Why it will be difficult or impossible for most freelancers to compete with translation brokers for direct clients.

Executive Summary:  With the widespread use of AI by translation agencies, the agency work model considered by most freelance translators to be a given is coming to an end, and low-paid post-editing will be attractive only to translators with no other choices. Acquisition of direct clients is one way to buy some time, but the hurdles are high, and some of the hurdles are probably unexpected by people who might casually say “Well, I guess I’ll just have to go after direct clients.”

As translation brokers order less and less translation from professional translators, one way for some professionals to buy some time before being swallowed up by an AI pogrom is to acquire direct clients, which are generally later adopters of AI than are the brokers, the latter being well on their way to full implementation of a business model calling for AI, followed by post-editing of some variety, to eliminate the need for them to pay for professional translation.

The smaller agencies that will remain and continue to order translation from professionals will be able to support only a tiny portion of the translator population. The big players using AI control most of the market.

But is acquisition of direct clients a reasonable goal for freelancers? It depends, of course, on the individual and on the specific market and language pair, but there are some very solid reasons why it will be extremely difficult or impossible for most individuals to acquire direct clients, and some of the reasons might not be obvious.

Your most valuable asset as a freelancer is the very thing that will make it extremely difficult to acquire direct clients. In competing with translation brokers, most of the major ones of which have extremely few and sometimes no working translators in their employ—this being particularly true of the US brokers selling JA-to-EN translation—your position as an individual translator can be a great selling point. It precisely because you are an individual, however, that you will face extremely difficult hurdles and will not have the freedoms brokers build for themselves by hiding behind an Internet presence and clueless project managers. The most important freedom of brokers is freedom to lie with impunity.

Unlike translation brokers, as a freelancer, lying will not work. Being an individual, you will inevitably be seen as an individual translator by clients, and that will make it impossible to adopt one of the most effective strategies used by brokers to sell their services, bare-faced lying.

Translation brokers have become accustomed to being able to tell the most ridiculous fairy tales about what they can do and how they do it, with almost no risk of being found out. Their clients seldom catch them out. I have discussed some of these lies with some of their clients, and there is usually a very surprised reaction. Many clients don’t think much about what happens after they issue a purchase order.

The lies you cannot get away with as a freelancer are various.

Unlike translation brokers, as a freelancer, you cannot lie about your capacity. Brokers can claim to “have” a team of thousands of translators and do not fear that clients will be able to judge the veracity of that claim. But you as an individual cannot do that.

You’re it. Prospects will know you are just you, even if you form a company and attempt to position yourself as a corporate entity. And you might think you can start up a website claiming to “have” hundreds or even thousands of translators on “your team.” This is a device used by the major translation brokers. The trick is that, although the brokers themselves generally “have” no translators to speak of, their clients have become accustomed to telling themselves that it’s ok, because the translation seller “has” all those translators. As a freelancer, you won’t be able to make that work, unless you wish to lie as do the brokers. Is that something you want to do?

Unlike translation brokers, you cannot hide behind an Internet presence to avoid direct client interaction. This reality is going to be troubling to many freelancers who have become used to working in front of their computers, from “registering” with a broker, through order receipt and job delivery. But that mode of operation needs to be discarded if you expect to succeed at acquiring direct clients. You need to learn what specific people to meet, to get out and meet those people, and to be convincing when you meet them.

Unlike translation brokers, as a freelancer, you cannot lie about your expertise. To avoid losing their clients to freelance translators, because translation brokers almost never allow clients to interact with or even know the names of the translators from whom the brokers purchase translations for resale to clients, clients almost never can verify the actual abilities of a translator, save for a translation test that might even be done by someone other than the translator to whom their work is subsequently given by the broker. Brokers can generally get away with telling clients that they “have” a large number of domain experts in “their” team of translators.

As a freelancer selling translations to a direct client, however, although you might have been able to fool a totally translation-clueless and domain-clueless project manager and survive if their clients are not sharp enough to detect problems, that’s not going to work if you are selling to translation consumers yourself.

If you are positioning yourself as an individual domain expert, as you will need to do to acquire direct clients, you are going to have to demonstrate that expertise in direct interaction with clients who themselves are domain experts. You don’t have the freedom to lie and hope for the best, as is done by translation brokers. And when faced with a direct client across the table, Googling terminology is not going to work.

Unlike translation brokers, as a freelancer, you cannot easily hide. Translation brokers can create beautiful looking websites and the larger once have actual physical addresses, although some mid-level and even major online-only brokers publish virtual addresses purporting to be the address of their company or publish no address at all. They can be very confident that they will not be phoned or visited by a client wanting to speak to a translator.

As an individual, that will not work. Potential clients will need to know where you are, sometimes for tax purposes. In fact, you will almost always need to meet potential clients directly (particularly in markets such as Japan) and to do that you will probably need to discard the idea of operating from a geographically remote location, as discussed below.

Unlike translation brokers, as a freelancer, where you live matters. This is because, unlike brokers, you will need to directly interact with specific individuals in selling your services. This is something that translation brokers can avoid, by simply inserting a clueless project manager into the loop and trusting that clients will assume the project manager can contact the translators for their work.

