What Does a Translation Management System (TMS) Do?

A translation management system is a technology with a misleading name. It does not translate text, it does not suggest words and it does not produce quality. What it does is build the administrative skeleton of a translation project: logging the request, assigning the file, review rounds, delivery and archiving.

Contents

For that reason a TMS is not a product the buying side sees directly. You feel its effect indirectly: deadlines that hold, the same document not being asked for a second time, a clear record of which version was approved. Its absence, on the other hand, is far more visible.

Below we set out what this system does, when it is needed and when it turns into an unnecessary layer of cost.

What is a translation management system?

A translation management system is a software layer that keeps a record of translation jobs from beginning to end. It logs when a job was requested, which language pair it involves, how many words it contains, who translated it, who reviewed it and on what date it was delivered.

Alongside that, it stores the shared assets: translation memories, termbases, style guides and client instructions. When these assets sit in a single place, the same rules apply regardless of who does the work. Knowledge tied to an individual turns into knowledge held by the organisation.

Its third function is reporting. Your annual translation volume, its distribution across language pairs, your average turnaround time and your repetition rate all come from here. Without this data, budget planning rests on guesswork.

In short, a TMS is not a translation tool but a supply chain tool. Keeping that distinction in mind also settles what you should expect from it. Expect quality from the system and you will be disappointed; expect order and visibility and you will get what you came for.

The difference between a TMS and a CAT tool

The CAT tool sits on the translator’s desk; it deals with sentences, terms and matches. The TMS sits on the project manager’s desk; it deals with files, dates, people and prices. The two can be connected to each other, but their functions are separate.

A practical example makes the difference clear. If a fifty-page manual is to be split between three translators, splitting the text and putting it back together is the work of the TMS; what keeps terminology consistent across the split parts is the shared memory of the CAT tool. One is a logistics job, the other a language job.

For the buying side, the consequence of this distinction is straightforward: when a supplier says “we use advanced translation technology”, it is worth asking which of the two they mean. The presence of a TMS shows capacity for coordination; the presence of a CAT tool shows capacity for linguistic consistency.

Smaller suppliers may not have a TMS, and on its own that is not a negative sign. In a relationship that involves a handful of documents a year, a TMS contributes little. What is decisive is the scale of the work.

What does a TMS take on in a project?

The table below shows the typical steps of a translation project and the work the system takes on at each of them. Every one of these steps can also be carried out without a TMS; the difference lies in whether they are done by hand or on the record.

The order of the steps hardly ever changes. The only thing that changes is how visible each step is. In a process run by hand, most of the steps live in one person’s memory; when that person goes on leave, the process stalls with them.

Step	What the system does	What it gives you
Request	Receives the file, counts volume and repetition	A fast, consistent quote
Assignment	Identifies the right translator and reviewer	Continuity with the same team
Production	Shares the memory and the terminology	Consistency across documents
Review	Records the review rounds and the comments	Corrections that can be traced
Delivery	Locks the version, releases the file	Less risk of the wrong version
Archiving	Stores the memory and the files	Lower cost on the next job

Traceability: who approved what, and when?

The least discussed but most valuable output of a TMS is traceability. Who approved the final state of a text, who asked for which correction and which version went live are all on the record. When something goes wrong, the discussion rests on the record rather than on recollection.

In regulated sectors this turns into a requirement. In fields such as pharmaceuticals, medical devices, finance and aviation, documentation of the translation process may be called for during audits. A process with no record cannot be evidenced, however well it was carried out.

Traceability also makes changing supplier easier. If the record of your past jobs and your memory files are kept in order, moving to a new supplier does not mean starting from scratch. In a relationship with no record, the cost of switching is high.

That is why a single clause in the contract is usually enough: that the project records and the memory files will be handed over to you on request. Putting that clause in at the outset is easier than arguing about it later.

Who needs a TMS and who does not?

The threshold of scale is settled by a simple question: how many separate translation requests come out of the same organisation in a year? If the number is in single figures, the system is unnecessary. Where there are dozens of requests, more than one language and more than one department, the cost of coordination exceeds the cost of the system.

The productivity gap becomes concrete at this point. In one benchmarking study the hourly output of a freelance translator was measured at 443 words and the hourly output of a corporate language service provider at 5,728 words (DePalma 2012). This gap is not individual speed; it is the result of an organised process.

At the scale of the individual translator the picture is different. The same source gives the average daily output per translator as 2,684 words and the benchmark used in the industry as 2,500 words. At this scale what is needed is not a system but an orderly file archive.

