At 10:37 p.m., Maya stared at two quotes for the same project and felt the peculiar dizziness of trying to compare apples to origami cranes. One agency promised a neat price per word; the other offered a steady per-page figure that made her spreadsheet sigh with relief. Her problem was painfully common: she had a firm budget and a deadline, yet no clear way to decide which quote honored both. Her desire was simple, too. She wanted a way to see the logic behind the numbers rather than guess at it. The promise of value in this story is the map itself: once you know how language-service agencies build per-word and per-page rates, you can choose the model that fits your document, manage costs without quality surprises, and negotiate like you know what happens behind the curtain.
She remembered the first time she bought language services: a pamphlet about an eco-friendly product line. The vendor sent a friendly PDF with a price per word and a handful of mysterious discounts for “repetitions” and “fuzzies.” Tonight, though, her files were messier: a scanned legal letter with stamps, a typed memo, and an image-heavy brochure. One agency said, We price by the word when we can count; by the page when we need to contain risk. The other said, Send editable files and we’ll save you money. That was the pivot. This post is the conversation Maya needed—how the math works, why it changes, and how to use it to your advantage.
Why agencies lean on per-word pricing: predictable effort in, predictable quality out. The per-word model exists because counting units of text gives both sides a shared yardstick for effort. But a “word” isn’t as simple as it seems. Does a date count? What about hyphenated compounds, numbers, or a string of product codes? Agencies typically rely on software that scans editable files and produces a consistent count. Those tools also categorize segments: exact repeats, near repeats, and brand-new text. The reason is practical. If Maya’s brochure reuses the same tagline fifty times, paying full price fifty times feels unfair. Weighted word counts solve this. For example, an agency might bill new words at 100% of the base rate, near repeats at 60%, and exact repeats at 10% to cover checking and consistency.
This pricing grid sits on top of a base influenced by language pair, subject complexity, and turnaround speed. A rare language or a highly specialized medical paper pushes the base up; a mainstream pair and a general audience push it down. A simple calculation illustrates the mechanics. Suppose the draft has 2,450 counted words: 1,600 new words, 500 near repeats, and 350 exact repeats. With a base of $0.12 per word, a 60% rate for near repeats ($0.072), and 10% for exact repeats ($0.012), the weighted total might be $192 + $36 + $4.20 = $232.20. Add a modest project management component, say 10%, and you have a transparent $255.42. Subtract a rush fee and page layout, and you have the spine of a fair quote.
Per-word pricing shines when files are editable and content is text-heavy, but it falters with dense imagery, scans, and design files that need extra hands. That’s where the next model enters.
Per-page pricing arrives when the word count lies, disappears, or hides inside images. Imagine Maya’s stamped legal letter, a bank statement with tables, or a birth record captured as a low-resolution scan. The software that powers per-word pricing either miscounts or cannot count at all. Suddenly the per-word model becomes guesswork, and guesswork is expensive. Agencies containerize this risk by pricing per page.
What does “per page” really mean? It can mean one of three things. First, a literal page: every PDF page—text-heavy or nearly blank—has a fixed price because the document requires certification, layout checks, or manual typing regardless of word density. Second, a “standard page,” commonly defined internally as 250 to 300 words of clean, editable text; multiple standard pages may live on a single PDF page. Third, a hybrid: per page for administrative rigors (stamps, seals, formatting, notarization), and a reduced per-word component for the main text extracted from editable sections. This hybrid is remarkably common for court bundles and academic packets.
The per-page model also bundles invisible effort. Someone must reconstruct tables, align signatures, preserve seals, and attach statements of accuracy. That is why official documents often come with a fixed per-page quote that includes layout and certification, sometimes labeled as certified translation. Here’s a concrete picture. Maya sends a three-page bank statement as scanned images. An agency might quote $45 per page, all-inclusive, because a specialist will transcribe numbers carefully, a reviewer will check alignment and totals, and a coordinator will ensure the final PDF mirrors the original, down to the footer. Another provider may define a standard page as 300 words and apply it only to the readable memo, billing the scans by actual PDF page. Neither is “more correct”; each is a way of telling you where effort hides and who carries the risk.
Per-page pricing is not a blunt instrument when used wisely. The best providers are explicit: they show the page definition, note what is included (layout, stamps, reviewer pass), and flag what is not (extensive desktop publishing, complex illustrations, or bilingual tables that require manual rebuilding). Once you see those levers, you can choose the model that maps to your document’s reality.
Turn your next quote into a controllable experiment by preparing files, asking for the right reports, and picking the model that fits the document’s physics. Start with file hygiene. Editable files lower costs because machines count well and linguists focus on language rather than transcription. If you only have scans, provide the clearest version and group similar documents together; batching reduces handling overhead.
Next, request visibility. Ask for a weighted word-count report when per-word pricing is proposed; it should list new content, near repeats, and exact repeats with their respective percentages. If per-page pricing is offered, ask for the page definition and what tasks are bundled—layout, review passes, and any administrative statements. In mixed projects, request a hybrid: per page for stamped or image-only sections, per word for the body text.
Then negotiate scope rather than numbers. Provide reference materials, past campaigns, or style guides, which improve consistency and reduce review time. If the schedule is flexible, consider staged delivery: urgent pages first, the remainder at a non-rush pace. Clarify quality tiers: a marketing tagline may deserve senior attention, while internal memos can use a standard pass. Ask whether prior work can be leveraged to reduce near repeats; even if you are new to the agency, they may create a memory of approved segments during the first project, lowering costs in future rounds.
Finally, align the model to the job. Dense legal or financial packets with seals and tables often make more sense per page; web copy, catalogs, and manuals usually favor per-word pricing with repetition discounts. Creative work that involves heavy rewriting may benefit from hourly or per-piece pricing. The key is to anchor each quote to the task that governs real effort—countable words for pure text, countable pages for administrative and layout burden.
Understanding how the numbers are built gives you more than a cheaper bill; it gives you control over outcomes. Per-word pricing rewards clarity and reuse, while per-page pricing protects both sides from the chaos of scans, stamps, and layout. The trick is not to pick a camp, but to read your documents as a project manager would, and then choose the model that aligns with the work.
Here is the takeaway for Maya and for you: when content is editable and repetitive, the word-based model with weighted discounts is your friend; when files are image-first or wrapped in official formatting, page-based pricing keeps risk contained. Ask for definitions and reports, separate the text from the admin tasks, and do not be afraid to request a hybrid when the document pile says you need both. If this helped demystify the quotes in your inbox, share a note about the document type that confuses you most and the model you chose. Your experience might be the story that helps the next person open a quote at 10:37 p.m. and feel calm instead of dizzy. Apply the checklist to your next request, and tell us what changed in your results and your budget. If you need assistance, consider reaching out to a translator who can help clarify the process for you.