In order to seek out and interact directly with potential direct clients, you should be operating in close geographic proximity to those clients. Some translators live in areas so remote that such interaction is virtually impossible or at least very costly.

Unlike translation brokers, you will need to seek out specific individuals as sales targets. Whereas the major translation brokers can get away with creating a killer website and waiting for email orders for translation to arrive, that is almost never going to work for individual freelancers, who need to be more pro-active in selling.

You need to approach specific individuals, but you need to learn who those individuals are. Sending emails to a general email address of a potential client is not going to work. General corporate email addresses, even if they are available, are bombarded by countless sales emails and are monitored by gatekeepers, who will effectively block your email from reaching anybody with a need for translation or ordering authority.

You need to expend the necessary effort to learn who the specific people to approach are. It’s known as sales, something which is apparently terrifying to many translators, and means you will need to interact with people outside of the translation business.

If you live in your source-language country and are unable to sell convincingly in the source language, you are at a serious disadvantage. This is one hurdle that will be a deal-breaker for most JA-to-EN translators I know living in Japan. Early-on acquisition of sales-ready spoken ability in your source language will go a long way to enabling your success if you live in your source-language country. Without that ability, you will probably be limited to approaching expats working in foreign firms in your country of residence, and that seriously limits the possibility of success.

The above is but a short catalog of reasons why acquiring direct clients is not a slam dunk. Inertia and the realization of these difficulties could be the major reason for translators going into denial about what is happening to translation because of AI, but it’s happening as sure as anything you can point to in the translation business.

At a translation conference in Cairns in 2019, I gave a presentation on getting from tier two (agency-dependent translation) to tier one (dealing with direct clients). I had some specific suggestions about what is expected of someone working with direct clients. That was significantly before the recent AI translation hype and adoption by agencies, and I must admit I did not realize the suggestions I made would grow in relevance in the ensuing five years.

Some Suggestions for Freelancer Translators

I have written more about this elsewhere, but here are some renewed suggestions to freelance translators who are studiously ignoring the elephant in their living room.

  • In the face of a rapidly crumbling translation agency work model, undertake to avoid denial and avoidance behaviors. They will get you nowhere. Believing something that feels good is not nearly as useful as realizing what is true.
  • Stop complaining about cheap translation brokers. The brokers that control the vast majority of the translation market don’t care, are not listening to you, and won’t change.
  • Discard the notion that continuing to deal with the above-noted translation brokers can be somehow made to work. It cannot. It’s ending.
  • Stop lecturing on social media (including LinkedIn) about how AI is not good enough and will lead to great problems. The people who you think you are addressing cannot be moved and are already figuring out that AI is good enough for them, and the people you need as clients don’t yet need the lectures and, in any event, are not going to social media to find translation providers; they purchase translation from people who actively sell to them.
  • Discard the notion that what you need to survive is to adapt to new technology. Technology alone doesn’t provide the more essential pieces to the puzzle of survival. Much more is needed.
  • Discard the notion that what you need to do to survive can be done sitting in front of your computer. It cannot. Because AI-using brokers have made aiming at direct clients an important survival strategy, the low-risk, recluse-translator approach is no longer useful or effective.
  • Invest in yourself by acquiring field-specific knowledge. You might be able to fool clueless project managers at a broker, but you cannot Google your way out of a tight spot when facing a direct client who is a subject-matter expert.
  • Learn the identities of specific individuals who control translation demand at specific direct clients and interact with those specific people.
  • Consider the path that you have traveled thus far and honestly assess whether you can do what you will need to do to survive, realizing that what it takes might be nearly impossible or extremely painful (sales, getting out from in front of your computer, meeting people, and more).

That’s it. Go for it, or figure out where else to go.

A Business Model for Manufacturing Automobiles That Are Good Enough

Custom Global Mobility is in the business of selling bespoke cars manufactured from the ground up to meet detailed customer requirements. This includes the body, the engine, the power train, and all other parts of the cars they sell.

Although CGM has no manufacturing facilities and has no knowledge or understanding of automotive technology, it positions itself as a car manufacturer producing high-quality vehicles to satisfy detailed customer requirements. In the process of doing that, however, there are a few things going on that their customers don’t know about.

Because the business model CGM follows does not require them to manufacture vehicles themselves and even explicitly calls for the outsourcing of every aspect of vehicle production, the company is aiming to drastically reduce the costs normally associated with a traditional manufacturing operation by outsourcing everything.

CGM boasts of a global team of automotive experts and claims to have thousands of expert automotive design and manufacturing people on their team, turning out best-in-class cars precisely meeting customer requirements. A look under the CGM hood reveals quite a different story.

Since CGM has no automotive manufacturing personnel, they must find and use vendors to design and manufacture the vehicles they sell. Those tasks are complicated by CGM themselves having almost no expertise in automotive manufacturing or even any knowledge of (and ability to evaluate) automobiles. But they undertake to overcome these deficiencies by what they call smart outsourcing.

