MTPE Pricing Calculator: Building Your Agency Rate Card¶

You just won a 500K-word technical documentation project from a Fortune 500 client. They’re excited, budget is tight, and they ask: “Can you do this as MTPE? What’s your rate?” You freeze. You don’t have an MTPE rate card. You don’t know if $0.05 per word is competitive or if you’ll lose money on every thousand words. You Google frantically, find three different pricing models, and guess. That’s not a business—that’s hope.

This guide walks you through building a defensible, profitable MTPE rate card from first principles. We’ll start with what the market actually pays, then move to the formulas that turn cost into price, and finally to the edge cases that trip up agencies most often. By the end, you’ll have a calculator and a strategy.

Understanding MTPE Pricing Tiers: Light PE vs Full PE¶

Before you set a single rate, you need to understand what you’re actually selling. ISO 18587 defines machine translation post-editing as “linguistic and technical assessment, editing and correction of machine-generated target language text,” but the standard splits this into two fundamentally different service tiers.

Full Post-Editing (FPE) aims for publication-ready quality—text that’s indistinguishable from human translation. Grammar, tone, terminology, cultural references, all corrected. ISO 18587 requires the final output to reach a quality level comparable to professional human translation, which means no shortcuts. If a machine translator writes “The company is founded in 1995 and it is very successful,” a full PE editor fixes the awkward pronoun, considers style, and ensures consistency with your client’s brand voice.

Light Post-Editing (LPE) targets comprehensibility. Major meaning errors gone, spelling fixed, sentences understandable—but you’re not polishing for publication. You’re ensuring the text doesn’t confuse the reader. Missing articles, occasional typos, less-than-perfect style: acceptable. LPE is the choice for internal technical documentation, process manuals, or anything where clarity matters more than elegance.

The pricing difference is stark:

A client asking for “MTPE” without specifying LPE or FPE is asking for ambiguity. Your first conversation should clarify which one you’re quoting.

The Market Rate Baseline: What Agencies Charge in 2026¶

If you’re building a rate card from scratch, you need to know where the market sits. The data below reflects 2025-2026 pricing from active translation agencies and service providers across North America and Europe.

Global MTPE Baseline¶

Per-word MTPE rates range from $0.05 to $0.15, with most agencies clustering around $0.08-$0.12 for full PE. This is roughly 50-70% of human translation rates, which typically range $0.15-$0.30 per word depending on language pair and complexity.

Light post-editing sits at $0.03-$0.08 per word, often undercut by high-volume MTPE specialists. If you’re pricing LPE at $0.08, you’re at the premium end and better have the quality to justify it.

Language Pair Variation¶

These rates are not one-size-fits-all. Common pairs (EN-DE, EN-FR, EN-ES) command lower rates because linguist supply is high and competition is fierce. Rare pairs shift dramatically:

If your core pair is EN-Finnish, charging $0.08 is insulting—the talent is rare. If it’s EN-German, $0.08 is fair to competitive.

Regional Rate Differences¶

European agencies quote in EUR and charge higher per-word rates due to labor costs and operating expenses:

• US Agencies: $0.05-$0.12/word (full PE), lower overhead, often outsource post-editors to lower-cost regions
• EU Agencies: €0.08-€0.14/word (full PE), higher local wages, more stringent labor regulations
• Boutique/Specialized: $0.12-$0.20/word (full PE) for legal, pharmaceutical, financial sectors with strict QA

Hourly Rates¶

Some agencies abandon per-word pricing for hourly, especially when MT quality is unpredictable. Hourly rates typically range $35-$85 per editor hour, depending on experience and geography. We’ll cover the hourly-to-per-word conversion later.

How Fast Is MTPE, Really? Productivity Metrics¶

Here’s where many agencies blow their pricing: they guess at productivity. A post-editor is not 10% faster than a translator or 50% faster—they’re often 2.8-4x faster, but only under specific conditions.

Professional translators working on financial texts delivered 700 words per hour when post-editing, compared to their normal 250 words per hour on full translation. For light MTPE projects, the expectation is somewhere around 600-1,000 words per hour depending on language pair and domain.

