How to Count a Document’s Translation Volume Yourself

A four-page contract can contain less text than a single crowded page of terms and conditions, so counting the sheets won’t tell you how much translation you’re asking for. To estimate the volume, count the text and check how the provider defines a billable page.

The useful number is often characters with spaces, but that isn’t a universal billing rule. A quote might use a different character limit, characters without spaces, a word count, or a fixed rate for a standard document. The steps below help you calculate a preliminary estimate and ask the right questions before you send a file for pricing.

What “translation page” means¶

A translation page is a billing unit set by the service provider; it doesn’t necessarily match a physical page in the document. A printed A4 sheet might have a short heading and a few lines, or it might be packed with small type, tables, and footnotes. The paper size stays the same while the text volume changes.

In some pricing arrangements, the provider calculates text in standard units called accounting pages. One possible definition is a set number of characters, with or without spaces. Other arrangements use words, a standard-document price, or a different character limit. Don’t assume a number from one price list applies to another.

The first question to ask is: What exactly does one billable page mean in this quote? Ask whether the count includes spaces, whether it’s based on the original or the completed translation, and how partial pages are treated. A “page” with no definition isn’t enough information to reproduce the estimate.

If a provider defines one translation page as 1,800 characters including spaces, the calculation is:

Characters with spaces ÷ 1,800 = estimated translation pages

For example, a hypothetical text containing 5,400 characters would equal 3 pages under that definition. The arithmetic is simple; the important part is confirming that the provider actually uses that unit. Ask about rounding and any minimum order before turning the result into a budget.

Why spaces matter¶

A character count with spaces includes the gaps between words. A count without spaces leaves those gaps out, so the two totals differ. Microsoft’s Word statistics distinguish between character counts and counts that include spaces in its Word statistic reference.

That distinction matters when a price list says “characters” without explaining which count it means. If your file has 9,000 characters with spaces and 7,600 without them, the two totals don’t describe the same billing basis. Don’t compare your number with a quote until you’ve checked which statistic the quote uses.

Word count and character count are also different measures. A document with many short words can have a high word count without the same character total as a document made up of long technical terms. A provider who charges per word needs a word count; a provider who charges per character needs the matching character statistic.

Microsoft Support explains that “Word counts the number of words in a document while you type. Word also counts pages, paragraphs, lines, and characters.” Read the Word count instructions.

The distinction helps you choose the right figure: page count, word count, and character count answer different questions. A translation estimate needs the measure the provider uses, not the one that happens to be easiest to find.

The document’s page count isn’t a substitute¶

Changing margins, line spacing, or font size can make a document occupy more or fewer pages without changing its text. A one-page form may have a dense block of terms, while a ten-page file may contain large headings and whitespace. The number of sheets alone can’t show how much text needs translation.

Physical pages still matter for other reasons. A provider may need to inspect layout, tables, stamps, handwriting, or image quality. A scan may take extra checking even if it contains little readable text. Those details can affect the work, but they don’t turn the PDF’s page count into a reliable character count.

How to count text in Word and Google Docs¶

For an editable Word file, start by checking the text that actually needs translation. Then open the statistics and choose the count that matches the provider’s billing unit. Microsoft says the word and page counts appear in Word’s status bar, and selecting the word-count indicator opens more statistics; the exact interface can vary by version, as noted in Microsoft’s instructions.

1. Open the file in Word. Confirm that the text is editable rather than a picture pasted into a document. If you can place the cursor inside a paragraph and select individual words, the file contains text that Word can count.
2. Decide what’s in scope. Mark the paragraphs, tables, and other content that the translator should include. If you need only part of the file translated, don’t use the whole-document total as your estimate.
3. Open the word-count statistics. Look for the word-count figure in Word’s status bar and open the statistics. Find the character count with spaces or without spaces, depending on the provider’s rule.
4. Check the selected text, if needed. Highlight the section you want translated. Word can report a count for selected text as well as for the entire document, according to Microsoft Support.
5. Write down the number and the unit. Record, for example, “characters with spaces,” not just “characters.” That label makes it easier to compare the count with a quote.
6. Apply the provider’s formula. Divide by the stated character limit if the price is per character-based page. Ask whether a partial unit is rounded up, rounded to a fraction, or handled some other way.

