Scanning a stack of papers produces a folder of images. Whether that folder is useful depends entirely on decisions made before you start — principally whether the text inside is searchable and whether the filenames mean anything.
Get those right and you have an archive. Get them wrong and you have a digital version of the drawer, with the same problem.
What changed in 2026
- Phone scanning became genuinely good. Automatic edge detection, perspective correction, and text recognition made dedicated scanners unnecessary for most documents.
- Text recognition became standard. Searchable output became the default rather than an extra step.
- Cloud storage search improved. Finding documents by their content became reliable across mainstream services.
- Original requirements persisted. Certain documents continued to require the physical original.
Make it searchable
The decision that determines whether the archive works.
A scan without text recognition is a picture. You can look at it and you cannot search for it. An archive of a thousand unsearchable images is functionally a filing cabinet you have to open one drawer at a time.
Text recognition — converting the image of text into actual text embedded in the file — makes the whole archive searchable by content. You search for a supplier name and find the invoice without knowing where you filed it.
Most modern scanning apps do this automatically and produce a searchable PDF. Confirm yours does, because it is the difference between an archive and a pile.
Where recognition struggles — handwriting, poor quality originals, unusual layouts — a descriptive filename compensates.
Name files consistently
The second decision, and simple rules work best.
Date first, in a sortable format. Year, then month, then day, so files sort chronologically by name automatically.
Then what it is. The organisation, the document type, and anything distinguishing.
That gives you names that sort usefully and read clearly, which handles the cases where text search does not.
The corollary: do not build an elaborate folder hierarchy. Nested folders require you to remember where you put things and require maintenance nobody sustains. A flat structure with a handful of broad folders, combined with searchable text and consistent names, is more findable and survives neglect.
Equipment
A phone scanning app handles the overwhelming majority of documents. Automatic detection and correction, text recognition, direct save to cloud storage. Fast enough that you will actually use it, which is the property that matters.
A flatbed scanner is better for fragile originals, photographs, and anything where quality matters more than speed.
A sheet-fed scanner earns its place only for a large backlog. If you have a filing cabinet to digitise, it saves real time; for ongoing use it mostly sits unused.
Start with the phone. Buy hardware only if a specific limitation makes you.
What to keep physically
Scanning is not a licence to shred everything.
Documents typically requiring the original include certain identity documents, some legal instruments where a signed original is needed, vehicle titles, and some property documents. Requirements vary by jurisdiction and by what you might need them for.
The workable rule: scan everything, keep originals where an original might be required, and shred the rest — securely, for anything with personal or financial information, per shredding schedule.
Originals worth keeping belong somewhere protected and known — see document vault.
And back the archive up. A digital archive on one device is one failure from gone — see photo backup strategy, which applies identically to documents.
Common mistakes
- Scans without text recognition. Unsearchable.
- Inconsistent naming. Sorting and finding become guesswork.
- Deep folder hierarchies. Maintenance nobody sustains.
- Shredding originals that are required. Some cannot be replaced.
- No backup of the archive. One failure from gone.
- Scanning everything at once. The backlog never gets finished; do it incrementally.
- Not scanning at the point of receipt. The backlog accumulates.
FAQ
What resolution should I use?
Enough that text is clearly legible, which for ordinary documents is modest. Higher resolution produces larger files with no benefit for text. Photographs and anything with fine detail warrant more.
PDF or image files?
PDF for documents, particularly multi-page ones, with embedded searchable text. Images for photographs.
Should I scan old records?
Selectively. Scanning everything historical is a large project that frequently stalls. Scanning what you might actually need, and going forward from now, is more achievable.
How do I handle receipts?
Photograph at the point of purchase rather than accumulating a pile to scan later — see receipt organization.
Where to go next
For where the archive and the originals belong, read document vault. For disposing of what you scanned, shredding schedule, and for keeping the archive safe, photo backup strategy.