The variance is crucial:

Full Post-Editing Productivity - Financial/technical (well-formed MT): 650-850 wph - Marketing/legal (trickier MT): 500-650 wph - Poorly sourced MT engine: 300-500 wph (editing time approaches human translation)

Light Post-Editing Productivity - High-quality MT output: 1,000-1,200 wph - Medium-quality output: 750-1,000 wph - Low-quality output: 500-750 wph

The productivity multiplier vs human translation is real: a single translator working without MT can deliver 2,000-3,000 words daily; with MTPE, output rises to 7,000 words per day. That’s a 2.3-3.5x boost.

But here’s the catch: this only holds when MT is trained well, terminology is consistent, and the post-editor isn’t second-guessing every decision. Bad MT—generic engines trained on news data translating medical documentation—can drag productivity to 400 wph, wiping out your cost savings entirely.

Action step: Before quoting a new language pair or domain, ask the client: What MT engine? Do you have a trained model? How’s the quality? If they say “we just run Google Translate and hope,” either increase your PE rate by 30-50% or decline the project.

Building Your Rate Card: Step-by-Step Formula¶

Now we build. Here are four practical methods to calculate your MTPE rates, from simplest to most data-intensive.

Method 1: Percentage Discount Off Human Translation (Quickest)¶

The easiest approach: take your standard human translation rate and apply a discount.

Formula:

MTPE Rate = (Human Translation Rate) × (1 − Discount %)

Example: - Your human translation rate: $0.20/word - Desired discount: 50% (because post-editors are faster) - MTPE rate: $0.20 × (1 − 0.50) = $0.10/word

This works if you have established human translation rates and your local market is transparent. The downside: it assumes your discount is right, which it often isn’t. A 50% discount for light PE might be fair; a 50% discount for full PE of poorly sourced MT is giving away margin.

When to use: You have an existing rate card and need a quick MTPE anchor. Use 30-40% discount for full PE, 50-65% for light PE as starting points, then adjust.

Method 2: Cost-Plus Markup (Most Common)¶

This is the standard in translation: calculate your cost, add desired margin, and price accordingly.

Formula:

MTPE Rate = Cost per Word ÷ (1 − Desired Margin %)

Example: - Post-editor cost: $30/hour - Productivity: 800 wph - Cost per word: $30 ÷ 800 = $0.0375/word - Your overhead per word: $0.01 (systems, management, QA, 15% of editing cost) - Total cost: $0.0475/word - Desired margin: 40% - MTPE rate: $0.0475 ÷ (1 − 0.40) = $0.0475 ÷ 0.60 = $0.079/word

Now you have a defensible rate. Your post-editor costs $0.0375, overhead costs $0.01, you price at $0.079, and you keep ~40% as gross margin ($0.031 per word). Subtract taxes, account receivable aging, and project management overhead, and your net margin lands around 20-25%—healthy for a services business.

When to use: You know your post-editor costs, you have a target margin (usually 30-50% gross), and you want a formula you can repeat. This is professional and scalable.

Method 3: Hourly-to-Per-Word Conversion (When Rates Are Hourly)¶

Sometimes you hire post-editors at $40/hour but need to quote per-word. Convert carefully.

Formula:

Per-Word Rate = (Hourly Rate) ÷ (Expected Productivity in WPH)

Example: - Post-editor cost: $45/hour - Expected full PE productivity: 700 wph - Per-word cost: $45 ÷ 700 = $0.064/word

Then apply Method 2 markup. If your desired margin is 40%, your rate: $0.064 ÷ 0.60 = $0.107/word, round to $0.11.

Critical caveat: That 700 wph is not guaranteed. On a bad-MT project, your post-editor might deliver 400 wph, and you’ve underpriced by 40%. Always build in a 15-20% buffer for downside risk.

Method 4: Project-Based Estimation (For Retainers & Ongoing Work)¶

For large retainers or ongoing volumes, calculate total project cost and price as a bundle.

Example: - Project scope: 250K words of full PE across 6 months - Post-editor availability: 40 hours/week - Productivity: 700 wph full PE - Monthly capacity: 40 hours × 4.3 weeks × 700 wph = 120,400 words/month - Total cost for post-editor: 6 months × $30/hour × 40 hours/week × 4.3 weeks = $31,320 - Overhead (QA, platform, management): $4,000 - Total cost: $35,320 - Desired margin: 35% - Project price: $35,320 ÷ (1 − 0.35) = $54,338 (or $0.217/word)

This looks high per-word, but for retainers you’re guaranteeing linguist availability and smoother workflow, so premium pricing is justified.

