cabinet will never be needed, and scanning it is work spent on nothing.
- Review after three months. If you cannot find things, the convention is wrong — fix it while there are hundreds of files rather than thousands.
Frequently asked questions
What is the most important step in going paperless?
Running OCR so documents are searchable. A scanned PDF with no text layer cannot be found by content — it is a picture whose retrieval depends entirely on its filename and location, which is the same constraint a filing cabinet had.
How should I name scanned documents?
ISO date first, then counterparty, then type: 2026-03-14_Wickes_invoice. The date format sorts chronologically as text, which no other format does. With a consistent convention a flat folder per year is usually enough.
Is going paperless more secure?
No — it changes the risks rather than reducing them. Paper can burn or be taken from a cabinet; digital files can be copied silently at scale by anyone with access. Treating it as automatically safer leads to skipping the encryption and backup work.
What should I not scan and shred?
Anything needing a wet signature to be valid, original certificates, and anything whose authority comes from a physical feature like a seal or embossed stamp — scanning discards exactly what makes it authoritative. Scan them for retrieval, keep the paper.
How long should I keep documents?
It varies by document and jurisdiction, and is worth looking up rather than guessing. Note that keeping personal data longer than necessary is a compliance question under GDPR and similar regimes, so "keep everything" is not a neutral default for a business.
Should I scan my whole filing cabinet first?
No — that is the usual reason these projects are abandoned half-done, which leaves nothing reliably anywhere. Start with new documents only, then scan old ones when you actually need them. Most of the cabinet will never be needed.