The selection feature is useful when a file contains an appendix, a bilingual section, or an internal note that doesn’t need translation. Select only the text in scope and check the result. Before relying on that figure, confirm with the provider whether tables, text boxes, headers, or other areas are included in the final count.

If you can’t find the word-count indicator, check the menus for a word-count or statistics command. Interface labels and locations can differ between Word versions, so follow the instructions for the version you’re using rather than expecting every screen to look identical.

Counting only part of a Word file¶

A whole-file count can overstate the volume when only selected sections need translation. A contract, for instance, might contain a source-language section followed by an existing translation, or an appendix that the recipient doesn’t require. Highlight the text in scope before checking the count.

Keep a record of what you excluded. A brief note such as “main text only; appendix excluded” helps you explain why your count differs from a quote based on the entire file. The provider may include additional elements you missed, so ask how they handle document content outside the main body.

Word’s count is a starting point, not a guarantee that every item in the file has been captured. Headers and footers, comments, shapes, text boxes, and text inside images may need separate checking depending on the file and the software version. If the estimate seems unexpectedly low, inspect the document’s less obvious text areas.

Counting in Google Docs¶

Google Docs on a computer provides a word-count panel through Tools → Word count. The panel includes pages, words, characters, and characters without spaces, according to Google Docs Editors Help. Choose the figure that matches the provider’s billing rule, and label it clearly.

Google notes that its count excludes headers, footers, and footnotes unless you select a section of the document. The Google Docs instructions explain this behavior. If any of those areas need translation, check them separately instead of assuming they’re already in the total.

Google’s guidance says, “Unless you select a section of the document, word count applies to everything except headers, footers, and footnotes.” That means a document’s visible text can exceed the figure shown in the count panel when those areas contain material for translation.

Before you send the number, make a short checklist: main body, tables, headers, footers, footnotes, text boxes, and any content inside images. Mark the items that apply. That way, you can tell the provider what you included and ask whether their count will follow the same scope.

The Google Docs word-count panel is for the computer version covered by Google’s instructions. If you’re working in another interface or a different application, check that version’s documentation rather than relying on the same menu path.

How to estimate volume in a PDF, scan, or photo¶

A PDF can contain selectable text, or it can be a picture of a page. The difference determines whether you can count its text directly. Try selecting and copying a sentence: if you can select individual words, the PDF contains text that a word processor or other tool may be able to count. If you can select only a whole page or nothing at all, the file may be image-based.

A scanned page is an image, not text that a character counter can read. OCR, or optical character recognition, analyzes the image and creates a text layer. Adobe describes how OCR can make a PDF searchable and editable in its scan and OCR guide.

Adobe describes the result of its OCR process this way: “The file is now a searchable, editable PDF file.” In practical terms, the recognized words can become available for search and selection, which gives you text to check and count.

OCR output needs review. A low-quality image can cause a recognition tool to miss letters, confuse characters, or read a stamp incorrectly. A tilted photo, faint print, blur, shadows, and handwriting can all make it harder to recognize the text accurately. A character count based on faulty OCR can understate or distort the amount of work.

Use this sequence when the source is a scan or photo:

1. Check whether the text is already selectable. Try copying a sentence from the PDF. If the text copies correctly, start with that digital text rather than running OCR unnecessarily.
2. Run OCR if the file is image-based. Choose an OCR tool that can recognize the document’s language. The output should add a text layer that can be searched or copied.
3. Check the recognized text against the page. Look at headings, numbers, names, stamps, signatures, and small print. Fix obvious recognition errors before using the count.
4. Count the corrected text. Open the recognized text in Word or another editor with a character-count feature. Choose characters with spaces or without spaces according to the provider’s rule.
5. Tell the provider that the source is scanned. A scan may need visual checking even if you have produced a count. Share the file so the provider can assess readability and confirm the scope.

The quick copy test won’t prove that a PDF is perfectly accurate or complete. A file may contain selectable text while still omitting text embedded in images or stamps. Inspect the pages, and tell the provider about content that’s difficult to read.

