Clicky

Fillfront
← All guides

How to turn a paper or scanned form into an online form

Scan the form once, upload the scan, and put a field wherever an answer goes. Fillfront picks the lines and boxes out of the image and reads the printed captions to name them, so you mostly check and correct. People fill it in on a phone, and what comes back is the same form with their answers printed on it. A straight, clean scan makes all of that much easier.

#What happens to a scan when you upload it

A scanned page is a photograph of paper. There are no words in the file, no lines, no boxes. Just pixels.

So we look at the pixels. Detection finds the things that look like places to write: ruled lines, drawn boxes, tick boxes, rows of dots. Then it reads the printed text around each one (that's the OCR part) and names the field after its caption. If the form isn't in English, it works out the language from that first read and, for the languages we support, reads the page again in that language. On a less common language, expect to type more of the field names yourself.

Diagram: on the left a scanned form with captions and blank lines; on the right the same form with green fields over each line, named Full name, Date of birth, Phone and Signature Lines and boxes come from the image; names come from the printed words beside them.

It works well on a clean scan. It's still reading a photograph, so expect to do more by hand than on a PDF someone made in Word. A field may arrive unnamed, or named after a stray smudge. Some faint lines won't be found at all.

What the checker will say. Run a scan through the free fillable PDF checker and it reports no form fields. That's the expected answer for a scan, not a problem with your file. Why can't I type in this PDF? explains how a scan differs from a flat PDF.

#How to scan a form so it works well

Most of the trouble we see with scanned forms traces back to the scan itself. Two minutes here saves twenty in the builder.

Lay it straight. Sideways and upside-down pages are fine; we turn those upright on upload, and there's a turn button in the builder if we get it wrong. A page that's tilted by a degree or two is the real problem. The fields you place are straight boxes, and the printed lines underneath them aren't, so answers drift off their lines towards one end.

Use greyscale or colour. Pure black-and-white scanning turns light grey boxes and shaded rows into nothing. If the scanner can't see a line, neither can we.

Get the contrast right. Detection tells ink from paper by how much darker it is. Pencil, a tired photocopy, or a scan with the brightness pushed up all leave lines too pale to find. White lines printed inside a dark band (common on order forms) are a known blind spot.

200 to 300 dpi. That's plenty for reading captions and for the final PDF to look sharp. Higher just makes a bigger file.

Start from the cleanest copy you have. A photocopy of a photocopy picks up specks, grey patches and broken lines. This scan will be the master copy for years, so it's worth hunting down a good original.

Diagram of four scan problems: a tilted page where answers drift off the lines, a pale line that isn't found, a white line inside a dark band that isn't found, and a dotted line with a gap that becomes two fields The four things we see most often on scanned forms that come out with missing or misplaced fields.

#Photographing it with a phone

That's fine, as long as you use a scanning mode rather than the camera. The iPhone Notes app has a document scanner, and on a Mac, Continuity Camera can scan straight into Preview. Both find the edges of the sheet and flatten the perspective. A plain photo keeps the angle you held the phone at, and that's the tilt problem again.

We see quite a few uploads that are exactly this: a phone scan, saved through Preview, one image per page. They work. Do it in good daylight, on a dark table, so the edges of the paper stand out.

#Why does my Canva PDF behave like a scan?

We see this one a lot. Of the real forms we test detection against, Canva files make up about a third, and over a third of those had no text in them at all. Nothing could be selected. They looked crisp and digital, and behaved exactly like a scan.

The likely cause is a checkbox. In Canva, Share → Download → PDF Print offers Flatten PDF, which turns every page into a single image.

The 10-second test. Open the PDF and try to drag across a heading. If nothing highlights, the page is an image, whatever app made it.

If you still have the Canva design, download it again with Flatten PDF unticked. Captions then come from real text rather than OCR, so the fields arrive with the right names more often. If you only have the flattened file, treat it like any other scan.