When to Discount MTPE, When to Premium¶

Now that you have a baseline, here’s where it flexes based on deal structure and risk.

When to Discount¶

Volume discounts: If a client commits to 500K+ words annually, consider 10-15% off your standard rate. The stable revenue justifies the discount.

Guaranteed MT quality: If the client has a trained, specialized MT engine with solid output, you can discount 15-20% because your post-editors will hit higher productivity. Negotiate access to the client’s TM and glossaries upfront.

Retainer/ongoing work: Monthly blocks of work at predictable volume allow you to staff more efficiently. Discount 10-20% for this stability.

Long-term relationship: If this is a growth account and you’re building a relationship, you can be slightly aggressive on entry pricing (e.g., $0.09/word instead of $0.11) to win and prove value.

When to Premium¶

Rush delivery: Same-day or next-day turnaround requires overtime or penalty rates from post-editors. Add 25-50% to your standard rate.

Low-quality MT output: If the client has a generic MT engine (Google Translate, untuned neural MT), your post-editors will slow down to 400-500 wph instead of 700-800. Add 20-30% to your standard rate to cover the extra effort.

Rare language pairs: EN-Finnish, EN-Icelandic, EN-Hungarian—linguist scarcity is real. Add 50-100% premium if you have to recruit or if your post-editor can only dedicate part-time hours.

Specialized domains: Legal, pharmaceutical, financial MTPE isn’t just faster human translation—it’s higher-risk. The client expects publication-ready output and your post-editor needs SME knowledge. Add 20-40% premium.

Tight terminology: If the client has extensive glossaries and CAT tool requirements (Trados, memoQ integration, QA checks), add 10-15% for tool overhead and compliance work.

Worked Example: When to Premium¶

You quote EN-German MTPE. Client says: “Our MT is from Google Translate. We need full PE. Two-week turnaround for 100K words.”

Your baseline for EN-DE full PE: $0.10/word.

Adjustments: - Bad MT (Google): +20% = $0.12/word - Two-week turnaround (tight): +15% = $0.138/word - Round up for confidence: $0.14/word

Revenue: 100K × $0.14 = $14,000.

Cost: Your post-editor at $40/hour, expected 500 wph on bad MT, 100K ÷ 500 = 200 hours = $8,000 cost. Overhead: $1,200. Total cost: $9,200. Margin: $4,800 (34%).

That margin compensates for the higher effort and schedule risk. You’re not giving away value.

Common MTPE Pricing Mistakes¶

Even with a formula, agencies still leave money on the table. Here are the traps.

Mistake 1: Underestimating Bad MT Impact

You quote based on 700 wph productivity. Client sends sample: it’s garbled, inconsistent terminology, poor punctuation. Your post-editors realistically do 450 wph. You’re now -35% on margin and the client is upset because you’re slow.

Fix: Always ask for sample content and run a 500-word trial PE before committing to rate.

Mistake 2: Pricing LPE and FPE the Same

LPE should be 30-50% cheaper than FPE, but some agencies charge the same rate and hope no one notices. Clients will eventually ask for full PE at light PE prices, and you’ll be trapped defending a bad decision.

Fix: Have separate rate cards. LPE at $0.05, FPE at $0.10. Make this explicit in every quote.

Mistake 3: Forgetting QA and Revision Cycles

MTPE still needs QA. You’re not accounting for time to review post-editor output, run automated checks, or handle revision requests. That’s 5-10% overhead you’re eating.

Fix: Charge for QA explicitly or build it into your per-word rate. If you’re quoting $0.10/word, that should include one quality pass. Revision rounds are separate line items.

Mistake 4: Not Adjusting for Linguist Skill

A junior post-editor might do 500 wph; a senior post-editor does 800+ wph. You’re paying the same way but getting different output quality. Either pay performance-based rates or accept that junior editors need higher-margin projects to justify their training.

Fix: Have tiered post-editor rates: junior ($25-30/hour), mid ($35-40/hour), senior ($50-60/hour). Price accordingly and match skill to project risk.

Mistake 5: Losing Track of Scope Creep

Client says “full PE” but means “make it perfect.” You deliver publication-ready output, they ask for three revision rounds, and you’ve blown your margin.