What if the PDF has selectable text and images?¶

Some PDFs combine a text layer with pictures, stamps, or scanned pages. The copy test may work on the main paragraph but fail on a signature block, a handwritten note, or a photograph of an attachment. Count the text you can extract, then list the visual areas that still need review.

A partial OCR layer can create a misleadingly low total. If the first page is selectable but later pages are images, the count may cover only the first part of the file. Test more than one page, especially if the document’s format changes partway through.

The safest estimate separates what you can count from what you can’t. For example: “The selectable text contains X characters; two pages are image-based and need review.” That statement is more useful than presenting an incomplete count as the full document volume.

What if you have only a phone photo?¶

A phone photo needs the same basic check as a scanned PDF: can you select the text, or are you looking at a picture? If the photo isn’t a document with selectable text, OCR must recognize the words before a character counter can process them.

Take care with each page’s framing and legibility. If the photo cuts off a line or blurs a stamp, OCR can’t reliably recover the missing information. A clean text count doesn’t solve a source-quality problem; the provider may need a clearer image or a manual review.

For several photos, keep the image order clear and check that every page is present. A missing page can make the total look lower even when the OCR is working correctly. Include a note about any unclear sections rather than guessing what they say.

Turn the count into a useful estimate¶

A character count helps only after you know the provider’s billing basis. The calculation below works for a character-based page; it doesn’t convert a word-based price or a fixed-document fee into the same unit.

Estimated pages = characters counted ÷ characters per page

If a provider sets the page at 1,800 characters with spaces, a hypothetical count of 7,200 characters gives 4 pages before any rounding rule. If the provider uses a different unit, substitute that unit’s limit. Don’t round the result yourself unless the price list or provider explains how partial pages are charged.

What the provider charges for	What to count	What to confirm
Characters with spaces	Character count with spaces	Whether the estimate uses the original or completed translation
Characters without spaces	Character count without spaces	Whether punctuation and numbers are included
Words	Word count	How tables, headings, and fragments are handled
A standard document	The relevant document type and version	Which fields or pages the fixed price covers
A physical page or another unit	The item specified in the quote	What qualifies as one billable unit

A fixed price for a standard document can be easier to estimate than a character count, but it applies only if your document fits the provider’s definition. A passport, certificate, diploma, or other template-based record may be priced as a document rather than by its measured character total. Ask whether the price covers every page and whether additions, handwritten entries, or attachments change the estimate.

Ask how the original and translation are counted¶

A count of the original can be a useful first estimate, but different languages may express the same content with different numbers of characters. The translation can end up longer or shorter than the source, depending on the language pair and text. A discussion of this difference appears in EVEREST’s explanation.

Ask whether the final bill is based on the original file or on the completed translation. If the provider calculates the final amount from the translated text, the source count can’t guarantee the final volume. Don’t apply a made-up expansion factor: the exact target text hasn’t been produced yet.

A helpful quote explains what happens after translation. Ask whether the provider will confirm a final count, what unit they’ll use, and how any difference from the initial estimate affects the bill. Those details matter more than a precise calculation based on the wrong billing rule.

Ask about rounding and minimum charges¶

A calculated volume may include a fraction of a page. Providers can handle partial units differently, and an order may have a minimum charge. Neither rule can be derived from the character count itself. Ask for the rounding method and minimum before comparing estimates.

For instance, an arithmetic result of 2.4 pages doesn’t tell you whether the invoice will use 2.4 pages, round up to 3, or apply another rule. A quote that includes a formula and a rounding policy is easier to check than a total with no explanation.

Separate translation from other services in your questions. Certification, notarization, layout work, or other requested services aren’t part of a character count. Ask whether they’re included in the quote or calculated separately, so you don’t mistake a service charge for extra translation volume.

A practical message you can send with your file¶

A clear request helps the provider assess the same material you counted. Include the file, the scope, the statistic you used, and the parts you’re unsure about. You can adapt this template:

“Please estimate translation of the attached file. I counted [number] characters [with/without spaces] in [Word/Google Docs/OCR output]. Please confirm the billing unit, whether you count the original or completed translation, how you handle partial units, and whether headers, footnotes, tables, and image-based text are included. The file contains [scan/photo/unclear section, if relevant].”

