Teamwork, not workflow, makes good translation
If you’ve ever read through translation company websites, you may have at some point been offered a service called Translation, Editing, and Proofreading, abbreviated “TEP.” The process itself sounds good in theory: a translator works on the initial file, and translation standards call for an error rate of no more than 1.5%. The next translator is an editor who will then, in theory, diligently edit the file and reduce the expected error rate all the way to 0%. A copywriter then comes in to clean up any writing problems, such as bad grammar, or translationese, turning it into a highly polished product. To an English speaker, the resulting file looks quite good, but only because a skilled copywriter spun straw into gold. Beneath the veneer of TEP lies a serious and fundamental problem. As research company, Common Sense Advisory, previously pointed out: quality assurance steps do not actually improve quality. This is especially true in my field, Chinese legal translation.
The industry manual General Theory of the Translation Company lays out the general TEP practice: the project manager gets the file to be translated, sends it to the translator, then receives the file, sends it to an editor, receives the file, and sends it to a proofreader. The yo-yo style workflow of e-mail tag routed repeatedly a single person acting as a nexus should seem bizarre to people from a professional setting. Why do translation managers do it? Essentially, the project manager needs to hide the identities of everyone working on the project from everyone else. This is the only way they can prevent employee or client poaching.
If you think about it, in my field of Chinese-to-English translation, if a translator, editor, and proofreader were introduced to each other and to a client, what would the first thing they notice be? They would notice that the project manager, who does not even speak Chinese, is taking a 70% gross commission on the project. That is huge, one of the highest commission rates you see in any industry. Poaching would be a nightmare scenario for the translation company, which is already spending almost half of its revenue on sales & marketing to acquire clients.
Therefore, the project workflow requires placing a Chinese Wall between all stakeholders. Any communications have to go through the project manager, who redacts all of the team member identities from the project. Therefore, it is completely impossible for any semblance of teamwork to emerge. Moreover, the team members have no meaningful input in who they work with; the other person is an anonymous translator the company assures is “very good” even if they are a poorly vetted fraudster.
The anonymity of the system encourages foul play from virtually everyone involved. This may not be true of European language projects, but it’s certainly true of Chinese projects and well known throughout the industry. I’ve audited the work of many translation companies and have seen a constant pattern of foul play in most legal translation related projects: work is constantly faked in Chinese translation at every level.
The project manager is the first one to get involved in foul play. The translator working in TEP scenarios generally has little or no incentive to do a good job, and in fact their goal is usually to patch up a machine translation as quickly as possible, to fake work without being caught cheating. Chinese translators routinely admit in TEP situations, quality work is “the editor’s responsibility,” so they usually rely on the machine translators, assuming that the editor has the knowledge and competence to fix everything. They also assume the proofreader will be responsible for English, so again no need to put any effort into grammar or quality writing. For reasons about to be discussed, the translator in a TEP environment can also easily and completely discredit any changes made by the editor that the project manager complains about. Thus, no matter how bad their quality, they can continue to do bad quality work.
In the second phase of the TEP scenario is the editor, who can personally and easily see that the translator is faking work. Moreover, the project managers will never trust the editor’s assessment, because since the translator and editors are total strangers, not friends or colleagues, they generally seek to destroy each other. This anonymization effect can be seen in the comments sections of blogs all over the internet: anonymous people depersonalize each other and are liable to attack. The editor’s goal, therefore, is to prove they did real work. Typically, they do this by using revision automators such as Grammarly, and text analysis tools to determine where a single ‘replace all’ will redden the text the most. They then do quick drag and drops all over the text, possibly without closely reading the source. Hundreds of changes go up in minutes.
The project manager gets the file covered with bad or even ridiculous changes, then sends it to the translator for their side of the story. The translator calls everything “preference edits” that have nothing to do with the substance of the document. The project manager can’t speak Chinese so usually will just leave it there. The lack of Chinese skill is what sets this language so vulnerable to fraud; there are plentiful French and Spanish speaking project managers will real translation skill in the industry. With notoriously difficult Chinese-to-English translators, the project manager is flying blind.