#Placing and checking the fields

Upload the scan and work down the field list. For each field:

  1. Check the name. The name is what your recipient reads on their phone, so "Field 7" or a string of odd characters needs renaming to something like "Emergency contact number".
  2. Check the type. Date, number, email, a choice from a list, a signature. Pick the one that matches the question.
  3. Check the size. Make each box a little more generous than the printed line, so a long name has room.

Then draw anything that was missed. Work top to bottom so the online form follows the order of the paper one.

A dotted line with a gap in it sometimes comes through as two or three short fields. Delete the extras and stretch one across the whole line.

#What about a multi-page form?

Scan every page into one PDF, in paper order. A four-page enrolment pack then becomes one link and comes back as one four-page PDF. Splitting it into four forms gets you four submissions to match up by hand, which is the problem you started with.

#Test it before you send it

Fill in your own link with obviously fake values. A row of X's in every text box is better than your real name, because it shows where a long answer would run out of room. Look at the PDF that comes back and nudge anything sitting high or low. Do it once, and the next two hundred come out right.

#Why keep the scan instead of redesigning the form?

Because people who use a form often don't read it. They navigate it by shape.

The volunteer who checks waivers at the gate knows the date is bottom left. The office that processes your enrolment forms knows which box to look at first. Redesign it and everyone who knew the old sheet is a beginner again.

There's also only one version. The paper copy at the front desk and the link on the poster are the same document with the same wording, so nobody has to work out which clause applied to which response.

Sometimes a clean rebuild is the better call, usually when the paper form was badly designed to begin with. If yours was fine, keep it.

#What your recipients see, and what comes back

They see a web page with your labels, one question after another, sized for their screen. Nobody pinches and zooms around a photo of a sheet of paper.

When they submit, their answers are printed onto the scan at the positions you marked. You get a PDF that looks like the form you have always filed, with the boxes written in. One per response, in one place, so a Saturday's worth of permission slips ends up as a list of PDFs rather than a folder of phone photos. How to collect form responses as completed PDFs covers what to do with them next, including the spreadsheet export.

Most organisations keep the paper version going for a while too. That's fine, since it's the same form. File the paper copies as you always have and the PDFs alongside them. How to make a PDF fillable without Acrobat covers the same process for forms that were never on paper.

#No form to scan?

If the process has always been done by hand and there's no sheet to scan, don't make one just to have an image. Start from a form that already exists.

The schools and childcare templates cover the paperwork that goes home in a book bag and comes back late, including enrolment, field trip permission slips and emergency contacts. The fields are already placed, so you can send a link the same afternoon.

Common questions

Do I have to rebuild my form from scratch?

No. Scan it once and upload the scan. The fields go on top of the image, so the form keeps the layout everyone already knows, and the completed PDF looks like the sheet you have always filed.

Does a scan have any fields to detect?

Not in the file. A scan is a picture, so there is nothing inside it to read. Fillfront looks at the picture instead, finding the lines, boxes and tick boxes, and reads the printed captions beside them to name each field. Expect to draw and rename more than you would on a PDF made in an app.

What resolution should I scan at?

200 to 300 dpi is plenty. Below that, small print gets hard to read. Above it the file just gets bigger. Use greyscale or colour rather than pure black and white if the form has any shading.

Do I need to run OCR on the scan first?

No. Fillfront reads the printed words itself to name the fields. Running OCR beforehand makes the file searchable, which is nice for an archive, but it doesn't change where answers go.

Can I photograph the form instead of scanning it?

Yes. Use a scanning mode rather than the plain camera, such as the document scanner in the iPhone Notes app or a phone scanning app, so the edges are found and the page is flattened. Save it as one PDF with the pages in order.

My PDF came from Canva, so why is it behaving like a scan?

Canva's PDF Print download has a Flatten PDF option that turns the whole page into one image. The result looks crisp but contains no text, so it behaves exactly like a scan. If you have the design, download it again with Flatten PDF unticked.

What comes back when someone fills it in?

The same scanned form with their answers printed onto it, as a PDF. One per response.

Keep reading

Templates for this

Free tools

Last updated