That message doesn’t assume that your count is the one the provider will use. It gives them a starting figure and asks them to confirm the rules. If you counted only a selected passage, name the excluded parts so both sides are estimating the same work.

Common counting mistakes and how to avoid them¶

A count can be mathematically correct and still be a poor estimate. The usual problem isn’t division; it’s a mismatch between what you counted and what the provider includes. Check the source, the scope, and the billing unit before drawing conclusions from the total.

Mistaking physical sheets for translation pages¶

A file’s page count reflects its layout, not just its text. Font size, spacing, tables, images, and margins all affect how much content fits on a sheet. Count the text or ask for a document-specific quote instead of multiplying the number of sheets by a guessed rate.

This matters especially with certificates and forms. A short form might contain little text but still require attention to stamps, labels, and layout. A dense report may fill many pages with text. Page count can help describe the file, but it can’t replace the provider’s billing unit.

Counting the wrong character statistic¶

A number without a label is hard to verify. “7,000 characters” could mean with spaces or without spaces, and it may come from a different selection than the one the provider priced. Save the exact statistic name and the scope alongside the result.

Word’s reference documentation distinguishes character statistics, and Google Docs’ word-count instructions describe counts with and without spaces. Choose the matching measure rather than assuming that all software reports the same field under the same label.

Forgetting text outside the main body¶

Headers, footers, footnotes, text boxes, tables, captions, comments, and images can all contain words that a reader expects to be translated. A word-count panel may not include every one of those areas. Google Docs explicitly excludes headers, footers, and footnotes unless a section is selected, according to its help page.

Inspect the document visually, then compare what you see with the counted text. If the provider’s quote includes material you left out, the final volume will be higher. If you don’t need an appendix or an existing translation included, say so before the provider estimates the work.

Treating OCR as ground truth¶

OCR creates text from pixels, but the result can contain recognition mistakes. A lower-case letter can be read as a number, a name can be misspelled, and a faint line can disappear. The resulting count may look precise while representing inaccurate text.

Review the recognized output against the scan, especially names, dates, figures, small print, and stamps. If the image is too unclear to verify, tell the provider. A manual estimate or a better source file may be more useful than a number built on uncertain recognition.

Assuming source and translation counts will match¶

The original text and the translated text don’t have to contain the same number of characters. The language pair and the wording affect the result. EVEREST’s explanation notes that character counts can change between the original and the translation.

Treat the source count as an estimate if the final amount depends on the completed translation. Ask how the provider will settle the final volume instead of adding an arbitrary percentage to your calculation.

Comparing quotes with different units¶

One price per page and another price per word can’t be compared directly without the relevant counts and definitions. Even two prices described as “per page” may use different character limits or rounding rules. Ask each provider to state the unit and apply it to the same file and scope.

A comparison should also separate the translation itself from additional work. A quote may include layout, certification, or another service that a second quote excludes. Confirm what each total covers before deciding which estimate is lower.

A quote is easier to understand when it names the file, the requested scope, the billing unit, the counting basis, and any separate services. If any of those details are missing, ask rather than filling in the blanks yourself.

FAQ¶

How do I count translation pages in Word?¶

Open Word’s word-count statistics and find the character count with spaces or without spaces, as required by the provider. Divide that figure by the provider’s stated characters per page, then ask how partial units are rounded.

How many characters with spaces are in one translation page?¶

There’s no universal definition. A provider might set a page by characters with spaces, characters without spaces, or another measure, so check the price list or quote before calculating.

How do I count a scanned PDF or document photo?¶

First check whether the text is selectable. If the file is image-based, run OCR to create text that can be counted, then compare the recognized words with the page because image quality can affect accuracy.

Why can one A4 page count as several translation pages?¶

A physical sheet can contain very different amounts of text. Translation pages are based on the provider’s billing unit, not on paper size, so a dense page may contain multiple units of text.

Why is the translation longer or shorter than the original?¶

Languages can express the same information with different numbers of characters. If the final bill is based on the translated text, the original count is an estimate rather than the final volume.