Fix: Define full PE clearly in the SOW. Example: “Full PE per ISO 18587 includes one quality assurance pass. Additional revisions are charged at $X per round.”

Mistake 6: Not Accounting for Vendor Lock-In

Client sends 50K words as a trial at $0.10/word. If it goes well, they’ll ask you to cut rate to $0.08 for the next 500K. You’re tempted because turning it down means zero revenue.

Fix: Price entry projects fairly. If $0.10 is your minimum profitable rate for that language/domain, don’t drop it. If you do, have a clear policy: volume discounts tier up at 250K+ words, not down from the first order.

Tools & Templates for Managing Your Rate Card¶

You need a system to track costs, maintain rates, and iterate as market conditions change. Here are practical options.

Spreadsheet Template (Simplest)¶

Build a single Google Sheet with these columns:

Update quarterly as post-editor rates change or as you gather actual productivity data. Version it; don’t let rates drift.

Python Calculator (If You Want Automation)¶

For agencies managing 20+ language pairs, a simple script beats spreadsheets:

def calculate_mtpe_rate(hourly_cost, expected_wph, overhead_pct, desired_margin_pct):
    cost_per_word = hourly_cost / expected_wph
    total_cost_per_word = cost_per_word * (1 + overhead_pct)
    rate = total_cost_per_word / (1 - desired_margin_pct)
    return round(rate, 4)

# Example
rate = calculate_mtpe_rate(
    hourly_cost=35,
    expected_wph=750,
    overhead_pct=0.15,
    desired_margin_pct=0.40
)
print(f"Your MTPE rate: ${rate}/word")  # Output: Your MTPE rate: $0.0783/word

Store your language pairs, domains, and target margins in a config dict, and you can regenerate your entire rate card in seconds when costs change.

Guiding Principle: When to Use Per-Word vs Hourly vs Project-Based¶

• Per-word: High-volume, standard MT quality, predictable domains. Best for most MTPE work.
• Hourly: One-off projects, extremely variable MT quality, where you need to bill for the actual time spent regardless of productivity.
• Project-based: Retainers, ongoing work, where you want predictable revenue and the client wants predictable cost.
• Hybrid: Quote per-word but include an hourly floor. Example: “Full PE at $0.10/word or $40/hour, whichever is higher.” Protects you if MT is worse than expected.

FAQ: Common Questions on MTPE Pricing¶

Q: How do I know if my MTPE pricing is competitive?

Research directly. Artlangs reports FPE €0.06-€0.12/word; WeGlot reports $0.08-$0.15/word. Check ProZ and TranslatorsCafe for local post-editor rates. Survey 5-10 competitors in your language pairs. Adjust for your positioning: if you guarantee faster turnaround or SME-level QA, price at the higher end. If you’re new to MTPE, price competitive—don’t undercut by 40%.

Q: What if a client’s MT is terrible?

You have three moves. Option 1: invest in better MT + glossaries and retrain the model specifically for their content. Upfront cost (€500-2,000), but your future projects are faster and you’re a hero. Option 2: charge at the high end of MTPE (€0.10-€0.12 for “full PE from raw machine output”). Option 3: decline the project or recommend full human translation. A client with poor MT and budget pressure is a liability.

Q: Should I charge more for rare language pairs?

Absolutely. EN-Finnish, EN-Icelandic, EN-Greek—linguist supply is tight and turnaround risk is high. Add 50-100% premium over your Tier 1 rates. A client requesting EN-Finnish at $0.08/word is either naïve or trying to squeeze you. Push back: “EN-Finnish full PE at $0.14/word; that’s fair for the available talent.”

Q: How do I handle revision rounds?

Define them in the SOW: “Quote includes one full QA pass. Revisions beyond scope: $X per round at post-editor hourly rate.” Typically, first revision is free if you messed up; subsequent revisions are $30-50/hour or bundled as 5-10% add-on if volume is high.

Q: Can I charge per-hour for MTPE if clients expect per-word?

Yes, but give them a hybrid option: “Full PE at $0.10/word or $45/hour, whichever is higher.” This protects you if MT is worse than expected (you fall back to hourly) while anchoring them to per-word for normal cases. Clients appreciate clarity.

Q: Should I offer discounts for retainers?

