Before you onboard a Turkish or Arabic subcontractor, verify five things: who translates and who reviews each job, which language variants the vendor covers natively, how quality is checked and reported, whether the team works inside your CAT tools and vendor portal and how your end client's data is protected. Confirm every answer with a test on real content from your own pipeline. El Turco is a regional vendor for LSPs: vetted native linguists in their own markets, in-house project management and QA, certified to ISO 17100, ISO 18587 and ISO/IEC 27001.
This guide is for LSP vendor managers adding or replacing a regional partner for Türkiye and the Arabic-speaking markets. It covers the language-specific risks to probe, a checklist you can paste into a vendor questionnaire, a test design, go and no-go criteria and the operational terms to agree before the first live job.
Why do Turkish and Arabic subcontractors need their own checklist?
A generic vendor questionnaire asks about rates, capacity and tools. Those answers matter, but they do not tell you whether a subcontractor can handle the problems that make Turkish and Arabic jobs fail in review. Your end client sees the result under your brand, so the checks must target the language itself.
What should you probe for Turkish?
Turkish is agglutinative. Suffixes follow vowel harmony and change with the word they attach to, so placeholders for names, numbers or product titles often need suffix handling or a rewritten string. Turkish has a dotted and a dotless i, which breaks case conversion in code that is not locale-aware. Brands must choose between the formal siz and the informal sen. Ask the vendor how its linguists handle each of these and how they flag strings that cannot be translated correctly as written.
What should you probe for Arabic?
Arabic runs right to left and mixes with Latin brand names, numbers and URLs, so bidirectional text needs checking in context. Modern Standard Arabic (MSA) is the shared written standard, while marketing and game content often call for a Gulf, Levantine, Egyptian or Maghrebi register. Arabic marks gender in second-person address and has a dual form. Ask which variants the vendor staffs with native linguists, how it decides between MSA and a dialect and how it tests right-to-left layouts.
What should you verify before you send a test?
Use the checklist below as a written questionnaire. Ask for evidence rather than adjectives. A vendor that answers with documents and named roles is easier to audit later when your own clients ask questions.
| Check | What to ask for | Evidence that counts | Red flag |
|---|---|---|---|
| Linguist sourcing | Who translates and who reviews each language pair | Native linguists in-house or in a vetted network, with a described vetting process | Work passed to unnamed third parties |
| Variant coverage | Which Arabic variants and which Turkish registers are covered natively | Named variants with native reviewers for each one | Arabic offered as one language with no dialect question |
| Review model | How many linguists touch each job | TEP with a second linguist reviewing every job | Self-review only or review on request |
| QA and reporting | Which automated checks run and what the delivery report contains | A sample QA report and an error typology for LQA | No report with deliveries |
| Certifications | Which ISO standards the vendor holds | Certificate copies for ISO 17100, ISO 18587 for MTPE and ISO/IEC 27001 | Logos without certificates or scope statements |
| Tool fit | Which CAT tools and portals the team works in | Live work in your tool and your vendor portal | A request to email files outside your TMS |
| Security | How end-client content is protected | NDA for every linguist and an ISMS under ISO/IEC 27001 | No answer on where files are stored |
| Capacity and urgency | How small urgent updates and large launches are staffed | A stated rush option and a named account manager | A turnaround promise before scoping |
How should you design the onboarding test?
A short test on your own content tells you more than a long questionnaire. Choose a sample from a real client account with the hard parts included. Send the same brief, style guide and glossary that a live job would carry.
- For Turkish, include strings with placeholders, a marketing line that needs adaptation and content that tests the siz or sen decision.
- For Arabic, include mixed-direction strings with numbers and Latin names, plus a screen or screenshot to review in context.
- Specify the target variant for Arabic and see whether the vendor asks questions about audience and register.
- Ask for the translation, a QA report and a list of queries raised during the work.
- Have the output scored by an independent native reviewer against your own error typology.
- Run the test inside your CAT tool and portal so you also test the handoff.
Which criteria decide go, conditional or no-go?
Agree the decision rules before the test results arrive. A written rule keeps the decision consistent across vendors and across the people on your team who score them.
Go
The test passes your error threshold with no critical errors. The vendor raised sensible queries, returned a QA report and worked inside your tool without workarounds. Certificates and NDA terms match what your end clients require.
Conditional
Quality is close to your threshold, or one area such as right-to-left review needs a process change. Onboard for a limited set of accounts, agree a corrective action and review results after the first live jobs.
No-go
Critical errors in meaning, the wrong Arabic variant, no QA evidence or refusal to work in your tool. Also decline when the vendor cannot say who translated the test.
What should you agree before the first live job?
Most partner problems start with terms that nobody wrote down. Settle the operational details at onboarding so the first urgent job runs on known rules.
- Pricing basis (per-word, hourly or project-based) and how each one is quoted.
- How turnaround is set and what the rush option covers.
- Who owns translation memories and termbases and how they are returned.
- How queries reach your project manager and the end client.
- Which QA checks run on every delivery and in which format the report arrives.
- Confidentiality terms that cover your end clients by name or by account.
- An escalation contact for quality issues and missed deadlines.
What warning signs should stop an onboarding?
Some answers show that a vendor is reselling capacity rather than running its own linguists. Treat these as reasons to pause.
- The same reviewer named for every Arabic variant.
- A price list sent before anyone has seen your content.
- Vague answers about who translated the test.
- No questions raised on a test that contains ambiguous source text.
- Reluctance to share certificate copies under NDA.
How El Turco handles this
El Turco works as a regional vendor for LSPs across 19 languages, including Turkish and Arabic with MSA plus Gulf, Levantine, Egyptian and Maghrebi variants. Work is done by 700+ vetted native linguists who live in the markets they translate for, with project management and QA kept in-house. Translation jobs follow TEP, with a second linguist reviewing every job.
Linguists join the network through screening, a paid test project and mentored onboarding before they are assigned to accounts. Every delivery runs through What's Wrong?, the translation QA tool from Still Needs a Human. Xbench or Verifika reports can be added on request. LQA services follow a documented methodology.
The team works in RWS Trados, memoQ, Phrase, XTM, Smartcat, Crowdin, Transifex and other CAT tools through your vendor portal, API or file exchange. LSPs get a dedicated account manager, one invoice and a free test translation of up to 300 words. Pricing is per-word, hourly or project-based, quoted after scoping. Turnaround is agreed at scoping. LSP references are available under NDA on request.