If the project manager calls in an outside opinion, the third translator will to the surprise of nobody do their best to discredit both the editor and the translator. The TEP process, instead of resembling a sane teamwork process, actually operates as a contemporary version of the Mad Max Thunderdome where each translator battles to come out on top. Very few translation companies can keep Chinese translators employed full time—they don’t have the volume for it—and each of those translators knows that if they discredit the others, fewer people will be fighting for a slice of the pie especially at the top of the value chain. This kind of office politics was portrayed on the highly controversial Chinese television show, Yanxi Palace, and basically reflects the reality of the translation practice.
Sometimes, the editors will tell an opposite lie: the translated file, which is often just a copy of a local Chinese machine translation engine’s output, is totally fine. Then they send it on to the client, who may rely on it to their detriment without noticing.
Amid the chaos, the project manager will often go rogue with some bad acts of their own. Finding every translator deemed unqualified, they begin deliberately sourcing underqualified translators for extremely low prices to throw at the editor who, anyway, is editing with the attitude of Rambo. The editor then uses automated editing software on incoherent English to make it sound decent, and the translation company thinks it’s getting a great deal. This actually turns into a vicious cycle, as it becomes increasingly easy for both the translator and editor to act as if quality is “someone else’s problem”—the translator relies on the editor to fix everything, and the editor can always blame the translator if the volume of mistakes are too high.
The project manager then brings on a proofreader, who will casually guess at what the translation means and put it into good English. They turn it into fantastic quality English that means something totally different from the original, and the most negligent part about it is they don’t even ask the translator if their guesses are right.
The client then gets the file, and often they are an employee at a company with the same flawed approach as the translators before them: translation quality is the translation company’s responsibility, and at any rate the translations don’t look too bad at all. After all, they were proofread. But in reality, they are grossly misleading, essentially being an English native speaker’s wild guess at what translators are trying to say. They treat the translation as being not a part of their core business, and to be outsourced. If the translation company fails them, then they switch translation companies, unaware that the same anonymous translators from the first company are all working at the second. Translation buyers rapidly discover the entire translation market a difficult place to find anything like quality. Chinglish covers everything from restaurant menus all the way to serious fraud litigation in American courts. The modern day e-mail tag workflow creates conflict because it prevents teamwork.
Teamwork: a better way
The revolutionary translation management process I want to introduce is the old fashioned concept I learned when I played sports as child: teamwork. This is something that the TEP model, by attempting to create a production line of Translation, Editing, and Proofreading, has destroyed. Unlike the factory environment where each step of the production line adds something to an unfinished product, the translator’s goal from the very beginning should be to create a finished product. As long as a translator is working in an environment where their work is viewed as an unfinished product, then they will deliberately begin producing unfinished translations.
If we look at firms like McKinsey or Skadden, known for their high quality of work, we see an entirely different attitude. In these firms, a highly experienced partner works with a number of both mid-level and entry-level and other to successfully handle sophisticated projects and deliver great results. The junior employees are trained and expected to do perfect work, but the reality is that only the partner-level expert has the experience and knowledge needed to deliver superior results. This model is used widely across industries, and in fact is the dominant professional service model. People work together, share ideas, and collaborate to avoid making mistakes. Unlike the translation companies to which they turn, these firms do not deliberately produce unfinished work that is “someone else’s problem.”
Conclusion
Many companies these days are offering TEP, short for Translation-Editing-Proofreading, as a translation service package to clients. When they claim it provides expertise, however, these companies cross the line into deceptive marketing practices. The reality of the TEP model is that a project manager centralizes all team member and client contact in themselves, shuffling documents between each contributor so that they are unable to solicit the staffing firm’s employees or its clients. The anonymity provided by this model results in an immense amount of dishonesty, cheating, and deception by the temporary employees provided by the translation staffing agency, leading time and again to extremely poor quality and unreliable translation work. The best translators do not collaborate with each other using the rigid TEP model, but rather operate on a teamwork model driven by trust and collegiality, and a commitment to put the client’s success above all else.
When selecting a translation provider, it’s very dangerous to use a legal translation provider that uses the TEP model instead of teamwork. When selecting a provider, ask to be sure they are completing projects by using teamwork, not through a TEP or similar conveyor-belt type workflow. At CBL, all of our translations are produced with a teamwork model similar to the best law firms.