Yes, 10-20% is standard. A client committing to 50K words/month at $0.10/word gets charged $0.085-$0.090. You lock in revenue, they lock in cost, and your post-editors get stable work. Win-win.

Q: How much should I budget for overhead (platform, QA, management)?

Add 12-20% on top of direct post-editor cost. This covers systems (CAT tools, TMS, MinIO storage if you’re validating post-edited translations with automated QA), project management, administrative overhead, and quality assurance. Don’t cheap out here—QA eats time.

Q: What’s the difference between MTPE pricing and MTQA (Machine Translation Quality Assurance)?

MTPE is the actual post-editing work. MTQA is a step after: automated or manual checking to ensure the PE was done right. Some agencies bundle MTQA into their MTPE rate; others charge it separately. Be explicit. Example: “Full PE at $0.10/word includes one automated QA pass; additional manual QA is $0.02/word.”

Putting It Together: Your First Rate Card¶

Let’s assume you’re a mid-size European agency specializing in EN-DE, EN-FR, and EN-ES technical translation. Your post-editors work at €35/hour and you target 40% gross margin.

Step 1: Calculate cost per word for full PE

• EN-DE: €35/hr ÷ 750 wph = €0.0467/word
• EN-FR: €35/hr ÷ 700 wph = €0.050/word
• EN-ES: €35/hr ÷ 750 wph = €0.0467/word

(Assuming post-editors are slightly slower on FR due to higher complexity.)

Step 2: Add overhead (15%)

• EN-DE: €0.0467 × 1.15 = €0.0537/word
• EN-FR: €0.050 × 1.15 = €0.0575/word
• EN-ES: €0.0467 × 1.15 = €0.0537/word

Step 3: Apply 40% margin

Using the formula: Rate = Cost ÷ (1 − 0.40)

• EN-DE: €0.0537 ÷ 0.60 = €0.0895/word → €0.09
• EN-FR: €0.0575 ÷ 0.60 = €0.0958/word → €0.10
• EN-ES: €0.0537 ÷ 0.60 = €0.0895/word → €0.09

Step 4: Light PE rates (40% discount off full PE)

• EN-DE LPE: €0.09 × 0.60 = €0.054 → €0.05
• EN-FR LPE: €0.10 × 0.60 = €0.060 → €0.06
• EN-ES LPE: €0.09 × 0.60 = €0.054 → €0.05

Step 5: Adjustments

• Premium for rush (48-hour turnaround): +30% (add €0.027/word to full PE rates)
• Premium for legal/financial domain: +20% (add €0.018/word)
• Discount for volume (250K+ words/year): -15% (subtract €0.0135/word)

Your final rate card:

Volume discount: -€0.014/word for 250K+ annual commitment.

Now you have a defensible card. Every rate is justified by cost + margin. You’re not guessing.

Monitoring and Iterating Your Rate Card¶

Your rate card isn’t static. Revisit it quarterly.

Collect this data as you work:

• Actual post-editor productivity (log hours and word counts per project).
• Actual revision cycle overhead (track revision time).
• Client churn by rate (do clients balk at €0.10 and accept €0.09?).
• Margin achievement (did you hit your 40% target or land at 25%?).

Adjust quarterly:

If post-editors are consistently hitting 900 wph instead of your assumed 750 wph, your cost per word drops and you can either increase margin or cut price slightly to win more work. If you’re hitting 30% margin instead of 40%, something’s wrong: productivity estimate, overhead category, or hidden revision cycles. Diagnose and fix.

If the market average for EN-FR FPE drops from €0.12 to €0.10 (competition heating up), don’t panic. Raise your productivity or reduce cost somewhere else. Knee-jerk price cuts destroy margins.

Conclusion: Pricing Is Not Guessing¶

MTPE pricing can feel opaque, but it’s not magic. Cost + overhead + desired margin = rate. That formula works for every language pair, every domain, every client. The hard part isn’t the math—it’s knowing your numbers: your post-editor costs, your realistic productivity, your actual overhead, and the margin you need to survive.

Build your rate card once. Test it against a few real projects. Iterate based on actual data. In six months, you’ll have a competitive, defensible pricing strategy that grows your margins without losing clients.

Start today. Open a spreadsheet, plug in your first language pair, and calculate. You’ll be surprised how quickly you move from guessing to strategizing.