Scanned form to a digital form

Pages become sections

Each page of a scanned form usually holds one topic, so pages come across as section headings with their fields underneath. That is the structure you want anyway, because it is what lets you split the form into steps afterwards. If a page carries two unrelated topics, say so in the prompt and it will be split.

Repeated blocks become one field

Paper handles repetition by printing the same block eight times: eight rows for eight items, or eight identical boxes for eight children. Online you want one table field with the right columns and an Add row button. Name that explicitly in the prompt, because a scan of eight identical blocks looks like eight separate questions unless you say otherwise.

What scan quality does

A 300 dpi flatbed scan of a printed original reads almost perfectly. A photocopy of a photocopy loses thin rules and small print, which is where the checkbox columns and the footer live. A skewed feed through a document scanner is the worst case, because rows drift across columns. If your scan is poor and the original still exists, rescan rather than editing 40 fields by hand.

Then cut it down

A five page paper form is almost never a five page web form. Once it is rebuilt, look for fields that only existed because paper cannot calculate, cannot branch and cannot validate. Serial numbers written twice, totals filled in by hand, and "office use only" blocks usually come straight out.

How it works, in three steps

1. Step 1

   Scan the whole form as one PDF

   One multi page PDF is easier than five images, and it keeps the page order unambiguous.

2. Step 2

   Describe the repetition

   Name the repeated blocks and the columns they should become. This is the single most useful sentence in the prompt for a scanned form.

3. Step 3

   Delete what paper forced on you

   Remove duplicate entries, hand totals and office use blocks. The online form does not need them.

The full walkthrough with screenshots is in the guide Convert a paper form into a web form.

Limits worth knowing

• A scan is read as an image, so small print, thin rules and skew all cost accuracy.
• Attaching files is a paid plan feature, five files at 10 MB each.
• Handwritten content is not read as data.
• Repeated blocks are only merged into a table field if the prompt asks for it.

See it on a finished piece

Patient intake form

25 fields

Rental application form

27 fields

Product order form

16 fields

Questions people ask

Related pages