How do I scan a document into Excel?

No, I can’t. And it’s not because I don’t want to. What you’re asking for is
very very difficult to do well, if at all. It all has to do with
exactly what it means to scan something, and what it means to be an Excel
spreadsheet.

The most important thing to realize is that a scan of a document is almost
exactly the same as taking a picture of the document. So what you end up with
is the equivalent of a “.jpg” picture of the document – even if the scanning
software doesn’t actually save it before performing other tasks.

A representation of a ‘picture’ of a light green letter ‘A’

A digital picture of a letter is just a collection of individual pixels that
are turned on in the appropriate color in the pattern of the original
letter.

The second most important thing to realize is that applications such as Word
or Excel don’t operate on pictures. They operate on documents in specific
formats and with data arranged in specific ways. So rather than operating on a
picture of the letters ABCD, Word operates on an internal representation of the
letter A, the letter B and so on. This way Word, and applications like it, make
it easy to edit – say replacing or deleting one of the letters, and
reformatting the entire document if needed to reflect the effects of that
change.

In Word and Excel and countless other applications, the letter “A” is
represented by the number 65.

So the first problem is converting those pictures of letters into the
computer’s internal representation of the equivalent letter.

Software to solve this problem exists, and is called “optical character
recognition” or OCR software. These programs, when given a picture of a page
with text, will analyze the image and produce a document with the program’s
best guess as to what those letters are and how they were organized. Many
scanners nowadays actually come with basic OCR software included.

The problem with OCR is that it’s never 100% accurate. In fact, it’s often
far less than that meaning that any OCR’d document will still required a fair
amount of cleanup once it’s been converted. OCR also has a particularly
difficult time with formatting, which can confuse it, or can just be lost.

And, of course, the quality of OCR is also dependent on the quality of the
original document scan.

But, after all’s said and done, OCR is often a good first step to converting
a scanned document back into some kind of editable form.

But what about spreadsheets like Excel?

The problem with spreadsheets is that they contain more than just text. The
rows and columns represent structure. OCR software is typically not
able to determine the structure of a document from just a scanned image. Some
do try, as in the case of trying to determine paragraphs in a more traditional
text document.

Unfortunately, scanning and converting the letters and numbers is one thing,
but understanding just how that data should be placed into which cells on a
spreadsheet is beyond the ability of most OCR software. In fact, I’m not
currently aware of any that can do so. (Check the comments to this article, as
I’m sure readers will add suggestions.)

Even as OCR technology continues to improve, I’d fully expect there to be
spreadsheet based documents that even the best software would never be able to
“understand”.

My recommendation today is to OCR the document into a the software’s best
guess at a regular word processing document, and then spend the time to create
your new spreadsheet by hand, copy/pasting the data in the ways that you know
the spreadsheet should be organized. It is work, but you’ll get the spreadsheet
you expect.