Since they lack in-house expertise, even the task of evaluating and certifying vendors is outsourced to other vendors. For a manufacturing vendor, the typical sequence is that a vendor candidate is given manufacturing drawings for an auto part (which are generated by outsourcing to one of CGM’s design vendors) and is asked to deliver a part for evaluation. For example, as a test, CGM might ask a candidate to manufacture a con rod.

Naturally, since CGM cannot judge the quality of the trial con rod and doesn’t itself have an incoming inspection department, it sends the part to another vendor—sometimes a vendor they use just for such very certification—for evaluation. If the test evaluation vendor gives thumbs up, the candidate is certified as a manufacturing vendor for CGM.

Such certified vendors have been able to sell parts or assembled vehicles to CGM under the condition that they use only the machine tools specified by CGM. The use of any other milling machines, lathes, or machining centers, for example, is prohibited.

Although a vendor might have some questions during the process of manufacturing, since CGM is totally unknowledgeable about automotive technology, it cannot themselves answer those questions, and usually tells the vendor to just manufacture in accordance with the drawings. For CGM, going back to the vehicle purchaser, who is not an automotive expert and just wants to buy a car, is not an option, since the purchaser would probably not be able to answer such questions from the CGM vendor and might (correctly) conclude that CGM doesn’t know what it’s doing.

Because of a lack of skilled vendors to work cheaply enough to satisfy CGM, they commonly need to use vendors that produce sub-standard products. Although that might sound problematical, the use of such vendors has actually reduced CGM’s overall manufacturing costs. These defective products are sent to other vendors to repair, followed by an outsourced “quality assurance” process, relying on what CGM calls “automotive quality assurance experts.”

In recent years, CGM has started using an automated manufacturing system to produce fully assembled vehicles in-house. They simply dump the client’s requirements for a vehicle into the system, and out pops an assembled vehicle.

Occasionally (and, more seriously, unpredictably), the automatic vehicle-production system used by CGM builds totally faulty parts into vehicles or assembles them incorrectly, so they use vendors (more of those automotive quality assurance experts) to find and correct these problems, sometimes including re-machining and assembly of numerous parts. CGM finds that the abundant availability of automotive quality assurance experts willing to work cheaply enough—combined with the low cost of the initial manufacturing of vehicles by using their automated in-house manufacturing systems—enables their business model to succeed. And the key to all of this is that the vehicles produced are good enough to satisfy customers and cause no safety problems.

If you are a freelance translator, the above should sound quite familiar. If you are a translation consumer, however, you might not realize what goes on after you order a translation from a translation broker, but be aware that, more and more these days, it is likely to be similar to CGM’s approach to manufacturing vehicles. There are better ways of providing products and services to clients.

AI bubble bursting? It’s of little concern and won’t provide relief to JA-EN translators.

Predictions that AI is a bubble that will shortly burst are becoming more common. These predictions appear to be not from investors, but rather mostly from people who stand to lose if AI succeeds.

The bubble for AI investors might indeed burst someday, but it is clear that AI has already succeeded in significantly reducing the need for translation brokers to purchase translations from professional translators, and the brokers are replacing professionals by using AI to create translations in-house, and then having them post-edited. The result has been that some professionals have been left with little translation work, have been reduced to doing low-paid post-editing, or have simply left translation as a way of making a living. These outcomes are a measure of AI’s success.

Professional translators, rather than anticipating with joy the bursting of the AI bubble, should think about current ways to survive in their chosen field of endeavor, in which the work sources (translation brokers) for the majority of freelancers are already rapidly replacing professional translators with AI systems they use themselves.

Neither freelance translators nor their organizations are giving sufficient attention and thought to this situation. Their focus appears to be on pointing out the failings of AI or claiming without evidence that everything will be fine if translators just “adapt” to the new technology. It is clear that things will not be fine and, in fact, it is clear that the model of freelancers getting translation work from brokers is already crumbling. The boiling frogs need to hop out of the AI pot before it is too late.

3D Reponses from Some Freelance Translators to the Use of AI by Translation Brokers

I have expounded at some length elsewhere on some better ways to meet the existential crisis presented to freelance translators by AI, but some translators have their own approaches.

Delusion that clients will come back when they discover how bad AI is (Reality: That AI continues to be used demonstrates that it is often good enough a large portion of the translation that is purchased), and delusion that freelancers can just adopt AI themselves to survive (Reality: They might be able to that, but not unless they radically switch their customer demographic to translation consumers and compete with agencies, and that is going to be possible for only a very small portion of the freelance translator population).

Denial of the major changes in progress and accelerating in the translation business, denial of acquiring direct clients as a useful survival strategy, and denial of the ongoing exodus of translators lacking translation work because of AI.

Desperation, quiet of course, because being openly desperate would reveal the denial behaviors to be no more than a face-saving defense mechanism.

Not all freelancers exhibit all of these reponses. Some are only suffering their desperation in silence.

I’ll say it again. The agency model for freelance translators is ending, and there is very little chance of preventing that from happening.