The criterion for the decision is this: if the process repeats, the system earns its keep; if every job is a one-off, it does not. Counting your translation requests over the past year is usually enough to settle the question.

The limits and hidden costs of a TMS

The first limit is this: the system does not produce quality. A bad translation is a bad translation even when it has been perfectly recorded and delivered on time. A TMS only increases the visibility of the process; responsibility for the content stays with people.

The second limit is the burden of setting it up. For the system to be of any use, term lists, style rules and approval flows have to be defined. Where those definitions are not made, the system turns into an expensive file folder and nobody uses it.

The third limit is dependency. If your data sits in the supplier’s system and the conditions for leaving have not been made clear, changing supplier becomes difficult. This is a contractual risk rather than a technical one, and it can be settled at the outset.

The fourth is that the system can make the process rigid. On urgent, small jobs the workflow steps can create needless delay. A well set-up system defines a short cut for small jobs; where none is defined, a two-page letter goes through the same bureaucracy as a fifty-page manual, and nobody wants to keep that up.

What does a TMS tell you when you are assessing a supplier?

A supplier’s use of a system is an indicator of its capacity for scale and continuity. On its own, though, it is not a sufficient criterion; to find out whether the system is genuinely used, you have to ask concrete questions.

The most useful question is about the past: can you show which version of a document you delivered six months ago was approved, and which terminology decisions were taken? A supplier who answers that clearly is using the system in practice, not on paper.

The second question is about handover: in what format and how quickly can you hand over the memory and terminology files? A vague answer shows that the data is not in fact under your control.

The third question is about capacity: how many language pairs and what volume can you run at the same time? The answer to this one shows how quickly your supplier can put a team together on an urgent multilingual job. The system is the most practical way of confirming that this capacity exists in practice rather than on paper.

To see which services call for this kind of process management, you can look at the scope descriptions on the translation services page.

The Türkiye context

Most of the businesses providing translation services in Türkiye are small teams. In this setting, process management is usually run by hand, over email and shared folders. For a relationship that involves a few documents a year that is enough; for organisations that export, that produce multilingual product documentation or that need regular regulatory translation, it falls short.

The second local particularity is applications to public bodies. In translations submitted to official institutions, the date of the document, the person who signed it and its annexes matter. Having this information on the record speeds the process up when a correction or an additional document is asked for later. A pile of files with no record puts the timetable of the application at risk.

The third is the importance of a processing record for documents falling within the scope of Law No. 6698 on the Protection of Personal Data. Who has seen a document containing personal data and when it was destroyed can be shown in a structure that keeps a record of the process; in one that does not, it cannot.

This article is for general information and is not legal advice.

What does this mean for you?

A TMS may not be a product you buy, but it is a good lens for gauging the process maturity of your supplier. Your real gain is the discipline of keeping records, and you can ask for that without a system. The steps below are the shortest route to establishing it.

1. Count how many translation requests you issued over the past year; no infrastructure decision can be taken without knowing the scale.
2. Gather your approved term list into a single file and send that same file with every job.
3. Have it written into the contract that the memory and terminology files belong to you and will be handed over on request.
4. Tie the authority to approve to a single person; approval by committee is the most common cause of delay.
5. Archive the files you receive with a date and a version label; this archive determines the cost of the next job.

If you would like to set up a working arrangement suited to the volume your organisation handles, you can tell us about your annual needs through the quote form, and you can reach similar guides through the knowledge centre.

Frequently asked questions

What is the basic difference between a TMS and a CAT tool?

A CAT tool works with the text itself, while a TMS manages the process: who received, approved and delivered which file, and when.

I have a few documents translated a year: do I need a TMS?

No. At this scale, email and an orderly file archive are enough; the cost of a TMS only pays for itself where there is continuity.

Is a supplier that uses a TMS more expensive?

Not necessarily. There is an infrastructure cost, but because it improves reuse and coordination it can bring the total cost down.

Can I get my own translation memory back through the TMS?

You can if the contract says so. It needs to be put in writing that the memory files belong to you and will be handed over on request.

References

• Pietrzak, P. & Kornacki, M. (2021). Using CAT Tools in Freelance Translation. Routledge Focus, Chapter 2.
• DePalma, D. (2012). Translation Productivity Benchmarks. Common Sense Advisory.
• Strandvik, I. (2017). Evaluation of outsourced translations. European Commission DGT.
• Law No. 6698 on the Protection of Personal Data (2016).